Draft Technical Assessment Report:

            Midterm Evaluation of Light-Duty
            Vehicle Greenhouse Gas Emission
            Standards and Corporate Average Fuel
            Economy Standards for Model Years
            2022-2025
&EPA
United States
Environmental Protection
Agency
California Environmental Protection Agency
©•Air Resources Board
RNHTSA

-------
 Draft Technical Assessment Report:

  Midterm Evaluation of Light-Duty
  Vehicle Greenhouse Gas Emission
Standards and Corporate Average Fuel
 Economy Standards for Model Years
               2022-2025
           Office of Transportation and Air Quality
           U.S. Environmental Protection Agency

         National Highway Traffic Safety Administration
            U.S. Department of Transportation

                    And

             California Air Resources Board
                                   EPA-420-D-16-900
                                   July 2016

-------
                                                              TOC and Abbreviations
TABLE OF CONTENTS

List of Acronyms	xix

EXECUTIVE SUMMARY	ES-1

Chapter 1: Introduction
  1.1   Purpose of this Report	1-1
  1.2   Building Blocks of the National Program	1-3
     1.2.1   Background on NHTSA's CAFE Program	1-3
     1.2.2   Background on EPA'sGHG Program	1-5
     1.2.3   Background on C ARE'sGHG Program	1-5
  1.3   Background on the National Program	1-5
  1.4   Agencies' Commitment to the Midterm Evaluation (MTE)	1-11
  1.5   Climate Change and Energy Security Drivers for the National Program	1-12
     1.5.1   Climate Change	1-13
      1.5.1.1  Overview of Climate Change Science and Global Impacts	1-13
      1.5.1.2  Overview of Climate Change Impacts in the United States	1-17
      1.5.1.3  Recent U.S. Commitments on Climate Change Mitigation	1-19
      1.5.1.4  Recent California Commitments on Climate Change	1-20
      1.5.1.5  Contribution of Cars and Light Trucks to the U.S. Greenhouse Gas Emissions
      Inventory	1-20
      1.5.1.6  Importance of the National Program in the U. S. Climate Change Program ... 1 -21
     1.5.2   Petroleum Consumption and Energy Security	1-22
      1.5.2.1  Overview of Petroleum Consumption and Energy Security	1 -22
      1.5.2.2  Recent U.S. Commitments on Petroleum  and Energy Security	1-23
      1.5.2.3  Contribution of Cars and Light Trucks to U.S. Petroleum Consumption	1-23
      1.5.2.4  Importance of National Program to Petroleum Consumption and Energy
      Security 1-23

Chapter 2: Overview of the Agencies' Approach to the Draft TAR Analysis
  2.1   Factors Considered in this Report	2-1
  2.2   Gathering Updated Information since the 2012 Final Rule	2-2
    2.2.1   Research Projects Initiated by the Agencies	2-2
    2.2.2   Input from Stakeholders	2-6
      2.2.2.1  Automobile Manufacturers	2-6
      2.2.2.2  Automotive Suppliers	2-6
      2.2.2.3  Environmental Non-governmental Organizations (NGOs) and Consumer Groups
              2-7
      2.2.2.4  State and Local Governments	2-7
    2.2.3   Other Key Data Sources	2-8
  2.3   Agencies' Approach to Independent GHG and CAFE Analyses	2-9

Chapter 3: Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
  3.1   Changes in the Automotive Market	3-2
    3.1.1   Fuel Economy and GHG Emissions	3-2

-------
                                                               TOC and Abbreviations
    3.1.2  Vehicle Sales	3-3
    3.1.3  Gasoline Prices	3-4
    3.1.4  Car and Truck Mix	3-5
    3.1.5  Vehicle Power, Weight, and Footprint	3-7
    3.1.6  Technology Penetration	3-11
  3.2    Compliance with the GHG Program	3-14
  3.3    Compliance with the CAFE Program	3-17
  3.4    Emerging Transportation Developments	3-22

Chapter 4: Baseline and Reference Vehicle Fleets
  4.1    EPA's Baseline and Reference Vehicle Fleets	4-1
    4.1.1  Why does the EPA Establish Baseline and Reference Vehicle Fleets?	4-1
    4.1.2  EPA's 2014 MY Baseline Fleet	4-2
       4.1.2.1   EPA's MY2014 Based MY2022-2025 Reference Fleet	4-9
         4.1.2.1.1  On What Data Are EPA's Reference Vehicle Fleet Volumes Based?	4-10
         4.1.2.1.2  How did the EPA develop the 2014 Baseline and 2022-2025 Reference
         Vehicle Fleet Volumes?	4-11
         4.1.2.1.3  How was the 2014 Baseline Data Merged with the IHS-Polk Data?	4-11
         4.1.2.1.4  How were the IHS-Polk Forecast and the Unforced AEO 2015 Forecast
         Used to Project the Future Fleet Volumes?	4-12
       4.1.2.2   What Are the Sales Volumes and Characteristics of the MY2014 Based
       Reference Fleet?	4-19
       4.1.2.3   What Are the Differences in the Sales Volumes and Characteristics of the
       MY2008 Based and the MY2014 Based Reference Fleets?	4-22
    4.1.3  Relationship Between Fuel Economy and Other Vehicle Attributes	4-26
       4.1.3.1   Recent Studies of the Engineering Tradeoffs between Power and Fuel Economy,
       and Increases in Innovation	4-29
       4.1.3.2   The Role of the Standards in Promoting Innovation	4-32
       4.1.3.3   Potential Ancillary Benefits of GHG-Reducing Technologies	4-34
       4.1.3.4   Estimating Potential Opportunity Costs and Ancillary Benefits	4-36
    4.1.4  Incorporation of the California Zero Emissions Vehicle (ZEV) Program into the
    EPA Reference Fleet	4-37
       4.1.4.1   The ZEV Regulation in OMEGA	4-37
       4.1.4.2   The ZEV Program Requirements	4-43
         4.1.4.2.1  Overview	4-43
         4.1.4.2.2  ZEV Credit Requirement	4-44
         4.1.4.2.3  Proj ected Representative of PHEV and BEV Characteristics for MY2021 -
         2025     4-45
         4.1.4.2.4  Calculation of Incremental ZEVs Needed for ZEV Program Compliance.. 4-
         49
  4.2    Development of the CAFE Light Duty Analysis Fleet	4-53
    4.2.1  Why did NHTSA Develop the Analysis Fleet?	4-53
    4.2.2  How the MY2015  Analysis Fleet Was Developed	4-53
       4.2.2.1   Background	4-53
    4.2.3  NHTSA Decision to use 2015 Foundation for Analysis Fleet	4-54
    4.2.4  Developments in 2015	4-55

-------
                                                                TOC and Abbreviations
    4.2.5  Manufacturer-Provided Information for 2015	4-56
    4.2.6  Other Data	4-57
       4.2.6.1  Redesign/Refresh Schedules	4-57
       4.2.6.2  Technologies	4-58
       4.2.6.3  Engine Utilization	4-58
    4.2.7  Estimated Technology Prevalence in the MY2015 Fleet	4-59
    4.2.8  Engine and Platform Sharing	4-62
       4.2.8.1  Platform Sharing	4-62
       4.2.8.2  Engine Sharing & Inheritance	4-63
    4.2.9  Class Types and Assignment	4-64
       4.2.9.1  Regulatory Class	4-64
       4.2.9.2  Safety Class	4-64
       4.2.9.3  Technology Class	4-64
       4.2.9.4  Technology Cost Class	4-65
    4.2.10  Mass Reduction and Aero Application	4-65
       4.2.10.1  Mass Reduction	4-65
         4.2.10.1.1   Mass Reduction Residual Analysis for Footprint	4-73
         4.2.10.1.2   Mass Reduction Residual Analysis for Low and High Price Platforms 4-76
         4.2.10.1.3   Mass Reduction Residual Trends for Company Heritage	4-78
       4.2.10.2  Aerodynamic Application	4-80
    4.2.11  Projecting Future Volumes for the Analysis Fleet	4-82

Chapter 5: Technology Costs, Effectiveness, and Lead-Time Assessment
  5.1    Overview	5-1
  5.2    State of Technology and Advancements Since the 2012 Final Rule	5-7
    5.2.1  Individual Technologies and Key Developments	5-7
    5.2.2  Engines: State of Technology	5-12
       5.2.2.1  Overview of Engine Technologies	5-13
       5.2.2.2  Sources of Engine Effectiveness Data	5-15
       5.2.2.3  Low Friction Lubricants (LUB)	5-16
       5.2.2.4  Engine Friction Reduction (EFR1, EFR2)	5-17
       5.2.2.5  Cylinder Deactivation (DEAC)	5-17
       5.2.2.6  Variable Valve Timing (VVT) Systems	5-17
         5.2.2.6.1  Intake Cam Phasing (ICP)	5-18
         5.2.2.6.2  Coupled Cam Phasing (CCP)	5-18
         5.2.2.6.3  Dual Cam Phasing (DCP)	5-18
         5.2.2.6.4  Variable Valve Lift (VVL)	5-18
       5.2.2.7  GDI, Turbocharging, Downsizing and Cylinder Deactivation	5-19
       5.2.2.8  EGR	5-28
       5.2.2.9  Atkinson Cycle	5-29
       5.2.2.10  Miller Cycle	5-33
       5.2.2.11  Light-duty Diesel Engines	5-36
       5.2.2.12  Thermal Management	5-39
       5.2.2.13  Reduction of Friction and Other Mechanical Losses	5-40
       5.2.2.14  Potential Longer-Term Engine Technologies	5-41
    5.2.3  Transmissions:  State of Technology	5-42
                                           in

-------
                                                            TOC and Abbreviations
  5.2.3.1  Background	5-42
  5.2.3.2  Transmissions: Summary of State of Technology and Changes since the FRM 5-
  43
  5.2.3.3  Sources of Transmission Effectiveness Data	5-44
  5.2.3.4  Sources of GHG Emission Improvements: Reduction in Parasitic Losses, Engine
  Operation, and Powertrain System Design	5-46
  5.2.3.5  Automatic Transmissions (ATs)	5-48
  5.2.3.6  Manual Transmissions (MTs)	5-51
  5.2.3.7  Dual Clutch Transmissions (DCTs)	5-52
  5.2.3.8  Continuously Variable Transmissions (CVTs)	5-53
  5.2.3.9  Transmission Parasitic Losses	5-56
    5.2.3.9.1   Losses in ATs	5-56
    5.2.3.9.2   Losses in DCTs	5-56
    5.2.3.9.3   Losses in CVTs	5-57
    5.2.3.9.4   Neutral Idle Decoupling	5-57
  5.2.3.10   Transmission  Shift Strategies	5-58
  5.2.3.11   Torque Converter Losses and Lockup Strategy	5-58
5.2.4   Electrification: State of Technology	5-59
  5.2.4.1  Overview of Electrification Technologies	5-62
  5.2.4.2  Non-Battery Components of Electrified Vehicles	5-64
    5.2.4.2.1   Propulsion Components	5-65
    5.2.4.2.2   Power Electronics	5-66
    5.2.4.2.3   Industry Targets for Non-Battery Components	5-69
  5.2.4.3  Developments in Electrified Vehicles	5-71
    5.2.4.3.1   Non-hybrid Stop-Start	5-71
    5.2.4.3.2   Mild Hybrids	5-74
    5.2.4.3.3   Strong Hybrids	5-79
    5.2.4.3.4   Plug-in Hybrids	5-82
    5.2.4.3.5   Battery Electric Vehicles	5-92
  5.2.4.4  Developments in Electrified Vehicle Battery Technology	5-103
    5.2.4.4.1   Battery Chemistry	5-104
    5.2.4.4.2   Pack Topology, Cell Capacity and Cells per Module	5-106
    5.2.4.4.3   Usable Energy Capacity	5-110
    5.2.4.4.4   Thermal Management	5-115
    5.2.4.4.5   Pack Voltage	5-116
    5.2.4.4.6   Electrode Dimensions	5-117
    5.2.4.4.7   Pack Manufacturing Volumes	5-118
    5.2.4.4.8   Potential Impact of Lithium Demand on Battery Cost	5-121
    5.2.4.4.9   Evaluation of 2012 FRM Battery Cost Projections	5-122
  5.2.4.5  Fuel Cell Electric Vehicles	5-128
    5.2.4.5.1   Introduction to FCEVs	5-128
    5.2.4.5.2   FCEV Cost Estimation	5-130
       5.2.4.5.2.1  Fuel Cell System Cost	5-131
       5.2.4.5.2.2  Hydrogen Storage Cost	5-135
       5.2.4.5.2.3  Combined Fuel Cell and Hydrogen Storage Systems Cost	5-135
       5.2.4.5.2.4  Market Projections	5-137
                                       IV

-------
                                                           TOC and Abbreviations
    5.2.4.5.3  FCEV Performance Status and Targets	5-139
    5.2.4.5.4  Onboard Hydrogen Storage Technology	5-141
    5.2.4.5.5  FCEV Commercialization Status	5-141
    5.2.4.5.6  Outlook for National FCEV Launch	5-142
5.2.5   Aerodynamics: State of Technology	5-143
  5.2.5.1  Background	5-143
  5.2.5.2  Aerodynamic Technologies in the FRM	5-143
  5.2.5.3  Developments since the FRM	5-144
    5.2.5.3.1  Industry Developments	5-145
    5.2.5.3.2  Joint Test Program with Transport Canada	5-147
    5.2.5.3.3  CARS Control-Tec Study	5-150
    5.2.5.3.4  EPA Study of Certification Data	5-150
    5.2.5.3.5  Conclusions	5-152
5.2.6   Tires: State of Technology	5-152
  5.2.6.1  Background	5-152
  5.2.6.2  Tire Technologies in the FRM	5-153
  5.2.6.3  Developments since the FRM	5-154
    5.2.6.3.1  Industry Developments	5-155
    5.2.6.3.2  Control-Tec Analysis of Trends in Tire Technologies	5-157
    5.2.6.3.3  Canada Tire Testing Program	5-157
  5.2.6.4  Conclusions	5-158
5.2.7   Mass Reduction: State of Technology	5-158
  5.2.7.1  Overview of Mass Reduction Technologies	5-158
  5.2.7.2  Developments since the 2012 FRM	5-162
  5.2.7.3  Market Vehicle Implementation of Mass Reduction	5-163
  5.2.7.4  Holistic Vehicle Mass Reduction and Cost Studies	5-166
    5.2.7.4.1  EPA Holistic Vehicle Mass Reduction/Cost Studies	5-169
       5.2.7.4.1.1  Phase 2 Low Development Midsize CUV Updated Study and
       Supplement 5-170
       5.2.7.4.1.2  Light Duty Pickup Truck Light-Weighting Study	5-173
    5.2.7.4.2  NHTSA Holistic Vehicle Mass Reduction/Cost Studies	5-176
       5.2.7.4.2.1  Updated Midsize Car Lightweight Vehicle Study	5-176
       5.2.7.4.2.2  Light Duty Pickup Truck Light-Weighting Study	5-179
    5.2.7.4.3  ARE Holistic Vehicle Mass Reduction/Cost Study	5-184
    5.2.7.4.4  Aluminum Association Midsize CUV Aluminum BIW Study	5-185
    5.2.7.4.5  DOE/Ford/Magna MMLV Mach 1 and Mach 2 Lightweighting Research
    Projects  5-187
       5.2.7.4.5.1  Mach I	5-189
       5.2.7.4.5.2  Mach 2	5-192
    5.2.7.4.6  Technical Cost Modeling Report by DOE/INL/IBIS on 40 Percent-45
    Percent Mass Reduced Vehicle	5-194
    5.2.7.4.7  Studies to Determine Mass Add for IfflS Small Overlap	5-195
       5.2.7.4.7.1  NHTSA Mass Add Study for a Passenger Car to Achieve a "Good"
       Rating on the IfflS Small Overlap	5-196
       5.2.7.4.7.2  Transport Canada Mass Add Study for a Light Duty Truck to Achieve a
       "Good" Rating on the IIHS Small Overlap	5-197

-------
                                                              TOC and Abbreviations
  5.2.8   State of Other Vehicle Technologies	5-200
    5.2.8.1   Electrified Power Steering: State of Technology	5-200
       5.2.8.1.1  Electrified Power Steering in the 2012 FRM	5-200
       5.2.8.1.2  Developments since the FRM	5-200
    5.2.8.2   Improved Accessories: State of Technology	5-200
    5.2.8.3   Secondary Axle Disconnect: State of Technology	5-201
       5.2.8.3.1  Background	5-201
       5.2.8.3.2  Secondary Axle Disconnect in the FRM	5-202
       5.2.8.3.3  Developments since the FRM	5-203
    5.2.8.4   Low-Drag Brakes: State of Technology	5-206
       5.2.8.4.1  Background	5-206
       5.2.8.4.2  Low Drag Brakes in the FRM	5-206
       5.2.8.4.3  Developments since the FRM	5-206
  5.2.9   Air Conditioning Efficiency and Leakage Credits	5-207
    5.2.9.1   A/C Efficiency Credits	5-208
       5.2.9.1.1  Background on the A/C Efficiency Credit Program	5-208
       5.2.9.1.2  Idle Test Procedure	5-208
       5.2.9.1.3  AC 17 Test Procedure	5-209
       5.2.9.1.4  Manufacturer Uptake of A/C Efficiency Credits since the 2012 FRM.. 5-210
       5.2.9.1.5  Evaluation of the AC 17 Test Procedure	5-211
       5.2.9.1.6  Conclusions and Future Work	5-215
    5.2.9.2   A/C Leakage Reduction and Alternative  Refrigerant Substitution	5-216
       5.2.9.2.1  Leakage	5-216
       5.2.9.2.2  Low-GWP Refrigerants	5-216
       5.2.9.2.3  Conclusions	5-218
  5.2.10  Off-cycle Technology Credits	5-218
    5.2.10.1   Off-cycle Credits Program	5-218
       5.2.10.1.1   Off-cycle Credits Program Overview	5-218
    5.2.10.2   Use of Off-cycle Technologies to Date	5-220
5.3     GHG Technology Assessment	5-223
  5.3.1   Fundamental Assumptions	5-223
    5.3.1.1   Technology Time Frame and Measurement Scale for Effectiveness and Cost... 5-
    223
    5.3.1.2   Performance Assumptions	5-224
    5.3.1.3   Fuels	5-227
    5.3.1.4   Vehicle Classification	5-228
  5.3.2   Approach for Determining Technology Costs	5-229
    5.3.2.1   Direct Manufacturing Costs	5-229
       5.3.2.1.1  Costs from Tear-down Studies	5-229
       5.3.2.1.2  Electrified Vehicle Battery Costs	5-231
       5.3.2.1.3  Specific DMC Changes since the 2012 FRM	5-232
       5.3.2.1.4  Approach to Cost Reduction through Manufacturer Learning	5-232
    5.3.2.2   Indirect Costs	5-237
       5.3.2.2.1  Methodologies for Determining Indirect Costs	5-237
       5.3.2.2.2  Indirect Cost Estimates Used in this  Analysis	5-239
    5.3.2.3   Maintenance and Repair Costs	5-243
                                         VI

-------
                                                            TOC and Abbreviations
     5.3.2.3.1  Maintenance Costs	5-243
     5.3.2.3.2  Repair Costs	5-244
  5.3.2.4   Costs Updated to 2013 Dollars	5-245
5.3.3   Approach for Determining Technology Effectiveness	5-245
  5.3.3.1   Vehicle Benchmarking	5-246
     5.3.3.1.1  Detailed Vehicle Benchmarking Process	5-246
       5.3.3.1.1.1  Engine Testing	5-247
       5.3.3.1.1.2  Transmission Testing	5-248
     5.3.3.1.2  Development of Model Inputs from Benchmarking Data	5-251
       5.3.3.1.2.1  Engine Data	5-251
       5.3.3.1.2.2  Engine Map	5-251
       5.3.3.1.2.3  Inertia	5-252
       5.3.3.1.2.4  Transmission Data	5-253
       5.3.3.1.2.5  Gear Efficiency and Spin Losses	5-253
       5.3.3.1.2.6  Torque Converter	5-254
     5.3.3.1.3  Vehicle Benchmarking Summary	5-255
  5.3.3.2   ALPHA Vehicle Simulation Model	5-256
     5.3.3.2.1  General ALPHA Description	5-256
     5.3.3.2.2  Detailed ALPHA Model Description	5-257
       5.3.3.2.2.1  Ambient System	5-258
       5.3.3.2.2.2  Driver System	5-258
       5.3.3.2.2.3  Powertrain System	5-259
       5.3.3.2.2.3.1   Engine Subsystem	5-259
       5.3.3.2.2.3.2  Electric Subsystem	5-260
       5.3.3.2.2.3.3   Accessories Subsystem	5-261
       5.3.3.2.2.3.4  Transmission Subsystem	5-261
       5.3.3.2.2.3.4.1   Transmission Gear Selection	5-261
       5.3.3.2.2.3.4.2   Clutch Model	5-262
       5.3.3.2.2.3.4.3   Gearbox Model	5-262
       5.3.3.2.2.3.4.4   Torque Converter Model	5-262
       5.3.3.2.2.3.4.5   Automatic Transmission & Controls	5-262
       5.3.3.2.2.3.4.6   DCT Transmission & Control	5-263
       5.3.3.2.2.3.4.7   CVT Transmission & Control	5-263
       5.3.3.2.2.3.4.8   Driveline	5-263
       5.3.3.2.2.3.5   Vehicle System	5-263
     5.3.3.2.3  Energy Auditing	5-264
     5.3.3.2.4  ALPHA Simulation Runs	5-265
     5.3.3.2.5  Post-processing	5-265
     5.3.3.2.6  Vehicle Component Vintage	5-266
     5.3.3.2.7  Additional Verification	5-267
  5.3.3.3   Determining Technology Effectiveness for MY2022-2025	5-268
  5.3.3.4   Lumped Parameter Model	5-271
     5.3.3.4.1  Lumped Parameter Model Usage in OMEGA	5-272
5.3.4   Data and Assumptions Used in GHG Assessment	5-275
  5.3.4.1   Engines: Data and Assumptions for this Assessment	5-275
     5.3.4.1.1  Low Friction Lubricants (LUB)	5-276
                                      vn

-------
                                                         TOC and Abbreviations
  5.3.4.1.2  Engine Friction Reduction (EFR1,EFR2)	5-276
  5.3.4.1.3  Cylinder Deactivation (DEAC)	5-277
  5.3.4.1.4  Intake Cam Phasing (ICP)	5-278
  5.3.4.1.5  Dual Cam Phasing (DCP)	5-279
  5.3.4.1.6  Discrete Variable Valve Lift (DVVL)	5-279
  5.3.4.1.7  Continuously Variable Valve Lift (CVVL)	5-279
  5.3.4.1.8  Investigation of Potential Future Non-HEV Atkinson Cycle Engine
  Applications	5-280
  5.3.4.1.9  GDI, Turbocharging, Downsizing	5-283
5.3.4.2  Transmissions: Data and Assumptions for this Assessment	5-294
  5.3.4.2.1  Assessment of Automated Transmissions (AT, AMT, DCT, CVT)	5-295
  5.3.4.2.2  Technology Applicability and Costs	5-299
5.3.4.3  Electrification: Data and Assumptions for this Assessment	5-300
  5.3.4.3.1  Cost and Effectiveness for Non-hybrid Stop-Start	5-300
  5.3.4.3.2  Cost and Effectiveness for Mild Hybrids	5-301
  5.3.4.3.3  Cost and Effectiveness for Strong Hybrids	5-302
  5.3.4.3.4  Cost and Effectiveness for Plug-in Hybrids	5-304
  5.3.4.3.5  Cost and Effectiveness for Electric Vehicles	5-304
  5.3.4.3.6  Cost of Non-Battery Components for xEVs	5-305
  5.3.4.3.7  Cost of Batteries for xEVs	5-313
    5.3.4.3.7.1   Battery Sizing Methodology for BEVs and PHEVs	5-313
    5.3.4.3.7.2   Battery Sizing Methodology for HEVs	5-341
    5.3.4.3.7.3   ANL BatPaC Battery Design and Cost Model	5-341
    5.3.4.3.7.4   Assumptions and Inputs to BatPaC	5-343
    5.3.4.3.7.5   Battery Cost Projections for xEVs	5-346
    5.3.4.3.7.6   Discussion of Battery Cost Projections	5-354
    5.3.4.3.7.7   Battery Pack Costs Used in OMEGA	5-355
    5.3.4.3.7.8   Electrified Vehicle Costs Used In OMEGA (Battery + Non-battery
    Items)      5-361
5.3.4.4  Aerodynamics: Data and Assumptions for this Assessment	5-363
5.3.4.5  Tires: Data and Assumptions for this Assessment	5-364
5.3.4.6  Mass Reduction: Data and Assumptions for this Assessment	5-365
  5.3.4.6.1  Cost Curves	5-365
    5.3.4.6.1.1   Cost Curve for Cars and CUVs	5-368
    5.3.4.6.1.2   Cost Curve for Light Duty Trucks	5-383
  5.3.4.6.2  Mass Reduction in the Baseline MY2014 Fleet	5-394
    5.3.4.6.2.1   Vehicles with MY2008 and MY2014 Production	5-395
    5.3.4.6.2.2   MY2014 Vehicles without MY2008 Counterparts	5-399
    5.3.4.6.2.3   MY2014 Cost Curve Adjustments Due to Vehicle Baseline MY2014-
    MY2008 Curb Weight Differences	5-399
    5.3.4.6.2.4   Safety Regulation Mass Increase Estimate Post MY2014	5-402
  5.3.4.6.3  Effectiveness of Mass Reduction	5-403
  5.3.4.6.4  Mass Reduction Costs used in OMEGA	5-403
5.3.4.7  Other Vehicle Technologies	5-411
  5.3.4.7.1  Electrified Power Steering: Data and Assumptions for this Assessment	5-
  411
                                   Vlll

-------
                                                             TOC and Abbreviations
       5.3.4.7.2  Improved Accessories: Data and Assumptions for this Assessment	5-412
       5.3.4.7.3  Secondary Axle Disconnect: Data and Assumptions for this Assessment.. 5-
       412
       5.3.4.7.4  Low Drag Brakes: Data and Assumptions for this Assessment	5-413
    5.3.4.8  Air Conditioning: Data and Assumptions for this Assessment	5-413
    5.3.4.9  Cost Tables for Individual Technologies Not Presented Above	5-413
5.4    CAFE Technology Assessment	5-415
  5.4.1   Technology Costs Used in CAFE Assessment	5-415
    5.4.1.1  Direct Costs	5-415
       5.4.1.1.1  Improved Low Friction Lubricants and Engine Friction Reduction Levels 2
       &3 (LUBEFR2 & LUBFFR3)	5-415
       5.4.1.1.2  Automatic Transmission Improvements Levels 1 & 2 (ATI1 & ATI2) 5-416
       5.4.1.1.3  High Compression Ratio Engine	5-416
       5.4.1.1.4  Advanced Diesel Engine (ADSL) Engine	5-416
       5.4.1.1.5  7-speed Manual Transmission	5-416
       5.4.1.1.6  6-speed Automatic Transmission	5-416
       5.4.1.1.7  8-speed Automatic Transmission	5-417
       5.4.1.1.8  6-speed Dual Clutch Transmission	5-417
       5.4.1.1.9  8-speed Dual Clutch Transmission	5-417
       5.4.1.1.10  Continuously Variable Transmission	5-418
       5.4.1.1.11  Belt Integrated Starter Generator	5-418
       5.4.1.1.12  Crank Integrated Starter Generator	5-418
       5.4.1.1.13  Electric Power Steering	5-418
       5.4.1.1.14  Improved Accessories (IACC1 & IACC2)	5-418
       5.4.1.1.15  Low Drag Brakes	5-419
       5.4.1.1.16  Secondary Axle Disconnect	5-419
       5.4.1.1.17  Low Rolling Resistance Tires	5-419
       5.4.1.1.18  Aerodynamic Drag Reduction	5-419
       5.4.1.1.19  Mass Reduction	5-419
         5.4.1.1.19.1  Light Duty Pickup Truck Light-Weighting Study	5-424
    5.4.1.2  Indirect Costs	5-428
       5.4.1.2.1  Methodologies for Determining Indirect Costs	5-428
       5.4.1.2.2  Indirect Cost Multipliers Used in this Analysis	5-430
       5.4.1.2.3  NHTS As Application of Learning Curves	5-434
    5.4.1.3  Technology Cost  Summary Tables	5-438
       5.4.1.3.1  Basic Gasoline Engine Costs	5-439
       5.4.1.3.2  Gasoline Turbo Engine Costs	5-443
       5.4.1.3.3  Other Advanced Gasoline Engine Technologies	5-446
       5.4.1.3.4  Diesel Engine Costs	5-448
       5.4.1.3.5  Transmission Costs	5-449
       5.4.1.3.6  Electric Vehicle and Accessory Costs	5-453
       5.4.1.3.7  Vehicle Technology Costs	5-456
  5.4.2   Technology Effectiveness Modeling Method and Data Used in CAFE Assessment 5-
  457
    5.4.2.1  Volpe  Model Background	5-458
    5.4.2.2  Autonomie Vehicle Simulation Tool	5-460
                                        IX

-------
                                                           TOC and Abbreviations
  5.4.2.2.1  Overview	5-460
  5.4.2.2.2  Plant Model Overview	5-462
     5.4.2.2.2.1  Internal Combustion Engine Model	5-462
     5.4.2.2.2.2  Transmission Models	5-464
     5.4.2.2.2.3  Electric Machine Models	5-467
     5.4.2.2.2.4  Energy Storage Models	5-467
     5.4.2.2.2.5  Chassis Models	5-468
     5.4.2.2.2.6  Tire Models	5-469
     5.4.2.2.2.7  Auxiliaries Model	5-469
     5.4.2.2.2.8  Driver Models	5-469
     5.4.2.2.2.9  Environment Models	5-469
  5.4.2.2.3  Control Overview	5-470
     5.4.2.2.3.1  Transmission Shifting Algorithm	5-470
     5.4.2.2.3.2  Torque Converter Lock-up Assumptions	5-478
     5.4.2.2.3.3  Fuel Cut-off Algorithm	5-480
     5.4.2.2.3.4  Vehicle Level Control for Electrified Powertrains	5-480
5.4.2.3  Vehicle Model Validation	5-491
  5.4.2.3.1  Vehicle Benchmarking	5-491
  5.4.2.3.2  Vehicle Validation Examples	5-495
     5.4.2.3.2.1  Transmission Shifting Algorithm	5-495
     5.4.2.3.2.2  Powersplit HEV	5-498
  5.4.2.3.3  Pre-transmission HEV	5-499
     5.4.2.3.3.1  Range Extender PHEV	5-500
5.4.2 A  Simulation Modeling Study Overview	5-501
5.4.2.5  Selection of Technologies for Modeling	5-502
5.4.2.6  Modeling Assumptions	5-503
  5.4.2.6.1  Vehicle Level	5-503
  5.4.2.6.2  Gasoline and Diesel Engines	5-504
5.4.2.7  Description of Engine Technologies Evaluated	5-512
  5.4.2.7.1  Friction reduction	5-512
  5.4.2.7.2  Cylinder Deactivation	5-513
  5.4.2.7.3  Turbocharged Engines	5-513
5.4.2.8  Transmissions	5-514
5.4.2.9  Torque Converter	5-520
5.4.2.10  Electric Machines	5-521
  5.4.2.10.1   Energy Storage Systems	5-523
  5.4.2.10.2   Fuel Cell Systems	5-524
5.4.2.11  Light-weighting	5-525
5.4.2.12  Rolling Resistance	5-525
5.4.2.13  Aerodynamic	5-525
5.4.2.14  Accessory Loads	5-526
5.4.2.15  Driver	5-526
5.4.2.16  Electrified Powertrains	5-526
  5.4.2.16.1   Electrified Powertrain Configurations	5-527
  5.4.2.16.2   Parallel Hybrid Vehicle	5-528
  5.4.2.16.3   Power Split Hybrid Vehicle	5-529

-------
                                                               TOC and Abbreviations
           5.4.2.16.3.1  Voltec Gen 1 Plug-in Hybrid Vehicle	5-530
         5.4.2.16.4  Series Fuel Cell HEV	5-532
         5.4.2.16.5  Powertrain Electrification Selection	5-533
       5.4.2.17  Drive Cycles and Vehicle Simulation Conditions	5-534
       5.4.2.18  Vehicle Sizing Process	5-534
       5.4.2.19  Autonomie Outputs	5-541
       5.4.2.20  Individual Vehicle Simulation Quality Check	5-543

Chapter 6: Assessment of Consumer Acceptance of Technologies that Reduce Fuel
Consumption and GHG Emissions
  6.1    Introduction	6-1
  6.2    Effects of the Standards on Vehicle Sales	6-1
     6.2.1   Overview of Vehicle Market	6-1
     6.2.2   Consumer Vehicle Choice Modeling and Recent Research	6-2
       6.2.2.1  EPA's Efforts in Developing and Assessing a Consumer Vehicle Choice Model
               6-3
  6.3    Conceptual Framework for Evaluating Consumer Impacts	6-5
  6.4    Consumer Response to Vehicles Subject to the Standards	6-9
     6.4.1   Recent New Vehicles	6-9
       6.4.1.1  Sales	6-9
       6.4.1.2  Evaluations from Professional Auto Reviewers	6-10
       6.4.1.3  Consumer Responses to New Vehicles	6-12
     6.4.2   MY2022-25 Vehicles	6-13
  6.5    Impacts of the Standards on Vehicle Affordability	6-16
     6.5.1   Effects on Lower-Income Households	6-16
     6.5.2   Effects on the Used Vehicle Market	6-16
     6.5.3   Effects on Access to Credit	6-19
     6.5.4   Effects on Low-Priced Cars	6-20
     6.5.5   Conclusion	6-22

Chapter 7: Employment Impacts
  7.1    Introduction	7-1
  7.2    Employment in the Auto Sector in Recent Years	7-1
  7.3    Current State of Knowledge of Employment in the Automotive Sector Based on the
  Peer-Reviewed Literature	7-4
     7.3.1   Regulatory Effects at the Firm Level	7-4
     7.3.2   Regulatory Effects at the Industry Level	7-5
     7.3.3   Peer-Reviewed Literature	7-6
  7.4    Employment Impacts in the Motor Vehicle and Parts Manufacturing Sector	7-7
     7.4.1   The Output Effect	7-7
     7.4.2   The Substitution Effect	7-7
     7.4.3   Summary of Employment Effects in the Motor Vehicle Sector	7-12
     7.4.4   Motor Vehicle Parts Manufacturing Sector	7-12
  7.5    Employment Impacts in Other Affected Sectors	7-12
     7.5.1   Effects on Employment for Auto Dealers	7-12
     7.5.2   Effects on Employment for Fuel Suppliers	7-13
                                          XI

-------
                                                                TOC and Abbreviations
    7.5.3  Effects on Employment due to Impacts on Consumer Expenditures	7-13
  7.6    Summary	7-13

Chapter 8: Assessment of Vehicle Safety Effects
  8.1    Safety Considerations in Establishing CAFE/GHG Standards	8-1
    8.1.1  Why Do the Agencies Consider Safety?	8-1
    8.1.2  How Do the Agencies Consider Safety?	8-3
  8.2    What is the Current State of the Research on Statistical Analysis of Historical Crash
  Data?  8-5
    8.2.1  Background	8-5
    8.2.2  Historical Activities Informing the 2017-2025 Final Rule	8-8
       8.2.2.1   2011 NHTSA Workshop on Vehicle Mass, Size and Safety	8-8
       8.2.2.2   Report by Green et. al., UMTRI - "Independent Review: Statistical Analyses of
       Relationship between Vehicle Curb Weight, Track Width, Wheelbase and Fatality
       Rates," April 2011	8-9
       8.2.2.3   2012NHTSA, LBNL, and DRI Reports	8-10
    8.2.3  Final Rule for Model Years 2017-2025	8-11
    8.2.4  Activities and Development since 2017-2025 Final Rule	8-11
       8.2.4.1   2013 Workshop on Vehicle Mass, Size and Safety	8-11
       8.2.4.2   Subsequent Analyses by LBNL	8-14
       8.2.4.3   2013 Presentations to NAS Subcommittee	8-15
       8.2.4.4   2015 National Academy of Sciences' Report	8-15
       8.2.4.5   2016 NHTSA/Volpe Study Reported in "Relationships between Fatality Risk,
       Mass, and Footprint in Model Year 2003-2010 Passenger Cars and LTVs: Preliminary
       Report," June 2016	8-16
       8.2.4.6   Report by Tom Wenzel, LBNL, "An Assessment of NHTSA's Report
       'Relationships between Fatality Risk, Mass, and Footprint in Model Year 2003-2010
       Passenger Cars and LTVs,'"2016	8-29
       8.2.4.7   Fleet Simulation Model	8-37
    8.2.5  Based on this Information, What do the Agencies Consider to be the Current State of
    Statistical Research on Vehicle Mass and Safety?	8-42
  8.3    How do the Agencies Think Technological Solutions Might Affect the Safety
  Estimates Indicated by the Statistical Analysis?	8-44
    8.3.1  Workshops on Technological Opportunities and Constraints to Improving Safety
    under Mass Reduction	8-45
       8.3.1.1   2011 Workshop on Vehicle Mass, Size and Safety	8-45
       8.3.1.2   2013 Workshop on Vehicle Mass, Size and Safety	8-47
    8.3.2  Technical Engineering Projects	8-49
       8.3.2.1   Honda Accord Study	8-49
       8.3.2.2   Second Honda Accord Study	8-50
       8.3.2.3   NHTSA Silverado Study and Light-Duty Fleet Analysis	8-51
       8.3.2.4   EPA Midsize CUV "Low Development" Study	8-53
       8.3.2.5   CARB Phase 2 Midsize CUV "High Development" Study	8-55
       8.3.2.6   EPA Light Duty Truck Study	8-56
  8.4    How have the Agencies Estimated Safety Effects for the Draft TAR?	8-57
    8.4.1  What was the Agencies' Methodology for Estimating Safety Effects?	8-57
                                          xn

-------
                                                                TOC and Abbreviations
     8.4.2  Why Might the Real-World Safety Effects be Less Than or Greater Than What the
     Agencies Have Calculated?	8-61
     8.4.3  What Are the Agencies' Plans Going Forward?	8-62

Chapter 9: Assessment of Alternative Fuel Infrastructure
  9.1    Overview	9-1
  9.2    Electric Vehicle Infrastructure	9-2
     9.2.1  Classification of Electric Vehicle Supply Equipment (EVSE)	9-3
       9.2.1.1  Level 1 EVSE	9-3
       9.2.1.2  Level 2 EVSE	9-4
       9.2.1.3  Direct Current (DC) Fast Charge	9-5
     9.2.2  Where People Charge	9-7
     9.2.3  Installation Costs  and Equipment Costs	9-11
       9.2.3.1  Installation Costs (Residential and Non-Residential)	9-12
       9.2.3.2  Installation Costs Trends	9-13
       9.2.3.3  EVSE Equipment Costs	9-13
       9.2.3.4  Equipment Costs Trends	9-15
     9.2.4  Status of National PEV Infrastructure	9-15
       9.2.4.1  Number of Connectors and Stations	9-15
       9.2.4.2  Trends, Growth	9-17
       9.2.4.3  Networks and Corridors	9-19
         9.2.4.3.1  West Coast Electric Highway (Baja California to British Columbia)	9-19
         9.2.4.3.2  Northeast Electric Vehicle Network (D.C. to Northern New England)... 9-20
         9.2.4.3.3  Tesla Super Charging Network (Coast to Coast)	9-20
         9.2.4.3.4  FAST Act - Nationwide Alternative Fuel Corridors	9-20
       9.2.4.4  Challenges and Opportunities with PEV Infrastructure	9-20
         9.2.4.4.1  Challenge - Multi-Unit Development (MuD)	9-20
         9.2.4.4.2  Challenge - Increasing Battery Capacity	9-21
         9.2.4.4.3  Challenge and Opportunity - Inductive Charging	9-21
         9.2.4.4.4  Opportunity - Vehicle Grid Integration (VGI)	9-22
         9.2.4.4.5  Opportunity - Utility Demand Response	9-22
       9.2.4.5  Further Analysis and Developments	9-22
       9.2.4.6  Status of Public PEV Infrastructure Network	9-23
       9.2.4.7  Summary of PEV Infrastructure	9-25
  9.3    Hydrogen Infrastructure Overview	9-25
     9.3.1  Hydrogen Network Development and Status	9-27
     9.3.2  Retail Experience	9-30
     9.3.3  Hydrogen Fueling Station Capacity	9-32
     9.3.4  Hydrogen Fueling Station Costs	9-32
     9.3.5  Paradigms for Developing Networks	9-35
     9.3.6  Challenges and Opportunities for Hydrogen Fueling Stations	9-39
  9.4    Fueling Infrastructure for Other Alternative Fuel Vehicles	9-41
  9.5    Summary of Alternative Fuel Infrastructure	9-41
                                          xni

-------
                                                              TOC and Abbreviations
Chapter 10: Economic and Other Key Inputs Used in the Agencies' Analyses
  10.1   The On-Road Fuel Economy "Gap"	10-1
     10.1.1   The "Gap" Between Compliance and Real World Fuel Economy	10-1
     10.1.2   Real World Fuel Economy and CCh Projections	10-2
  10.2   Fuel Prices and the Value of Fuel Savings	10-4
  10.3   Vehicle Mileage Accumulation and Survival Rates	10-6
  10.4   Fuel Economy Rebound Effect	10-9
     10.4.1   Accounting for the Fuel Economy Rebound Effect	10-9
     10.4.2   Summary of Historical Literature on the LD V Rebound Effect	10-10
     10.4.3   Review of Recent Literature on LDV Rebound since the 2012 Final Rule	10-15
     10.4.4   Basis for Rebound Effect Used in the Draft TAR	10-19
  10.5   Energy Security Impacts	10-21
     10.5.1   Implications of Reduced Petroleum Use on U.S. Imports	10-21
     10.5.2   Energy Security Implications	10-24
       10.5.2.1   Effect of Oil Use on the Long-Run Oil Price	10-25
       10.5.2.2   Macroeconomic Disruption Adjustment Costs	10-28
       10.5.2.3   Cost of Existing U.S. Energy Security Policies	10-33
       10.5.2.4   Military Security Cost Components of Energy Security	10-34
  10.6   Non-GHG Health and Environmental Impacts	10-35
     10.6.1   Economic Value of Reductions in Particulate Matter	10-36
  10.7   Greenhouse Gas Emission Impacts	10-41
  10.8   Benefits from Reduced Refueling Time	10-50
  10.9   Benefits and Costs from Additional Driving	10-53
     10.9.1   Travel Benefit	10-53
     10.9.2   Costs Associated with Crashes, Congestion and Noise	10-53
  10.10  Discounting Future Benefits and Costs	10-54
  10.11  Additional Costs of Vehicle Ownership	10-55
     10.11.1  Maintenance & Repair Costs	10-55
     10.11.2  Sales Taxes	10-55
     10.11.3  Insurance Costs	10-56

Chapter 11: Credits, Incentives and Flexibilities
  11.1   Overview	11-2
  11.2   Averaging, Banking, and Trading Provisions	11-3
  11.3   Air Conditioning System Credits	11-4
  11.4   Off-cycle Technology Credits	11-5
  11.5   Incentives for Advanced Technology Vehicles	11-6
  11.6   Advanced Technology Incentives for Large Pickups	11-8
  11.7   Harmonized CAFE Incentives and Flexibilities	11-9

Chapter 12: EPA's Analysis of the MY2022-2025 GHG Standards
  12.1   EPA's Estimates of Costs  per Vehicle & Technology Penetrations Based on OMEGA
         12-2
     12.1.1   Central Analysis Results	12-5
       12.1.1.1   CO2 Targets and Achieved Values	12-5
         12.1.1.1.1  Reference Case	12-5
                                         XIV

-------
                                                             TOC and Abbreviations
       12.1.1.1.2  Control Case	12-6
       12.1.1.1.3  Off-Cycle, Pickup Incentive and A/C Credits in OMEGA	12-10
       12.1.1.1.4  Projected 2-Cycle CO2	12-12
    12.1.1.2  Cost per Vehicle	12-14
       12.1.1.2.1  Reference & Control Case	12-14
    12.1.1.3  Technology Penetration	12-18
       12.1.1.3.1  Reference Case	12-18
       12.1.1.3.2  Control Case	12-24
    12.1.1.4  Comparisons to the 2012 Final Rule	12-34
  12.1.2   Sensitivity Analysis Results	12-36
    12.1.2.1  Reference Case: CO2 Targets	12-36
    12.1.2.2  Control Case: CO2 Targets	12-37
    12.1.2.3  Cost per Vehicle and Technology Penetrations	12-37
    12.1.2.4  Observations on Sensitivity Analyses	12-40
  12.1.3   Payback Period & Lifetime Savings	12-41
12.2   EPA's Projected Impacts on Emissions Inventories & Fuel Consumption	12-47
  12.2.1   Analytical Tools Used	12-47
  12.2.2   Inputs to the Emissions and Fuel Consumption Analysis	12-47
    12.2.2.1  Methods	12-47
    12.2.2.2  Global Warming Potentials	12-48
    12.2.2.3  Years Considered	12-49
    12.2.2.4  Fleet Activity	12-49
       12.2.2.4.1  Vehicle Sales, Survival Schedules, and VMT	12-49
    12.2.2.5  Upstream Emission Factors	12-49
       12.2.2.5.1  Gasoline Production and Transport Emission Rates	12-49
       12.2.2.5.2  Electricity Generation Emission Rates	12-50
    12.2.2.6  Reference Case CO2 g/mi & kWh/mi	12-51
    12.2.2.7  Control Case CO2 g/mi & kWh/mi	12-53
    12.2.2.8  Criteria Pollutant and Select Toxic Pollutant Emission Rates	12-55
  12.2.3   Outputs of the Emissions and Fuel Consumption Analysis	12-56
    12.2.3.1  Calendar Year Results	12-57
    12.2.3.2  Model Year Lifetime Results	12-61
  12.2.4   Sensitivity Analysis Results	12-62
    12.2.4.1  Calendar Year Case Comparison Results	12-63
    12.2.4.1  Model Year Lifetime Case Comparison Results	12-63
12.3   EPA's Benefit-Cost Analysis Results	12-64
  12.3.1   Model Year Analysis	12-64
    12.3.1.1  AEO 2015 Reference Fuel Price Case Using ICMs	12-64
    12.3.1.2  AEO 2015 Reference Fuel Price Case Using RPEs	12-67
    12.3.1.3  AEO 2015 High Fuel Price Case Using ICMs	12-69
    12.3.1.4  AEO 2015 Low Fuel Price Case Using ICMs	12-71
    12.3.1.5  Summary of MY Lifetime Benefit-Cost Analysis Results	12-74
  12.3.2   Calendar Year Analysis	12-74
    12.3.2.1  AEO 2015 Reference Fuel Price Case Using ICMs	12-75
    12.3.2.2  AEO 2015 Reference Fuel Price Case Using RPEs	12-76
    12.3.2.3  AEO 2015 High Fuel Price Case Using ICMs	12-76
                                        XV

-------
                                                              TOC and Abbreviations
       12.3.2.4  AEO 2015 Low Fuel Price Case Using ICMs	12-77
       12.3.2.5  Summary of CY Benefit-Cost Analysis Results	12-78
  12.4   Additional OMEGA Cost Analyses	12-79
     12.4.1   Cost per Vehicle Tables - Absolute and Incremental Costs	12-79
     12.4.2   Cost per Percentage Improvement in CCh	12-82

Chapter 13: Analysis of Augural CAFE Standards
  13.1   Significant Assumptions and Inputs to theNHTSA Analysis	13-2
     13.1.1   MY2015 Analysis Fleet	13-2
     13.1.2   Assumptions about Product Cadence	13-5
     13.1.3   Assumptions about Consumer Behavior	13-8
     13.1.4   Updated Mileage Accumulation Schedules for the Draft TAR	13-11
       13.1.4.1  Updated Schedules	13-11
       13.1.4.2  Data Description	13-16
       13.1.4.3  Estimation	13-19
       13.1.4.4  Comparison to previous schedules	13-20
       13.1.4.5  Future direction	13-21
     13.1.5   Other Assumptions of Note	13-21
  13.2   CAFE Model (aka "Volpe Model") Overview and Updates Since the 2012 Final Rule
         13-23
     13.2.1   Updates to 2012 Final Rule Version of the CAFE Model	13-24
       13.2.1.1  Integrating Vehicle Simulation Results into the CAFE Model	13-28
     13.2.2   Overview and Technology Application	13-37
     13.2.3   Simulating Manufacturer Compliance with Standards	13-48
     13.2.4   Simulating the Economic and Environmental Effects of CAFE Standards	13-55
  13.3   Simulation Results for Augural MY2022 - 2025 Standards	13-56
     13.3.1   Industry Impacts	13-57
     13.3.2   Consumer Impacts	13-93
     13.3.3   Social and Environmental Impacts	13-99
     13.3.4   Overall Benefits and Costs	13-102

Appendix A               CARB Analysis of Vehicle Load Reduction Potential for
Advanced Clean Cars      	A-l

Appendix B               Mass Reduction Technologies
  B.I    Design Optimization	B-4
  B.2    Material Advancements - Steel	B-6
  B.3    Material Advancements - Aluminum	B-24
  B.4    Material Advancements - Magnesium	B-40
  B.5    Material Advancements - Plastics	B-46
  B.6    Material Advancements - Composites	B-52
  B.7    Material Advancements - Glass	B-76
  B.8    Multi-Material Technology Examples	B-77
  B.9    Additional Vehicle Level Cost Analysis	B-81
  B.10   Mass Reduction Technology Adoption Trends in the Marketplace	B-84
                                         XVI

-------
                                                              TOC and Abbreviations
Appendix C               EPA's OMEGA Model
  C.I    OMEGA Pre-Processors, Vehicle Types & Packages	C-l
    C.I.I  Vehicle Types	C-2
    C.I.2  Technology Packages, Package Building & Master-sets	C-3
    C.I.3  Master-set Ranking & the Technology Input File	C-ll
    C.I.4  Applying Ranked-sets of Packages to the Projected Fleet	C-14
    C.I.5  New to OMEGA since the 2012 FRM	C-15
  C.2    OMEGA Overview	C-15
  C.3    OMEGA Model Structure	C-17
                                         XVll

-------
                                    TOC and Abbreviations
This Page Intentionally Left Blank
             XVlll

-------
                                                             TOC and Abbreviations
List of Acronyms
      2MHEV
      ABS
      ABT
      AC
      A/C
      ACEEE
      AEO
      AER
      AFDC
      AGM
      AHSS
      ALPHA
      AMI
      ANL
      ARE
      AST
      ASL
      ASM
      AT
      Avg
      AWD
      BenMAP
      BEV
      BISG
      BIW
      BLS
      BMEP
      BOM
      BSFC
      BTE
      BTU
      CAA
      CAD
      CAD/CAE
      CAE
      CAFE
      CARS
      CBD
      CBI
      CCP
      CDPF
      CEC
2-Mode Hybrid
Anti-lock Braking System
Averaging, Banking, and Trading
Alternating Current
Air Conditioning
American Council for an Energy-Efficient Economy
Annual Energy Outlook
All-Electric Range
Alternative Fuels Data Center
Absorbent Glass Mat
Advanced High Strength Steel
Advanced Light-Duty Powertrain and Hybrid Analysis Tool
Automated Manual Transmission
Argonne National Laboratory
California Air Resources Board
Area Specific Impedance
Aggressive Shift Logic
Annual Survey of Manufacturers
Automatic Transmissions
Average
All Wheel Drive
Benefits Mapping and Analysis Program
Battery Electric Vehicle
Belt Integrated Starter Generator
Body-In-White
Bureau of Labor Statistics
Brake Mean Effective Pressure
Bill of Materials
Brake Specific Fuel Consumption
Brake-Thermal Efficiency
British Thermal Unit
Clean Air Act
Computer Aided Designs
Computer Aided Design And Engineering
Computer Aided Engineering
Corporate Average Fuel Economy
California Air Resources Board
Center for Biological Diversity
Confidential Business Information
Coupled Cam Phasing
Catalyzed Diesel Particulate Filter
California Energy Commission
                                        XIX

-------
                                                      TOC and Abbreviations
CES          Consumer Expenditure Survey
CFD          Computational Fluid Dynamics
CFR          Code of Federal Regulations
CFLt          Methane
CISG         Crank Integrated Starter Generator
CNG         Compressed Natural Gas
CO           Carbon Monoxide
CCh          Carbon Dioxide
CCheq        CO2 Equivalent
COP          Coefficient of Performance
CSM         Conceptual Site Model
CSV          Comma-separated Values
CUV         Crossover Utility Vehicles
CVT          Continuously Variable Transmission
CY           Calendar Year
DC           Direct Current
DCFC        Direct Carbon Fuel Cell
DCP          Dual Cam Phasing
DCT          Dual Clutch Transmission
DEAC        Cylinder Deactivation
DFMA™      Design for Manufacturing and Assembly
DGS          California Department of General Services
DICE         Dynamic Integrated Climate and Economy
DMC         Direct Manufacturing Costs
DoE          Department of Energy
DOE         Design of Experiments
DOHC        Dual Overhead Camshaft Engines
DOT         Department of Transportation
DRI          Dynamic Research, Inc.
DRLs         Daytime Running Lamps
DVVL        Discrete Variable Valve Lift
EGR         Exhaust Gas Recirculation
EHPS         Electrohydraulic Power Steering
              Energy Information Administration (part of the U.S. Department of
              Energy)
EISA         Energy Independence and Security Act
EIVC         Early Intake Valve Closing
EPA          Environmental Protection Agency
EPCA        Energy Policy and Conservation Act
EPRI         Electric Power Research  Institute
EPS          Electric Power Steering
EPS          Energy Power Systems
EREV        Extended Range Electric Vehicle
                                  xx

-------
                                                      TOC and Abbreviations
ERM         Employment Requirements Matrix
ESC          Electronic Stability Control
EV           Electric Vehicle
EVSE         Electric Vehicle Supply Equipment
PARS         Fatality Analysis Reporting System
FCEV         Fuel Cell Electric Vehicle
FCPM         Fuel Cost Per Mile
FCEV         Fuel Cell Electric Vehicle
FE           Finite Element
FEV1         Functional Expiratory Volume
FHWA        Federal Highway Administration
FMEP         Friction Mean Effective Pressure
FMVSS       Federal Motor Vehicle Safety Standards
FR           Federal Register
FRIA         Final Regulatory Impact Analysis
FRM         Federal Rulemaking
FRM         Federal Reference Method
FTP          Federal Test Procedure
gal/mi         Gallon/Mile
GCWR        Gross Combined Weight Rating
GDI          Gasoline Direct Injection
GDP          Gross Domestic Product
GEM         Greenhouse gas Emissions Model
GHG         Greenhouse Gases
              Greenhouse Gases, Regulated Emissions, and Energy Use in
GREEI        m       , , •
              Transportation
GVW         Gross Vehicle Weight
GWP         Global Warming Potential
GWU         George Washington University
HD           Heavy-Duty
HEV         Hybrid Electric Vehicle
HFC          Hydrofluorocarbon
HFET         Highway Fuel Economy Dynamometer Procedure
HIL          Hardware-In-Loop
hp            Horsepower
hrs           Hours
HP/WT        Horsepower Divided by Weight
HVAC        Heating, Ventilating, And Air Conditioning
hz            Hertz
IACC         Improved Accessories
IAM          Integrated Assessment Models
IATC         Improved Automatic Transmission Control
1C            Indirect Cost
                                  XXI

-------
                                                       TOC and Abbreviations
ICCT         International Council on Clean Transport
ICF           ICF International
ICM          Indirect Cost Multiplier
ICMs         Indirect Cost Markups
IHX          Internal Heat Exchanger
IMA          Improved Mobile Assist
EVIAC         Improved Mobile Air Conditioning
INL           Idaho National Laboratory
IOU          Investor Owned Utilities
IPCC         Intergovernmental Panel on Climate Change
IPM          Integrated Planning Model
ITC           Institute of Transportation Studies
IWG          Interagency Working Group
k             Thousand
kg            Kilogram
kW           Kilowatt
kWh          kilowatt-hour
L             Liter
Ib             Pound
LBNL         Lawrence Berkeley National Laboratory
LD           Light-Duty
LEV          Low-Emission Vehicle
LHD          Light Heavy-Duty
LDV          Light Duty Vehicle
LNT          Lean NOx Trap
LRR          Lower Rolling Resistance
LT           Light Trucks
LWT         Lightweighted Pickup Truck
MAD         Minimum Absolute Deviation
MBPD        Million Barrels Per Day
MD           Medium-Duty
MDPV        Medium-Duty Passenger Vehicles
MEMA       Motor Equipment Manufacturers Association
Mg           Megagrams
mg           Milligram
MHEV        Mild Hybrid
mi            mile
min           minimum
min           Minute
MM          Million
MMLV       Multi-Material Lightweight Vehicle
MMT         Million Metric Tons
MOVES       Motor Vehicle Emissions Simulator
                                  xxn

-------
                                                      TOC and Abbreviations
mpg          Miles per Gallon
mph          Miles per Hour
MPV         Multi-Purpose Vehicle
MSRP        Manufacturer's Suggested Retail Price
MTE         Mid Term Evaluation
MuD         Multi-Unit Development
MY          Model Year
N2O          Nitrous Oxide
NA           Not Applicable
NAAQS       National Ambient Air Quality Standards
NADA        National Automobile Dealers Association
NAS          National Academy of Sciences
NCA         National Climate Assessment
NCAP        New Car Assessment Program
NEMS        National Energy Modeling System
NESHAP      National Emissions Standards for Hazardous Air Pollutants
NFs          Nitrogen Trifluoride
NGO         Non-Governmental Organization
NHTSA       National Highway  Traffic Safety Administration
NiMH        Nickel Metal-Hydride
NFs          Nitrogen Trifluoride
NOX         Nitrogen Oxides
NO2          Nitrogen Dioxide
NOx          Oxides of Nitrogen
NPRM        Notice of Proposed Rulemaking
NRC          National Research  Council
NRC-CAN     National Research  Council of Canada
NREL        National Renewable Energy Laboratory
NVH         Noise Vibration and Harshness
NVPP        National Vehicle Population Profiles
OAR         EPA's Office of Air and Radiation
OEM         Original Equipment Manufacturer
OECD        Organization for Economic Cooperation and Development
OHV         Overhead Valve
OLS          Ordinary Least Squares
OMB         EPA's Office of Management and Budget
OPEC        Organization of Petroleum Exporting Countries
ORNL        Oak Ridge National Laboratory
OTAQ        EPA's Office of Transportation and Air Quality
PAGE        Policy Analysis of the Greenhouse Effect
PC            Passenger Car
P/E           Power-to-Energy
PEF          Peak Expiratory Flow
                                 XXlll

-------
                                                       TOC and Abbreviations
PEV          Plug-in Electric Vehicle
PFCs         Perfluorocarbons
PFI           Port-fuel-injection
PGM         Platinum Group Metal
PHEV        Plug-in Hybrid Electric Vehicle
PLM         Planar Layered Matrix
PM           Particulate Matter
PM2.5         Fine Particulate Matter (diameter of 2.5 jim or less)
PMSMs       Permanent-Magnet Synchronous Motors
PSHEV       Power-split Hybrid
PSI           Pounds per Square Inch
PWM         Pulse-width Modulated
R&D         Research and Development
RFS2         Renewable Fuel Standard 2
RIA          Regulatory Impact Analysis
RPE          Retail Price Equivalent
RPM         Revolutions per Minute
RSM         Response Surface Models
RTI          RTI International (formerly Research Triangle Institute)
SA           Strategic Analysis, Inc.
SAB          Science Advisory Board
SAB-         Science Advisory Board Environmental Economics Advisory
EEAC        Committee
SAE          Society of Automotive Engineers
SCCb         Soak Control third iteration
SCC          Social  Cost of Carbon
SCR          Selective Catalyst Reduction
SFe           Sulfur  Hexafluoride
SGDI         Stoichiometric Gasoline Direct Injection
SHEV        Strong Hybrid Electric Vehicles
SI            Spark-Ignition
SIDI          Spark Ignition Direct Inj ection
SIL           Software-In-Loop
SMDI        Steel Market Development Institutes
SNAP        Significant New Alternatives Policy
SNPRM       Supplemental Notice of Proposed Rulemaking
SO2          Sulfur  Dioxide
SOx          Sulfur  Oxides
SOC          State of Charge
SOHC        Single Overhead Cam
SOL          Small Overlap
SPR          Strategic Petroleum Reserve
                                  XXIV

-------
                                                       TOC and Abbreviations
Std           Standard
SUV          Sport Utility Vehicle
TAR          Technical Assessment Report
TC           Total Costs
TCIP          Tire Consumer Information Program
TDC          Top Dead Center
Tds           Direct Solar Transmittance
TFECIP       Tire Fuel Efficiency Consumer Information Program
TPE          Total Primary Energy
TRBDS       Turbocharging and Downsizing
TSD          Technical Support Document
UMTRI       University of Michigan Transportation Research Institute
UTQGS       Uniform Tire Quality Grading Standards
V2V          Vehicle-To-Vehicle
VGI          Vehicle Grid Integration
VIF           Variance Inflation Factor
VMT          Vehicle Miles Traveled
VOC          Volatile Organic Compound
VSL          Vehicle Speed Limiter
VVL          Variable Valve Lift
VVT          Variable Valve Timing
WT/FP        Weight Divided By Footprint
                                  XXV

-------

-------
                                                                    Executive Summary
Executive Summary

   The Environmental Protection Agency (EPA) and the Department of Transportation's
National Highway Traffic Safety Administration (NHTSA) have established a coordinated
program for Federal standards for greenhouse gas (GHG) emissions and corporate average fuel
economy (CAFE) for light-duty vehicles.1  This program was developed in cooperation and
alignment with the California Air Resources Board (CARB) to ensure a single National Program.
The National Program established standards that increase in stringency year-over-year from
model year (MY) 2012 through MY2025 for EPA and through MY2021 for NHTSA.  California
adopted the first in the nation GHG standards for light-duty vehicles in 2004 for MY2009-2016,
and in 2012 for MY2017-2025, followed by amendments that allow compliance with the Federal
GHG standards as compliance with the California GHG standards, in furtherance of a single
National Program. Under the National Program, consumers continue to have a full range of
vehicle choices that meet their needs, and, through coordination with the California standards,
automakers can build a single fleet of vehicles across the U.S. that satisfies all GHG/CAFE
requirements. In the agencies' 2012 final rules establishing the MY2017-2025 standards for
EPA and 2017-2021 final  and 2022-2025 augural standards for NHTSA, the National Program
standards were projected by MY2025 to double fuel economy and cut GHG emissions in half,
save 6 billion metric tons of carbon dioxide (CCh) pollution and 12 billion barrels of oil over the
lifetime of MY2012-2025 vehicles, and deliver significant savings for consumers at the gas
pump.

   The rulemaking establishing the National Program for MY 2017-2025 light-duty vehicles
included a regulatory requirement for EPA to conduct a Midterm Evaluation (MTE) of the GHG
standards established for MYs 2022-2025.' The 2012 final rule preamble also states that "[t]he
mid-term evaluation reflects the rules' long time frame, and, for NHTSA, the agency's statutory
obligation to conduct a de novo rulemaking in order to establish final standards for MYs 2022-
2025." NHTSA will consider information gathered as part of the MTE record, including
information submitted through public comments, in the comprehensive de novo rulemaking it
must undertake to set CAFE standards for MYs 2022-2025." Through the MTE, EPA must
determine no later than April 1, 2018 whether the MY2022-2025 GHG standards, established in
2012, are still appropriate  under section 202 (a) of the Clean Air Act, in light of the record then
before the Administrator, given the latest available data and information.111 EPA's decision could
go one of three ways: the standards remain appropriate, the standards should be less stringent,  or
the standards should be more stringent. EPA and NHTSA also are closely coordinating with
CARB in conducting the MTE to better ensure  the continuation of the National Program.  The
MTE will be a collaborative, data-driven, and transparent process and must entail a holistic
assessment of all the factors considered in the initial standards  setting.1V

   This Draft Technical Assessment Report (TAR), issued jointly by EPA, NHTSA, and CARB
for public comment, is the first formal step in the MTE process/ In this Draft TAR, the agencies
examine a wide range of technical issues relevant to GHG emissions and augural CAFE
standards for MY2022-2025, and share with the public the initial technical analyses of those
issues.  This is a technical report, not a policy or decision document. The information in this
1 The agencies finalized the first set of National Program standards covering model years (MYs) 2012-2016 in May
  20101 and the second set of standards, covering MYs 2017-2025, in October 2012.
                                         ES-1

-------
                                                                    Executive Summary
report, and in the comments we receive on it, will inform the agencies' subsequent determination
and rulemaking actions. The agencies will fully consider public comments on this Draft TAR as
they continue to update and refine the analyses for further steps in the MTE process.

   In this Draft TAR, EPA provides its initial technical assessment of the technologies available
to meet the MY2022-2025 GHG standards and one reasonable compliance pathway, and
NHTSA provides its initial assessment of technologies available to meet the augural MY2022-
2025 CAFE standards and a different reasonable compliance pathway. Given that there are
multiple possible ways that new technologies can be added to the fleet, examining two
compliance pathways provides valuable additional information about how compliance may
occur.  NHTSA and EPA also performed multiple sensitivity analyses which show additional
possible compliance pathways.  The agencies' independent analyses complement one another
and reach  similar conclusions:

      A wider range of technologies exist for manufacturers to use to meet the MY2022-2025
      standards, and at costs that are similar or lower, than those projected in the  2012 rule;
      Advanced gasoline vehicle technologies will continue to be the predominant technologies,
      with modest levels of strong hybridization and very low levels of full electrification (plug-
      in vehicles) needed to meet the standards;
   -   The car/truck mix reflects updated  consumer trends that are informed by a range of factors
      including economic growth, gasoline prices, and other macro-economic trends.  However,
      as the standards were designed to yield improvements across the light duty  vehicle fleet,
      irrespective of consumer choice, updated trends are fully accommodated by the footprint-
      based standards.
   Additionally, while the Draft TAR analysis focuses on the MY2022-2025 standards, the
agencies note that the auto industry, on average, is over-complying with the first several years of
the National Program.  This has occurred  concurrently with a period during which the
automotive industry successfully rebounded after a period of economic distress. The industry
has now seen six consecutive years of increases and a new all-time sales record in 2015,
reflecting positive consumer response to vehicles complying with the standards.

   A summary of each chapter of the Draft TAR follows.

   Chapter 1:  Introduction. This chapter provides a broad discussion of the National
Program, explains further the MTE process and timeline, and provides additional background on
NHTSA's CAFE program, EPA's GHG program, and California's GHG program. This chapter
also includes an update on what the latest science tells us about climate change impacts, and the
U.S.'s and California's commitments on actions to address climate change. Chapter 1  also
provides a discussion of petroleum consumption and energy security.

   Chapter 2:  Overview of Agencies' Approach to Draft TAR Analysis. The  agencies are
committed to conducting the MTE through a collaborative, data-driven, and transparent process.
In gathering data and information for this Draft TAR,  the agencies drew from a wide range of
sources to evaluate how the automotive industry has responded into the early years of the
National Program, how technology has developed, and how other factors affecting the light-duty
vehicle fleet have changed since the final  rule in 2012. The agencies found that there is a wealth
of information since the 2012 final rule upon which to inform this Draft TAR,  and this
                                          ES-2

-------
                                                                    Executive Summary
information is detailed throughout the document. Chapter 2 describes these sources, including
extensive state-of-the-art research projects by experts at the EPA National Vehicle and Fuel
Emissions Laboratory, as well as consultants to the agencies, data and input from stakeholders,
and information from technical conferences, published literature, and studies published by
various organizations. A significant study informing the agencies' analyses is the National
Academy of Sciences 2015 report" on fuel economy technologies, which the agencies highlight
in Chapter 2, and discuss throughout this document.

   The analyses presented in this Draft TAR reflect the new data and information gathered by the
agencies thus far, and the agencies will continue to gather and evaluate more up-to-date
information, including public comments on this Draft TAR, to inform our future analyses. The
agencies have conducted extensive outreach with a wide range of stakeholders - including auto
manufacturers, automotive suppliers, non-governmental organizations (NGOs), consumer
groups, labor unions, automobile dealers, state and local governments, and others.

   Chapter 3: Recent Trends in the Light-Duty Vehicle Fleet since the 2012 Final Rule.
This chapter summarizes trends in the light-duty vehicle market in the four years since the 2012
final rule, including changes in fuel economy/GHG emissions, vehicle sales, gasoline prices,
car/truck mix, technology penetrations, and vehicle power, weight and footprint. Since the 2012
final rule, vehicle sales have been strong, hitting an all-time high of 17.5 million vehicles in
2015, gas prices have dropped significantly, and truck share has grown. At the same time, fuel
economy technologies are entering the market at rapid  rates. The agencies provide the latest
available projections for vehicle sales, gasoline prices,  and fleet mix out to 2025, and compare
those to projections made in the 2012 final rule.  This chapter also highlights compliance to date
with the GHG and CAFE standards, where, for the first three years of the program (MY2012-
2014), auto manufacturers have over-complied with the program.

   Chapter 4: Baseline and Reference Vehicle Fleets. This chapter describes the agencies'
methodologies for developing a baseline fleet of vehicles and future fleet projections out to
MY2025.  The GHG analysis uses a baseline fleet based on the MY2014 fleet, the latest year
available for which there are final GHG  compliance data. The CAFE analysis uses a MY2015
baseline fleet based on MY2015 data and sales projections provided by manufacturers in the
latter half of MY2015, when production  was well underway. These data sets complement one
another and each yield important perspective, with the MY2014 data  having the benefit of
validation through compliance data, and the MY2015 data providing  more recent perspective.
The GHG and CAFE analysis fleets utilized similar, but separate, purchased projections  from
IHS-Polk for the future vehicle fleet mix out to 2025, thereby representing some of the
uncertainty inherent in all reference case projections. Both analyses used data from the Energy
Information Administration's Annual Energy Outlook 2015 (AEO 2015) as the basis for total
vehicle sales projections to 2025,  as well as for the car and truck volume mix.  Although the
agencies have relied on different data sources in development of the baseline fleets, we believe
this combination of approaches strengthens our results  by showing robust results across a range
of reference case projections.

   Chapter 5: Technology Costs, Effectiveness, and Lead-Time Assessment.  This chapter is
an in-depth assessment of the state of vehicle technologies to improve fuel economy and reduce
GHG emissions, as well as the agencies' assessment of expected future technology developments
through MY2025.  The technologies evaluated include all those considered for the 2012  final
                                          ES-3

-------
                                                                    Executive Summary
rule, as well as new technologies that have emerged since then. Every technology has been
reconsidered with respect to its cost, effectiveness, application, and lead-time considerations,
with emphasis on assessing the latest introductions of technologies to determine if and how they
have changed since the agencies' assessment in the 2012 final rule. These efforts reflect the
significant rate of progress made in automotive technologies over the past four years since the
MY2017-2025 standards were established.  Technologies considered in this Draft TAR include
more efficient engines and transmissions, aerodynamics, light-weighting, improved accessories,
low rolling resistance tires, improved air conditioning systems, and others. Beyond the
technologies the agencies considered in the 2012 final rule, manufacturers are now employing
several technologies, such as higher compression ratio, naturally aspirated gasoline engines, and
greater penetration of continuously variable transmissions (CVTs); other new technologies are
under active development and are expected to be in the fleet well before MY2025, such  as 48-
volt mild hybrid systems.

   In Chapter 5, the agencies also provide details on the specific technology assumptions used
respectively by EPA for the GHG assessment and by NHTSA for the CAFE assessment in this
Draft TAR, including the specific assumptions that EPA and NHTSA each made for each
technology's cost and effectiveness, and lead-time considerations. The agencies' estimates of
technology effectiveness were informed by vehicle simulation modeling approaches; NHTSA
utilized the Autonomie model developed by Argonne National Laboratories for the Department
of Energy (DOE), and EPA used its Advanced Light-duty Powertrain and Hybrid Analysis
(ALPHA) model.  The agencies look forward to public comment in this and other areas to help
advance collective forecasting of technology effectiveness in  the out years of the program.

   It is clear that the automotive industry is innovating and bringing new technology to market at
a rapid pace and neither of the respective agency analyses reflects all of the latest and emerging
technologies that may be available in the 2022-2025 time frame. For example, the agencies were
not able for this Draft TAR to evaluate the potential for technologies such as electric turbo-
charging, variable compression ratio, skip-fire cylinder deactivation, and P2-configuration mild-
hybridization. These technologies may provide further cost-effective reductions in GHG
emissions and fuel consumption. The agencies will continue  to update their analyses throughout
the MTE process as new information becomes available.

   Chapter 6: Assessment of Consumer Acceptance of Technologies that Reduce Fuel
Consumption and GHG Emissions. This chapter reviews issues surrounding consumer
acceptance of the vehicle technologies expected to be used to meet the MY2022-2025 standards.
Since the program has been in effect since MY2012, the  agencies focus on the evidence to date
related to consumer acceptance of vehicles subject to the National Program standards. This
evidence includes an analysis of how professional auto reviewers assess fuel-saving
technologies. For each technology, positive evaluations exceed negative evaluations, suggesting
that it is possible to implement these technologies without significant hidden costs.  To date,
consumer response to vehicles subject to the standards is positive. Chapter 6 also discusses
potential impacts of the standards on vehicle sales and affordability, which are closely
interconnected with the effects of macroeconomic and other market forces. Based on the
agencies' draft assessments, the reduced operating costs from fuel savings over time are expected
to far exceed the increase in up-front vehicle costs, which should mitigate any potential adverse
effects on vehicle sales and affordability.
                                          ES-4

-------
                                                                     Executive Summary
   Chapter 7: Employment Impacts. This chapter discusses the effects of employment in the
automotive sector to date, and the projected effects of the MY 2022-2025 standards on
employment.  Employment in the automotive industry dropped sharply during the Great
Recession, but has increased steadily since 2009.  The primary employment effects of these
standards are  expected to be found in several key sectors: auto manufacturers, auto parts
manufacturing, auto dealers,  fuel production and supply, and consumers. The MY2025
standards are  likely to have some effect on employment, due to both the effects of the standards
on vehicle sales, and the need to produce new technologies to meet the standards.  Nevertheless,
the net effect  of the standards on employment is likely to be small compared to macroeconomic
and other factors affecting employment.

   Chapter 8: Assessment of Vehicle Safety Effects. This chapter assesses the  estimated
overall crash  safety impacts of the MY 2022-2025 standards.  In this chapter, the agencies first
review the relationships between mass, size, and fatality risk based on the statistical analysis of
historical crash data, which includes the new analysis performed by using the most recent crash
data.  The updated NHTSA analysis develops five parameters for use in both the NHTSA and
EPA assessments to calculate the estimated safety impacts of the modeled mass reductions over
the lifetimes of new vehicles in response to MY 2022-2025 GHG standards and augural CAFE
standards. Second, to examine the impact of future lightweight vehicle designs on safety, the
agencies also  reviewed a fleet crash simulation study that examined frontal crashes using
existing and future lightweight passenger car and cross-over utility vehicle designs. The study
found a relationship between vehicle mass reduction and safety that is directionally consistent
with the overall risk for passenger cars from the NHTSA 2016 statistical analysis  of historical
crash data.  Next, the agencies investigate the amount of mass reduction that is affordable and
feasible while maintaining overall fleet safety and as well as functionality such as durability,
drivability, noise, vibration and handling (NVH), and acceleration performance. Based on those
approaches, the agencies further discuss why the real world safety effects might be less than or
greater than calculated safety impacts, and what new challenges these lighter vehicles might
bring to vehicle safety and potential countermeasures available to manage those challenges
effectively.

   Chapter 9: Assessment of Alternative Fuel Infrastructure. This chapter assesses the
status of infrastructure for alternative fueled vehicles, with emphasis on two technologies the
agencies believe will be important for achieving longer-term climate and energy goals - plug-in
electric vehicles (PEVs) and  fuel cell electric vehicles (FCEVs). The agencies also discuss
infrastructure for ethanol (E85) flex-fueled vehicles and natural gas vehicles.  The agencies'
assessment is that, as we concluded in the 2012 rule, high penetration levels of alternative fueled
vehicles will not be needed to meet the MY2025 standards, with the exception of a very small
percentage of PEVs, and that infrastructure is progressing sufficiently to support vehicles from
those manufacturers choosing to produce alternative fueled vehicles to meet the MY2022-2025
standards. The majority of PEV charging occurs at home, and national PEV infrastructure in
public and work locations is progressing appropriately. Hydrogen infrastructure developments
are addressing many of the initial challenges of simultaneously launching new vehicle and
fueling infrastructure markets, and current efforts in California and the northeast states will
facilitate further vehicle and infrastructure rollout at the national level.

   Chapter 10:  Economic and Other Key Inputs Used in  the Agencies' Analyses. This
chapter describes many of the economic and other inputs used in the agencies' analyses. This
                                          ES-5

-------
                                                                    Executive Summary
chapter discusses the methodologies used to assess inputs such as the real-world fuel
economy/GHG emissions gap, vehicle miles traveled (VMT), vehicle survival rates, the VMT
rebound effect, energy security, the social cost of carbon and other GHGs, health benefits,
consumer cost of vehicle ownership, and others.

   Chapter 11: Credits, Incentives and Flexibilities.  The National Program was designed
with a wide range of optional compliance flexibilities to allow manufacturers to maintain
consumer choice,  spur technology development, and reduce compliance costs, while achieving
significant GHG and oil reductions. Chapter 11 provides an informational overview of all of
these compliance flexibilities, with particular emphasis on those flexibility options likely to be
most important in the MY2022-2025 timeframe.

   Chapter 12: Analysis of the MY2022-2025 GHG Standards;  and Chapter 13: Analysis of
Augural CAFE Standards.  Chapters 12 and 13 provide results, respectively, of EPA's initial
technical assessment of the technologies available to meet the MY2022-2025 GHG standards
(i.e., the footprint-based standard curves) and their costs, and NHTSA's initial technical
assessment of technologies capable of meeting CAFE standards corresponding to the augural
standards for MY2022-2025, and these technologies' costs. CARB has not conducted an
independent analysis, but has participated in both EPA's and NHTSA's analyses.  Although all
three agencies have been working collaboratively in an array of areas throughout the
development of this  Draft TAR, the EPA GHG and NHTSA CAFE assessments were done
largely independently. These independent analyses were done in part to recognize differences in
the agencies'  statutory authorities and to reflect independent choices regarding some of the
modeling inputs used at this initial stage of our evaluation. The agencies believe that
independent and parallel analyses can provide complementary results.  The agencies further
believe that, for this  Draft TAR which is the first step of the Midterm Evaluation process, it is
both reasonable and  advantageous to make use of different data sources and modeling tools, and
to show multiple pathways for potential compliance with the MY 2022-2025 GHG standards and
augural CAFE standards.

   As noted above, although CARB did not perform its own modeling assessment of the costs
and technologies to meet the 2022-2025 GHG and  CAFE requirements, it was integrally
involved in analyzing the underlying technology cost and effectiveness inputs to the EPA and
NHTSA modeling.  CARB believes that the analyses presented in this Draft TAR appropriately
present a range of technologies that could be used to meet the requirements. However, as
discussed above, there are, and will continue to be, emerging technologies that may well be
available in the 2022-2025 timeframe and could perform appreciably better or be lower cost than
the technologies modeled in this Draft TAR.  Such technologies are exemplified by recent
advancements already seen in the marketplace yet not anticipated by the agencies' rule four years
ago (e.g., expanded use of higher compression ratio, naturally aspirated gasoline engines).
Vehicle manufacturers have historically outpaced agency expectations and CARB believes it is
likely that industry will continue to do so.

   In this Draft TAR, NHTSA does not present alternatives to the augural standards because, as
the first stage of the  Midterm Evaluation process, the TAR is principally an exploration of
technical issues — including assumptions about the effectiveness and cost of specific
technologies,  as well as other inputs, methodologies and approaches for accounting for these
issues. The agencies seek comment from stakeholders to further inform the analyses, in advance
                                         ES-6

-------
                                                                     Executive Summary
of the NHTSA rulemaking and the EPA Proposed Determination.  For the purposes of clearly
reflecting the impacts of updated technology assumptions relative to a familiar point of
comparison, both agencies have run their respective models using the stringency levels included
in NHTSA's augural standards, and EPA's existing GHG standards through MY2025. However,
the technology assumptions and other analyses presented in this Draft TAR, which will be
informed by public comment, will support the development of a full range of stringency
alternatives in the subsequent CAFE Notice of Proposed Rulemaking.

  In this Draft TAR, the EPA GHG and NHTSA CAFE assessments both show that the
MY2022-2025 standards can be achieved  largely through the use of advanced gasoline vehicle
technologies with modest penetrations of lower cost electrification (like 48 volt mild hybrids
which include stop/start) and low penetrations of higher cost electrification (like strong hybrids,
plug-in hybrid electric vehicles, and all electric vehicles). Given the rapid pace of automotive
industry innovation, the agencies may consider effectiveness and cost of additional technologies
as new information, including comments on this Draft TAR, becomes available for further steps
of the Midterm Evaluation.

  Based on various assumptions including the Annual Energy Outlook 2015 (AEO 2015)
reference case projections of the car/truck mix out to 2025, the footprint-based GHG standards
curves for MY2022-2025 are projected to achieve an industry-wide fleet average CCh target of
175 grams/mile (g/mi) in MY2025, and the augural CAFE standards are projected to result in
average CAFE requirements increasing from 38.3 mpg in MY2021 to 46.3 mpg in MY2025.
The projected fleet average CCh target represents a GHG emissions level equivalent to 50.8 mpg
(if all reductions were achieved exclusively through fuel economy improvements).2

  Table ES-1 below compares two additional AEO 2015 scenarios in addition to the AEO 2015
reference case:  a low fuel  price case and a high fuel price case. As shown, these fuel price cases
translate into different projections for the car/truck fleet mix (e.g., with a higher truck share
shown in the low fuel price case, and a lower truck share shown in the high fuel price case),
which in turn leads to varying projections for the estimated fleet wide CAFE requirements and
GHG CO2 targets and MPG-e levels projected for MY2025, from 169 g/mi (52.6 mpg-e) under
the high fuel price case to 178 g/mi (49.9 mpg-e) under the low fuel price case.  These estimated
GHG target  levels and CAFE requirements reflect changes in the latest projections about the
MY2025 fleet mix compared to the projections in 2012 when the agencies first established the
standards. Under the footprint-based standards, the program is designed to ensure significant
GHG reductions/fuel economy improvements across the fleet, and each automaker's standard
automatically adjusts based on the mix (size and volume) of vehicles it produces each  model
year.  In the  agencies' current analyses for this Draft TAR, we are  applying the same footprint-
based standards established in the 2012 final rule to the updated fleet projections for MY2025.  It
is important  to keep in mind that the updated MY2025 fleet wide projections reflected in this
Draft TAR are still just projections (as were the fleet projections in the 2012 rule) — based on the
latest available information, which may continue to change with future projections — and that the
actual GHG  emissions/fuel economy level achieved in MY2025 won't be determined until the
2 The projected MY 2025 target of 175 g/mi represents an approximate 50% decrease in GHG emissions relative to
  the fuel economy standards that were in place in 2010. It is clear from current GHG manufacturer performance
  data that many automakers are earning air conditioner refrigerant GHG credits that reduce GHG emissions, but do
  not increase fuel economy. Accordingly, the projected MY 2025 target of 175 g/mi represents slightly less than a
  doubling of fuel economy relative to the standards that were in place in 2010.


                                          ES-7

-------
                                                                        Executive Summary
manufacturers have completed their MY2025 production.  The agencies will continue to assess
the latest available projections as we continue the Midterm Evaluation process.
Table ES-1 Projections for MY2025:  Car/Truck Mix, CO2 Target Levels, and MPG-equivalent1


Car/truck mix
CAFE (mpg)2
CO2 (g/mi)
MPG-e

2012 Final Rule
67/33%
48.7
163
54.5
AEO 2015 Fuel Price Case
AEO Low
48/52%
45.7
178
50.0
AEO Reference
52/48%
46.3
175
50.8
AEO High
62/38%
47.7
169
52.6
Notes:
1 The CAFE, COa and MPG-e values shown here are 2-cycle compliance values. Projected real-world values are
detailed in Chapter 10.1; for example, for the AEO reference fuel price case, real-world EPA CCh emissions
performance would be 220 g/mi and real-world fuel economy would be 36 mpg.
2 Average of estimated CAFE requirements.
3 Mile per gallon equivalent (MPG-e) is the corresponding fleet average fuel economy value if the entire fleet were
to meet the CO2 standard compliance level through tailpipe CO2 improvements that also improve fuel economy.
This is provided for illustrative purposes only, as we do not expect the GHG standards to be met only with fuel
efficiency technology.

   The agencies' updated assessments provide projections for the MY2022-2025 standards for
several key metrics, including modeled "low-cost pathway" technology penetrations, per-vehicle
average costs (cars, trucks, and fleet, by manufacturer and total industry-wide), industry-wide
average costs, GHG and oil reductions, consumer payback, consumer fuel savings, and benefits
analysis.

   Based on the extensive updated assessments provided in this Draft TAR, the projections for
the average per-vehicle costs of meeting the MY2025 standards (incremental to the costs already
incurred to meet the MY2021 standard) are, for EPA's analysis of the GHG program, $894 -
$1,017, and, for NHTSA's analysis of the CAFE program, $1,245 in the primary analysis using
Retail Price Equivalent (RPE), and $1,128 in a sensitivity case analysis using Indirect Cost
Multipliers (ICM).  In the 2012 final rule, the estimated costs for meeting the MY2022-2025
GHG standards (incremental to the costs for meeting the MY2021 standard in MY2021) was
$l,070.3'vii
! This cost estimate from the 2012 final rule was based on the use of Indirect Cost Multipliers (ICMs) in 2010$.
                                            ES-8

-------
                                                                      Executive Summary
    Table ES- 2 Per-Vehicle Average Costs to Meet MY2025 Standards: Draft TAR Analysis
       Costs Shown are Incremental to the Costs to Meet the MY2021 Standards

Car
Truck
Combined
GHG1 in MY2025
Primary Analysis
$707
$1,099
$894
RPE Sensitivity
Case3
$789
$1,267
$1,017
CAFE in MY 2028
Primary Analysis2
$1,207
$1,289
$1,245
ICM Sensitivity
Case3
$1,156
$1,096
$1,128
Notes:
1 The values reported for the GHG analysis to account for indirect costs reflect the use of Indirect Cost Multipliers
for the primary analysis, and Retail Price Equivalent for the sensitivity case.
2 The values reported for CAFE primary analysis reflect the use of RPE and include civil penalties estimated to be
incurred by some OEMs as provided by EPCA/EISA. Estimated technology costs (without civil penalties) average
$ 1,111, $ 1,246, and $ 1,174, respectively for MY2028 passenger cars, light trucks, and the overall light-duty fleet.
3 Note that Chapter 12 (GHG) and Chapter 13 (CAFE) include a wide range of additional sensitivity cases.

   In  Table ES-2, NHTSA's estimates are provided for MY2028 because NHTSA's analysis,
which is conducted on a year-by-year basis, indicates that manufacturers could make use of
EPCA/EISA's provisions allowing credits to be earned and carried forward to be applied toward
ensuing model years.  Therefore, NHTSA's analysis indicates that a "stabilized" response to the
augural standards might not be achieved until approximately 2028 (see Chapter 13 for additional
detail). EPA estimates are provided for MY2025 because EPA's analysis projects that each
manufacturer would comply in MY2025 with that year's standards (see Chapter 12  for additional
details).

   Table ES-3 shows fleet-wide penetration rates for a subset of the technologies the agencies'
project could be utilized to comply with the MY2025 standards. While all three agencies have
been working collaboratively on an array of issues throughout this initial phase of the Midterm
Evaluation, much of the EPA GHG and NHTSA CAFE assessments were done largely
independently, as reflected in the different technology pathways shown in Table ES-3 (see
Chapter 2.3  for additional detail). The agencies' analyses each project that the MY2022-2025
standards can be met largely through improvements in gasoline vehicle technologies, such as
improvements in engines, transmissions, light-weighting, aerodynamics, and accessories. The
analyses further indicate that only modest amounts of hybridization, and very little full
electrification (plug-in hybrid electric vehicles (PHEV) or electric vehicles (EV)) technology
will be needed to meet the standards. This initial assessment of potential technology paths is
similar to the agencies' projections made in the 2012 final rule, and is consistent with the
findings of the National Academy of Sciences report from June 2015 (discussed in Chapter 2).
                                           ES-9

-------
                                                                        Executive Summary
                Table ES- 3 Selected Technology Penetrations to Meet MY2025 Standards1

Turbocharged and downsized
gasoline engines
Higher compression ratio, naturally
aspirated gasoline engines
8 speed and other advanced
transmissions2
Mass reduction
Stop-start
Mild Hybrid
Full Hybrid
Plug-in hybrid electric vehicle3
Electric vehicle3
GHG
33%
44%
90%
7%
20%
18%
<3%
<2%
<3%
CAFE
54%
<1%
70%
6%
38%
14%
14%
<1%
<2%
Notes:
1 Percentages shown are absolute rather than incremental. These values reflect both EPA and NHTSA's primary
analyses; both agencies present additional sensitivity analyses in Chapter 12 (GHG) and Chapter 13 (CAFE). For
EPA this includes a pathway where higher compression ratio naturally aspirated gasoline engines are held at a 10%
penetration, and the major changes are turbocharged and downsized gasoline engines increase to 47% and mild
hybrids increase to 38% (See Chapter 12.1.2)
2 Including continuously variable transmissions (CVT)
3 In EPA's modeling, the California Zero Emission Vehicles (ZEV) program is considered in the reference case
fleet; therefore, 3.5% of the fleet is projected to be full EV or PHEV in the 2022-2025 timeframe due to the ZEV
program and the adoption of that program by nine additional states.

   Although some of the differences in costs are expected as EPA and NHTSA conducted two
independent analyses, the consideration of CARS's program also led to one important
difference. As noted in the footnote for Table ES-3, EPA's analysis included consideration for
compliance with other related state regulations including CARB's ZEV regulation that has also
been adopted by nine other states under Section 177 of the Federal Clean Air Act. CARB's ZEV
program requires a portion of new light-duty vehicle sales to be ZEVs and collectively, CA and
these states represent nearly 30 percent of nationwide  sales of light-duty vehicles. CARB
worked with EPA to include ZEVs reflecting compliance with California's ZEV program within
the reference fleet used by EPA.  NHTSA's analysis did not.  This  accounts for at least part of
the cost differences in the two agencies' analyses as well as for  some of the difference in
technology penetration rates for full hybrids.

   EPA's analysis indicates that, compared to the MY2021  standards, the MY2025 standards
will result in a net lifetime consumer savings  of $1,460 - $1,620 and a payback of about 5  to 5 1A
years.4  NHTSA's primary analysis indicates that net lifetime consumer savings could average
$680 per vehicle, such that increased vehicle purchase costs are paid back within about 6 /^
years, and $800 with payback within about 6 years in a sensitivity case analysis using ICMs.
4 Based on the AEO 2015 reference case gasoline price projections, 3 percent discount rate, and ICMs.
                                           ES-10

-------
                                                                     Executive Summary
     Table ES- 4 Payback Period and Lifetime Net Consumer Savings for an Average Vehicle Compared to
                                        the MY2021 Standards

Payback period
(years)
Net Lifetime
Consumer Savings
($, discounted at 3%)
GHG
MY2025 Vehicle
Primary Analysis
5
$1,620
RPE Sensitivity Case
5/2
$1,460
CAFE
MY2028 Vehicle
Primary Analysis
6/2
$680
ICM Sensitivity Case
6
$800
* Note that Chapter 12 (GHG) and Chapter 13 (CAFE) include a wide range of additional sensitivity cases.
   Over the lifetimes of MY2021-2025 vehicles, EPA estimates that under the GHG standards,
GHG emissions would be reduced by about 540 million metric tons (MMT) and oil consumption
would be reduced by 1.2 billion barrels. Over the lifetimes of MY2016-2028 vehicles, NHTSA
estimates that under the augural MY2022-2025 CAFE standards,  GHG emissions would be
reduced by about 748 MMT and oil consumption would be reduced by about  1.6 billion barrels.
NHTSA's estimates span a wider range of model years for two reasons, as discussed in Chapter
13: first, the NHTSA analysis projects that manufacturers may take some "early action" prior to
MY2022; second, as discussed above, the response to the augural standards might not be
"stabilized" until after MY2025.  Differences in these values also result from  differences in the
agencies' estimates of annual mileage accumulation by light-duty vehicles.5
         Table ES- 5  Cumulative GHG and Oil Reductions for Meeting the MY2022-2025 Standards
Lifetime Reductions
CO2e reduction
(million metric tons, MMT)
Oil reduction (billion barrels)
GHG
( MYs 2021-2025 vehicles)
540
1.2
CAFE
(MYs 2016-2028 vehicles)
748
1.6
   For the EPA GHG analysis, total industry-wide costs of meeting the MY2022-2025 GHG
standards are estimated at $34 to $38 billion.  Societal monetized benefits of the MY2022-2025
standards (exclusive of fuel savings to consumers) range from $40 to $41 billion.  Consumer pre-
tax fuel  savings are estimated to be $89 billion over the lifetime of vehicles meeting the
MY2022-2025 standards. Net benefits (inclusive of fuel savings) are estimated at $90 to $94
billion.  These values are all at a 3 percent discount rate; values at a 7 percent discount rate are
shown in Table ES-6 below.
5 The agencies' methods for assessing vehicle mileage accumulation are discussed in Chapter 10.3 for EPA, and
  Chapter 13 for NHTSA.
                                          ES-11

-------
                                                                        Executive Summary
       Table ES- 6  GHG Analysis of Lifetime Costs & Benefits to Meet the MY2022-2025 Standards (for
                            Vehicles Produced in MY2021-2025)* (Billions of 2013$)

Vehicle Program
Maintenance
Fuel
Benefits*
Net Benefits
3 Percent Discount Rate
Primary Analysis
-$34
-$2
$89
$41
$94
RPE Sensitivity
Case
-$38
-$2
$89
$40
$90
7 Percent Discount Rate
Primary
Analysis
-$24
-Si
$49
$30
$54
RPE Sensitivity
Case
-$27
-Si
$49
$30
$51
Note:
These values reflect AEO 2015 reference fuel price case. The Primary Analysis reflects ICMs and the Sensitivity
Case reflects RPEs.  All values are discounted back to 2015; see Chapter 12.3 for details on discounting social cost
of GHG and non-GHG benefits. Note that Chapter 12 also includes a number of additional sensitivity cases.

   NHTSA's primary analysis shows that compared to the No Action alternative, the augural
CAFE standards could entail additional costs totaling $87 billion during MYs 2016-2028
(reasons for this span of MYs are discussed above), and a sensitivity case using ICM shows total
costs  of $79 billion. The primary analysis shows benefits totaling $175 billion, and the ICM
sensitivity case shows $178 billion.  Consumer fuel savings are estimated to be $67 billion to
$122  billion over the lifetime  of vehicles meeting the MY2022-2025 standards.  Thus, net
benefits (inclusive of fuel savings) could total $88 billion based on the primary analysis and $99
billion for the ICM sensitivity case.  These are estimates of the present value (in 2015) of costs
and benefits, based on a 3 percent discount rate. NHTSA has also conducted analysis using a 7
percent discount rate, and a broader sensitivity analysis to examine the impact of other key
analysis inputs, as discussed in Chapter 13. Below, Table ES-7 provides an overall summary of
costs  and benefits observed in NHTSA's analysis.
      Table ES- 7 CAFE Analysis of Lifetime Costs & Benefits to Meet the MY2022-2025 Standards (for
                             Vehicles Produced in MY2016-2028) (Billions of 2013$)

Vehicle Program1*
Benefits (Fuel)
Benefits (Other)
Net Benefits
3 Percent Discount Rate
Primary Analysis2
-$87
$120
$55
$88
ICM Sensitivity Case3
-$79
$122
$56
$99
7 Percent Discount Rate
Primary Ana lysis
-$60
$67
$43
$50
Notes:
1 Includes changes in maintenance costs (small relative to cost of additional technology).
2 The Primary Analysis reflects RPE.
3 Note that Chapter 13 includes a wide range of additional sensitivity cases.

As noted above, because EPA and NHTSA developed independent assessments of technology
cost, effectiveness, and reference case projections, the compliance pathways and associated costs
that result are also different.  Consideration of these two results provides greater confidence that
compliance can be achieved through a number of different technology pathways.
                                           ES-12

-------
                                                                              Executive Summary
References
1 See 40 CFR 86.1818-12(h).
11 See 40 CFR 86.1818-12(h).
111 See 40 CFR section 86.181-12(h).
1V See 77 FR 62784 (Oct. 12, 2012).
v See 40 CFR 86.1818-12(h)(2)(i).
V1 National Academy of Sciences, National Research Council to the National Academies, "Cost, Effectiveness and
Deployment of Fuel Economy Technologies for Light-Duty Vehicles," June 2015.
vl1 Regulatory Impact Analysis:  Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas Emissions
Standards and Corporate Average Fuel Economy  Standards, EPA-420-R-12-016, Table 5.1-8, page 5-8.
                                               ES-13

-------
                                                                          Introduction
Table of Contents

Chapter 1:  Introduction	1-1
   1.1    Purpose of this Report	1-1
   1.2    Building Blocks of the National Program	1-3
     1.2.1   Background on NHTSA's CAFE Program	1-3
     1.2.2   Background on EPA'sGHG Program	1-5
     1.2.3   Background on C ARE'sGHG Program	1-5
   1.3    Background on the National Program	1-5
   1.4    Agencies' Commitment to the Midterm Evaluation (MTE)	1-11
   1.5    Climate Change and Energy Security Drivers for the National Program	1-12
     1.5.1   Climate Change	1-13
       1.5.1.1   Overview of Climate Change Science and Global Impacts	1-13
       1.5.1.2   Overview of Climate Change Impacts in the United States	1-17
       1.5.1.3   Recent U.S. Commitments on Climate Change Mitigation	1-19
       1.5.1.4   Recent California Commitments on Climate Change	1-20
       1.5.1.5   Contribution of Cars and Light Trucks to the U.S. Greenhouse Gas Emissions
       Inventory	1-20
       1.5.1.6   Importance of the National Program in the U. S. Climate Change Program ... 1 -21
     1.5.2   Petroleum Consumption and Energy Security	1-22
       1.5.2.1   Overview of Petroleum Consumption and Energy Security	1 -22
       1.5.2.2   Recent U.S. Commitments on Petroleum and Energy Security	1-23
       1.5.2.3   Contribution of Cars and Light Trucks to U.S. Petroleum Consumption	1-23
       1.5.2.4   Importance of National Program to Petroleum Consumption and Energy
       Security 1-23

Table of Figures
Figure 1.1 CAFE Target Curves for Passenger Cars	1-9
Figure 1.2 CAFE Target Curves for Light Trucks	1-10
Figure 1.3 CO2 (g/mile) Passenger Car Standards Curves	1-10
Figure 1.4 CO2 (g/mile) Light Truck Standards Curves	1-11

-------
                                                                          Introduction
Chapter 1: Introduction

1.1    Purpose of this Report

   The Environmental Protection Agency (EPA) and the National Highway Traffic Safety
Administration (NHTSA) have conducted two joint rulemakings to establish a coordinated
National Program for stringent Federal corporate average fuel economy (CAFE) and greenhouse
gas (GHG) emissions standards for light-duty vehicles. The National Program builds on over 35
years of the National Highway Traffic Safety Administration's (NHTSA's) issuance and
enforcement of the Nation's fuel economy standards under the Energy Policy and Conservation
Act (EPCA), and responds to a 2007 Supreme Court decision determining that greenhouse gases
(GHGs) can be regulated under the Clean Air Act (CAA) and EPA's endangerment finding.  The
agencies finalized the first set of National Program standards covering model years (MYs) 2012-
2016 in May 20101 and the second set of standards, covering MYs 2017-2025, in October 2012.2
The National Program establishes standards that increase in stringency year-over-year from
MY2012 through MY2025, projected to reach a level by 2025 that nearly doubles fuel economy
and cuts GHG emissions in half as compared to MY2008. Through the coordination of the
National Program with the California standards, automakers can build one single fleet of vehicles
across the U.S. that satisfies all GHG/CAFE requirements, and consumers can continue to have a
full range of vehicle choices that meet their needs. In the agencies' October 2012 final rules, the
National Program was estimated to save 6 billion metric tons of carbon dioxide (CO2) pollution
and 12 billion barrels of oil over the lifetime of MY2012-2025 vehicles. In addition, the final
standards are projected to provide significant savings for consumers due to reduced fuel use.

   The rulemaking establishing the National Program for model year MY2017-2025 light-duty
vehicles included a regulatory commitment from EPA to conduct a Midterm Evaluation (MTE)
of the GHG standards established  for MYs 2022-2025.3  The 2012 final rule states "The mid-
term evaluation reflects the rules' long time frame, and, for NHTSA, the agency's statutory
obligation to conduct a de novo rulemaking in order to establish final standards for MYs 2022-
2025. NHTSA will use the MTE as part of the rulemaking it must undertake to set standards for
MYs 2022-2025._Through the MTE, EPA will determine whether the GHG standards for model
years 2022-2025, established  in 2012, are still appropriate, within the meaning of section 202 (a)
of the Clean Air Act, in light of the record then before the Administrator, given the latest
available data and information. See 40 CFR section 86.181-12(h). EPA's decision could go one
of three ways: the standards remain appropriate, the standards should be less stringent, or the
standards should be more stringent. In order to align the agencies' proceedings for MYs 2022-
2025 and to maintain a joint national program, EPA and NHTSA will finalize their actions
related to MYs 2022-2025 standard concurrently. If the EPA determination is that the standards
may change, the agencies will issue a joint NPRM and joint final rules." See 77 FR at 62628
(Oct. 15,2012).

   The MTE is a collaborative, data-driven, and transparent process that will be "a holistic
assessment of all of the factors considered in standards setting," and "the expected impact of
those factors on manufacturers' ability to comply, without placing decisive weight on any
                                             1-1

-------
                                                                             Introduction
particular factor or projection."  See 77 FR 62784 (Oct. 15, 2012). A The MTE analysis is
intended to be as robust and comprehensive as that in the original setting of the MY2017-2025
standards. Id. EPA and NHTSA also are closely coordinating with the California Air Resources
Board (CARB) in conducting the MTE to better ensure the continuation of the National Program.
Id.  The agencies fully expect that any adjustments to the standards will be made in consultation
with CARB.  The details of National Program and the MTE are discussed in Sections 1.2 and 1.3
respectively, below.

   The 2012 final rule preamble also states "Prior to beginning NHTSA's rulemaking process
and EPA's mid-term evaluation, the agencies plan to jointly prepare a Draft Technical
Assessment Report (TAR) to examine afresh the issues and, in doing so, conduct similar
analyses and projections as those considered in the current rulemaking, including technical and
other analyses and projections relevant to each agency's authority to set standards as well as any
relevant new issues that may present themselves."  See 77 FR 62965 (Oct. 15, 2012).  This Draft
Technical Assessment Report (TAR) is the first formal step in the MTE process and is being
issued jointly by EPA, NHTSA, and CARB for public comment. EPA is required to prepare and
seek public comment on the TAR.4  The Draft TAR is a technical report, not a decision
document. It is an opportunity for all three agencies to share with the public the initial technical
analyses of the MY2022-2025 standards.  The Draft TAR is a first step in the process that will
ultimately inform whether the MY2022-2025 GHG standards adopted by EPA in 2012 should
remain in place or should change, and what MY2022-2025 CAFE standards would be maximum
feasible for NHTSA. EPA's regulations require it to consider in the Draft TAR a wide range of
factors relevant to the MY2022-2025 standards5 including:

       •  Powertrain improvements for gasoline and diesel engines
       •  Battery developments for hybridization, electrified vehicles
       •  Technology costs
       •  Vehicle light-weighting and impacts on safety
       •  Market penetration of fuel efficient technologies
       •  Fuel prices
       •  Fleet mix (cars v. trucks)
       •  Employment impacts
       •  Infrastructure for electric vehicle charging, alternative fuels
       •  Consumer acceptance
       •  Consumer payback periods
       •  Any other factors deemed relevant
   The agencies have conducted extensive research and analyses to support the MTE, as
discussed in Chapter 2 and throughout the document. As part of gathering robust data and
information to inform the MTE, the agencies also have conducted extensive outreach with a wide
range of stakeholders - including auto manufacturers, automotive suppliers, NGOs, consumer
A 40 CFR section 86.1818 (h) (1) lists factors which EPA must consider, including "availability and effectiveness of
  the technology;" "the appropriate lead time for introduction of technology;" the feasibility and practicability of
  the standards;" "the impact of the standards on reduction of emissions, oil conservation, energy security and fuel
  savings by consumers;" "the impact of the standards on the automobile industry;" and "the impacts of the
  standards on automobile safety."
                                               1-2

-------
                                                                           Introduction
groups, labor unions, state and local governments, the academic and research communities, and
others. Among other things, the Draft TAR presents analyses reflecting this research and
information obtained during the agencies' outreach, presents updated assessments since the 2012
final rule, including a 2015 assessment by the National Academies of Science, and offers an
opportunity for public comments on our work thus far. The agencies will fully consider public
comments on this Draft TAR as they continue the MTE process, discussed below.

1.2    Building Blocks of the National Program

   The National Program is both needed and possible because the relationship between
improving fuel economy and  reducing CCh tailpipe emissions is very direct and close. The
amount of those CCh emissions is essentially constant per gallon combusted of a given type of
fuel. Thus, the more fuel efficient a vehicle is, the less fuel it burns to travel a given distance.
The less fuel it burns, the less CCh it emits in traveling that distance. While there  are emission
control technologies that reduce the pollutants (e.g., carbon monoxide) produced by imperfect
combustion of fuel by capturing or converting them to other compounds, there is currently no
such technology for CCh. Further, while some of those pollutants can also be reduced by
achieving a more complete combustion of fuel, doing so only increases the tailpipe emissions of
CCh.  Thus, there is a single pool of technologies for addressing these twin problems, i.e., those
that reduce fuel consumption and thereby reduce CCh emissions as well. As noted in the 2012
final rule, the rates of increase in stringency for the CAFE standards are lower than EPA's rates
of increase in stringency for GHG standards for purposes of harmonization and in reflection of
several statutory constraints on the CAFE program.6'6

1.2.1   Background on NHTSA's CAFE Program

   The establishment of national fuel economy standards followed directly from passage of the
Energy Policy and Conservation Act (EPCA) of 1975.  The Act directed the Secretary of
Transportation to set standards separately for passenger cars and light trucks at the maximum
feasible levels in each model  year (with the passenger car standard not to exceed 27.5 mpg), and
provided additional direction regarding many aspects of the program.  The Secretary has
delegated this responsibility to the National Highway Traffic Safety Administration (NHTSA).
The first fuel economy standards took effect in MY 1978.

   Congress has amended EPCA several times to provide further direction.  Through the Energy
Independence and Security Act (EISA) of 2007, Congress directed the Secretary to, among other
things, define future standards in terms of vehicle attributes related to fuel economy and ensure
that those standards cause the overall fleet to achieve an average fuel economy level of at least
35 mpg by 2020. EISA did not otherwise change the requirement that the Department of
Transportation (DOT) set standards separately for passenger  cars and light trucks at the
maximum feasible levels in each model year. NHTSA can only set standards for up to five
model years at a time and standards must be set at least eighteen months before the beginning of
the model year.7

   In the  late 1970s, NHTSA issued regulations to establish and significantly increase the
stringency of fuel economy standards through 1985.  In the 1980s, the Department relaxed the
B For a fuller discussion of these issues, see 77 FR 62639, October 15, 2012.
                                              1-3

-------
                                                                           Introduction
passenger car standards for model years 1986-1989 and then increased the standard to 27.5 mpg.
In 1994, NHTSA issued a notice proposing to explore higher fuel economy standards for light
trucks. However, starting with the fiscal year 1996 and continuing through fiscal year 2001,
Congress prohibited NHTSA from using any funds to increase fuel economy standards.  In 2003,
NHTSA increased light truck standards during model years 2005-2007. In 2006, NHTSA
increased light truck standards during model years 2008-2011 and required an attribute-based
standard in 2011.

   Following EISA and a 2007 decision by the United  States Court of Appeals for the Ninth
Circuit8 (requiring that, when issuing CAFE standards, the Department issue Environmental
Impact Statements and assign an economic value to avoided CCh emissions), the Department
proposed in April 2008 to establish more stringent attribute-based standards for both passenger
cars and light trucks during model years 2011-2015. The Department subsequently completed
work on a rule to finalize these standards; however, with the automobile industry experiencing a
steep decline during 2008, the Department withdrew the rule. Under President Obama, the
Department promulgated the model year 2011 standards in April 2009, and began work on
harmonized DOT fuel economy and EPA GHG standards referred to here as the National
Program.

   As shown below, as required fuel economy standards have increased, passenger car (PC) and
light truck (LT) average fuel economy levels achieved by manufacturers have improved:
      60 -r
      50 --
               CAFE Requirement and Achieved: Overall U.S. Fleet
    QO
    Q.
    E
    >.40 ""
    E
    o
    c
    o
    u
    ^ 30 +
    00
    ra
      20 --
      10 --
-Passenger Car Requirement

•Passenger Car Achieved

•Light Truck Requirement

•Light Truck Achieved
                                              Requirements for 2015 - 2021 are estimated.
                                              Requirements for 2022 - 2025 are based on augural standards.
        1970
 1980
1990
  2000
Model Year
2010
2020
2030
                  Figure 1.1 Average Required and Achieved Fuel Economy Levels
                                              1-4

-------
                                                                          Introduction
   It is important to note that the CAFE fuel economy values (both the required and the
achieved) shown in this chart are based on EPA 2-cycle city and highway tests as required by
Congress.  Accordingly, these values are a minimum of 25 percent higher than the typical fuel
economy values shown on fuel economy labels (which are based on 5-cycle testing that reflects a
much broader range of driving conditions) and achieved by consumers in real world driving.

1.2.2  Background on EPA's GHG Program

   Under the Clean Air Act, EPA is responsible for addressing emissions of air pollutants from
motor vehicles.  On April 2, 2007, the U.S. Supreme Court issued its opinion in Massachusetts v.
EPA,9 a case involving EPA's 2003 denial of a petition for rulemaking to regulate GHG
emissions from motor vehicles under section 202(a) of the Clean Air Act (CAA).10 The Court
held that GHGs fit within the definition of air pollutant in the Clean Air Act and further held that
the Administrator must determine whether or not emissions from new motor vehicles cause or
contribute to air pollution which may reasonably be anticipated to endanger public health or
welfare, or whether the science is too uncertain to make a reasoned decision.  The Court rejected
the argument that EPA cannot regulate CCh from motor vehicles because to do so would de facto
tighten fuel economy standards, authority over which has been assigned by Congress  to DOT.
The Court stated that "[b]ut that DOT sets mileage standards in no way licenses EPA to shirk its
environmental responsibilities. EPA has been charged with protecting the public's 'health' and
'welfare', a statutory obligation wholly independent of DOT's  mandate to promote energy
efficiency." The Court concluded that "[t]he two obligations may overlap, but there is no reason
to think the two agencies cannot both administer their obligations and yet avoid inconsistency."11
The case was remanded back to the Agency for reconsideration in light of the Court's decision.12

   On December 15, 2009, EPA published two findings (74 FR 66496): That emissions of GHGs
from new motor vehicles and motor vehicle engines contribute to GHG air pollution,  and that
GHG air pollution may reasonably be anticipated to endanger public health and welfare of
current and future generations in the U.S.

1.2.3  Background on CARB's GHG Program

   Recognizing the increasing threat of climate change to the well-being of California's citizens
and the environment, in 2002 the state  legislature passed assembly bill 1493 (AB 1493) which
directed CARB to adopt the maximum feasible and cost-effective reductions in GHG emissions
from passenger cars and light-duty trucks beginning in the 2009 model year. Accordingly, in
2004, CARB adopted the first in the  nation GHG emission requirements for passenger cars and
light-duty trucks for model years 2009-2016. In January 2012, CARB adopted additional light-
duty vehicle GHG emission requirements for model years 2017-2025.  These additional
requirements were developed in a joint effort with EPA and NHTSA on the development of
corporate fuel economy and federal GHG emission standards for model year 2017 and beyond.

1.3    Background on  the National  Program

   NHTSA and EPA have conducted two joint rulemakings to  establish a National Program for
corporate average fuel  economy (CAFE) and GHG emissions standards.  Together, the two rules
established  strong and coordinated Federal  GHG and fuel economy standards for passenger cars,
light-duty trucks, and medium-duty passenger vehicles (hereafter light-duty vehicles or LDVs).
Each agency adopted standards covering MYs 2012-2016 in May 201013 and covering MY2017
                                             1-5

-------
                                                                            Introduction
and beyond in October 2012.14  The MYs 2012-2016 rule represented the first time EPA
established standards for GHG emissions under its Clean Air Act authority.  The Federal GHG
and fuel economy standards for MY2017 and beyond were developed in a joint effort with
CARB. And, subsequent to the adoption of California-specific GHG standards for MYs 2017-
2025 and the adoption of the Federal standards for MY2017 and beyond, CARB adopted a
"deemed to comply" provision whereby compliance with the Federal GHG standards would be
deemed as compliance with California's GHG program in furtherance of a single National
Program. The National Program approach, combined with California standards, helps to better
ensure that all manufacturers can build a single fleet of vehicles that satisfy all requirements
under both federal programs and under California's program, which helps to reduce costs and
regulatory complexity for auto manufacturers. In addition, the National Program provides
significant environmental and climate benefits, energy security, and consumer savings to the
general public. Most stakeholders strongly supported the National Program, including the auto
industry, automotive suppliers, state and local governments, labor unions, NGOs, consumer
groups, veterans groups, and others.

   Together, light-duty vehicles, which  include passenger cars, sport utility vehicles, crossover
utility vehicles, minivans, and pickup trucks, are presently responsible for approximately 60
percent of all U.S. transportation-related GHG emissions and fuel consumption.15 The 2012 final
rule projected that combined, the National Program standards, and NHTSA's MY2011 CAFE
standards, result in MY2025 light-duty vehicles with nearly double the fuel economy and half
the GHG emissions compared to MY2010 vehicles. Collectively, these represented some of the
most significant federal actions ever taken to reduce GHG emissions and improve fuel economy
in the U.S.  In the 2012 final rule, based on future assumptions including car/truck share, EPA
projected that its standards would lead to an average industry fleet wide emissions level of 163
grams/mile of carbon dioxide (CCh) in model year 2025 (compared to 326 g/mile in MY 2011),
which is equivalent to 54.5 mpg if this level were achieved solely through improvements in fuel
economy.C'D  In the same notice, NHTSA estimated that, if proposed and subsequently finalized
at levels announced on an augural basis for model years 2022-2025, CAFE standards could
increase industry-wide fuel economy to 48.7-49.7 mpg by model year 2025, depending on a
range of factors.

   In the 2012 final rule, the agencies projected that, in meeting the MY2025 standards, a wide
range of vehicles would continue to be available, preserving consumer choice.  The agencies
projected that the MY2025 standards would be met largely through advancements in
conventional vehicle technologies, including advances in gasoline engines (such as
downsized/turbocharged engines) and transmissions, vehicle weight reduction, improvements in
c 163g/mi would be equivalent to 54.5 mpg, if the entire fleet were to meet this CC>2 level through tailpipe COa and
  fuel economy improvements. However, the agencies projected in the 2012 rulemaking analysis that a portion of
  these improvements will be made through improvements in air conditioning refrigerant leakage and the use of
  alternative refrigerants, which would contribute to reduced GHG emissions but would not contribute to fuel
  economy improvements. This is why NHTSA's 48.7-49.7 mpg range differs from EPA's projected 54.5 mpg
  standard.
D Real-world CC>2 is typically 25 percent higher and real-world fuel economy is typically 20 percent lower than the
  CO2 and CAFE compliance values discussed here.
                                               1-6

-------
                                                                            Introduction
vehicle aerodynamics, more efficient vehicle accessories, and lower rolling resistance tires. The
agencies also projected that vehicle air conditioning systems would continue to improve by
becoming more efficient and by increasing the use of alternative refrigerants and lower leakage
systems. The agencies estimated that some increased electrification of the fleet would occur
through the expanded use of stop/start and mild hybrid technologies, but projected that meeting
the MY2025  standards would require only about five to nine percent of the fleet to be full hybrid
electric vehicles (HEVs) and only about two to three percent of the fleet to be electric vehicles
(EV) or plug-in hybrid electric vehicles (PHEVs).E  All of these technologies were available at
the time of the final rule, some on a limited number of vehicles while others were more
widespread, and the agencies projected that manufacturers would be able to meet the standards
through significant efficiency improvements in the technologies, as well as through increased
usage of these and other technologies across the fleet.

   In the 2012 final rule, EPA adopted standards through MY2025, with the MY2022-2025
standards subject to the midterm evaluation process established in the EPA regulations. As
mentioned above, NHTSA adopted standards only through MY2021, due to a statutory
requirement of the Energy Policy and Conservation  Act (EPCA) of 1975, as amended by the
Energy Independence and Security Act (EISA) of 2007, which allows NHTSA to set CAFE
standards for  only up to five model years at a time.  Due to this statutory requirement, NHTSA
must conduct a full de novo rulemaking to establish standards for MYs 2022-2025. In the 2012
final rule, NHTSA thus presented MY2022-2025 standards as "augural," reflecting the agency's
best judgment of what standards would have been maximum feasible at the time  of the final rule,
based on the information then available.  The future rulemaking to set MY2022-2025 CAFE
standards must be based on the best information, data, and analysis available at the time of the
new rulemaking.

   The MY2012-2016 and MY2017 and beyond CAFE and GHG emissions standards are
attribute-based standards/ using vehicle footprint as the attribute.  Footprint is defined as a
vehicle's wheelbase multiplied by its average track width16—in other words, the  area enclosed
by the points  at which the wheels meet the ground.  The standards are therefore generally based
on a vehicle's size: larger vehicles have numerically less stringent fuel economy/GHG
emissions targets and smaller vehicles have numerically more stringent fuel economy/GHG
emissions targets.

   Under the  footprint-based standards, the footprint curve defines a GHG or fuel economy
performance target for each separate car or truck footprint. Individual vehicles or models,
however, are  not required to meet the target on the curve. To determine its compliance
obligation, a vehicle manufacturer would average the curve targets for a given year for each of
its footprints  of its vehicle models produced in that year, as weighted by the number of vehicles
it produced of each model.0 Each manufacturer thus will have a GHG and  CAFE average
E For comparison to vehicles for sale today, an example of a mild HEV is GM's eAssist (Buick Lacrosse), a strong
  HEV is the Toyota Prius, an EV is the Nissan Leaf, and a PHEV is the Chevrolet Volt.
F Attribute-based standards are required by EISA (49 U.S.C. 32902(b)(3)) and allowed by the CAA. NHTSA first
  used the footprint attribute in its Reformed CAFE program for light trucks for model years 2008-2011 and
  passenger car CAFE standards in MY2011.
G See, e.g., 49 CFR 531.5 for the curve equations for passenger car CAFE standards.
                                               1-7

-------
                                                                             Introduction
standard that is unique to each of its car and truck fleets, depending on the footprints and
production volumes of the vehicle models produced by that manufacturer in a given model year.
A manufacturer will have separate footprint-based standards for passenger cars (like sedans,
station wagons, and many 2WD sport-utility vehicles and crossovers) and for light trucks (like
most 4WD and heavier 2WD sport-utility vehicles, minivans, and pickup trucks)H. The curves
are mostly sloped, so that generally, vehicles with larger footprints will be subject to higher CCh
grams/mile targets and lower CAFE mpg targets than vehicles with smaller footprints.  This is
because, generally speaking, smaller vehicles are more capable of achieving lower levels of CCh
and higher levels of fuel economy than larger vehicles. Although a manufacturer's fleet average
standards  could be estimated throughout the model year based on the projected production
volume of its vehicle fleet (and are estimated as part of the EPA certification process), the final
standards  with which each manufacturer must comply are determined by its final model year
production figures.  A manufacturer's calculation of its fleet average standards as well as its
fleets' average performance at the end of the model year will thus be based on  the production-
weighted average target and performance of each model in its fleet.1

   The footprint curves for the MY2012-2025 CAFE standards are shown below in Figure 1.1
and Figure 1.2 and GHG standards are shown below in Figure 1.3 and Figure 1.4. As noted
above, NHTSA has only adopted standards through MY2021.  The CAFE MY2022-2025 curves
provided below were presented as augural attribute curves in the MY2017-2025 rule, and will
have to be re-evaluated as part of the upcoming rulemaking to establish final CAFE standards for
those model years. Although the general model of the target curve equation is  the same for each
vehicle category and each year, the parameters of the curve equation differ for  cars and trucks.
Each parameter also changes on a model year basis, resulting in the yearly increases in
stringency.17
H This is required for the CAFE program under 49 U.S.C. § 32902.
1 A manufacturer may have some models that exceed their target, and some that are below their target. Compliance
  with a fleet average standard is determined by comparing the fleet average standard (based on the production
  weighted average of the target levels for each model) with fleet average performance (based on the production
  weighted average of the performance for each model).
                                               1-8

-------
                                                                                       Introduction
   70
   60
1
                                   »          t   ^
	2025
     2024
     202 J
   •  2022
   -- 2021
   •  2020
    -2019
   — 2018
     2017
	2016
   -  2015
   -  2014
   •- 20U
	2012
   20
     30
                                         50       55
                                               Footprint (sf)
                                                                    65
                                                                             .'ii
                                                                                      75
                        Figure 1.1 CAFE Target Curves for Passenger Cars
                                                    1-9

-------
                                                                             Introduction
  70
  1,0
  50
•f 40
  30
  20
== -=_—-=-' =. >! ^ * *^"*$^*^
~-------------^^^	
	2025
	 2024
	2023
 -  2022
 • -2021
- •  2020
	2019
- •  2018
	2017
	 2016
    2015
  - 2014
  - 2013
	2012
               40      45      50      55      60      65
                                  Footprint (sf)

                 Figure 1.2 CAFE Target Curves for Light Trucks
                                                                  70
                                                                          7S
                                                                                 80
                                                                                         — 201
                                                                                         ••• 2013
                                                                                         . — 2014
                                                                                         -  2015
                                                                                         •  2016
                                                                                         — 2017
                                                                                         -  2018
                                                                                         . 2019
                                                                                         ...2020
                                                                                         	2021
                                                                                         — 2022
                                                                                         — 2023
                                                                                         »  2024
                                                                                         	2025
                                       55        60
                                        Footprint (sf)
                   Figure 1.3  CCh (g/mile) Passenger Car Standards Curves
                                             1-10

-------
                                                                           Introduction
                                         55       60
                                          Footprint (sf)
                      Figure 1.4  CCh (g/mile) Light Truck Standards Curves

   Footprint-based standards help to distribute the burden of compliance across all vehicle
footprints and across all manufacturers. Manufacturers are not compelled to build vehicles of
any particular size or type, and each manufacturer has its own fleetwide standard for each fleet in
each year that reflects the light-duty vehicles it chooses to produce. This approach also preserves
consumer choice, as the standards do not constrain consumers' opportunity to purchase the size
of vehicle with the performance, utility and safety features that meet their needs.

1.4    Agencies' Commitment  to the Midterm Evaluation (MTE)

   Given the long time frame at issue in setting standards for MY2022-2025 light-duty vehicles,
and given NHTSA's statutory obligation to conduct a de novo rulemaking in order to establish
final standards for vehicles for the 2022-2025 model years, the agencies committed in the 2012
final rule to conduct a comprehensive mid-term evaluation for the MY2022-2025 standards.
The MY2017-2025 final rule noted that in order to align the agencies' proceedings for MYs
2022-2025 and to maintain a joint national program, EPA and NHTSA will finalize their actions
related to MY2022-2025 standards concurrently.

   As noted above, through the MTE, EPA will determine whether the GHG standards for model
years 2022-2025, established in 2012, are still "appropriate," within the meaning of section 202
(a)(l) of the Clean Air Act, given the latest available data and information.  EPA's decision
could go one of three ways: the standards remain  appropriate, the standards should be less
stringent, or the standards should be more stringent. Public input on the Draft TAR, along with
any new data and information, will inform the next step in the MTE process — EPA's Proposed
Determination. The Proposed Determination will be the EPA Administrator's proposal on
                                             1-11

-------
                                                                          Introduction
whether the MY2022-2025 standards are appropriate. The Proposed Determination will be
available for public comment, as required by EPA's regulations.  If the Administrator's proposal
is that the standards should change (either more or less stringent), then this action will be a
Notice of Proposed Rulemaking. Public input on the Proposed Determination, as well as new
data and information available, will inform the next step — EPA's Final Determination.  The
Final Determination will be the Administrator's final decision on whether or not the MY2022-
2025 standards are appropriate, in light of the record then before the Administrator. EPA is
legally bound to make a final determination, by April 1, 2018, on whether the MY2022-2025
GHG standards are appropriate under section 202(a), in light of the record then before the
agency. See generally 40 CFR 86.1818-12(h).

   As stated above, EPCA limits NHTSA to setting CAFE standards for up to five years at a
time, so that the MY2022-2025 CAFE provisions are only "augural," reflecting NHTSA's best
judgment of what standards would have been maximum feasible at the time of the final rule,
based on the information then available.  The MTE is closely coordinated with NHTSA's plan to
conduct a CAFE rulemaking to establish MY2022-2025 standards and NHTSA committed to
fully participate in the MTE process, including this Draft Technical Assessment Report (TAR).
77 FR 62784. NHTSA's rulemaking will consider all relevant information and fresh balancing
of statutory factors in order to determine the maximum feasible CAFE standards for MYs 2022-
2025. In order to maintain a joint national program by aligning the agencies' proceedings for
MYs 2022-2025, if the EPA determination is that its standards will not change, NHTSA will
issue its final rule concurrently with the EPA final determination.  If the EPA determination is
that standards may change, the agencies will issue a joint NPRM and joint final rule similar to
the previous two joint rulemakings. The public input on the research and analysis presented in
the Draft TAR will inform NHTSA's proposed rule as well as EPA's MTE determination
process.

   NHTSA and EPA are conducting this mid-term evaluation in close coordination with the
California Air Resources Board (CARB), given our commitment to maintaining a National
Program to address GHG emissions and fuel economy. California adopted its own GHG
standards for MYs 2017-2025 in 2012 prior to NHTSA and EPA finalizing the GHG and fuel
economy standards for the National Program.  Through direction from its Board in 2012, CARB
both adopted a 'deemed to comply' provision allowing compliance with EPA's GHG standards
in lieu of CARB's standards, and committed to participating with NHTSA and EPA in
conducting the mid-term evaluation. EPA subsequently granted California's waiver request
under the Clean Air Act on January 9, 2013 for its MY2017-2025 GHG standards.18 To date,
CARB has been involved with the preparation of this Draft TAR to inform the mid-term
evaluation of the National  Program.

   Additionally, CARB  is  scheduled to provide an update to its Board in late 2016 regarding the
status of the mid-term evaluation as well  as a review of California-specific elements of the
CARB Advanced Clean Cars (ACC) program.19

1.5   Climate Change and Energy Security Drivers for the National Program

   The two primary policy drivers for the National Program are to reduce the U.S. contribution
to global climate change (the legal basis for EPA's GHG emissions standards) and to reduce
                                             1-12

-------
                                                                          Introduction
petroleum consumption and improve U.S. energy security (the legal basis for NHTSA's CAFE
standards).

1.5.1   Climate Change

1.5.1.1 Overview of Climate Change Science and Global Impacts

   According to the National Research Council, "Emissions of CCh from the burning of fossil
fuels have ushered in a new epoch where human activities will largely determine the evolution of
Earth's climate. Because CCh in the atmosphere is long lived, it can effectively lock Earth and
future generations into a range of impacts, some of which could become very severe. Therefore,
emission reduction  choices made today matter in determining impacts experienced not just over
the next few decades, but in the coming centuries and millennia."20

   In 2009, based on a large body of robust and compelling scientific evidence, the EPA
Administrator issued the Endangerment Finding under CAA section 202(a)(l).21 In the
Endangerment Finding, the Administrator found that the current, elevated concentrations of
GHGs in the atmosphere—already at levels unprecedented in human history—may reasonably
be anticipated to endanger public health and welfare of current and future generations in the U.S.
The D.C. Circuit later upheld the Endangerment Finding from all challenges. Coalition for
Responsible Regulation v. EPA, 684 F. 3d 102, 116-26 (D.C. Cir. 2012).

   Since the administrative record concerning the Endangerment Finding closed following the
EPA's 2010 Reconsideration Denial, the climate has continued to change, with new records
being set for a number of climate indicators such as global average surface temperatures, Arctic
sea ice retreat, CCh concentrations, and sea level rise. Additionally, a number of major scientific
assessments have been released that improve understanding of the climate system and strengthen
the case that GHGs endanger public health and welfare both for current and future generations.
These assessments, from the Intergovernmental Panel on Climate Change (IPCC), the U.S.
Global Change Research Program (USGCRP), and the National Research Council (NRC),
include: IPCC's 2012 Special Report on Managing the Risks of Extreme Events and Disasters to
Advance Climate Change Adaptation (SREX) and the 2013-2014 Fifth Assessment Report
(AR5), the USGCRP's 2014 National Climate Assessment, Climate Change Impacts in the
United States (NCA3), and the NRC's 2010 Ocean Acidification:  A National  Strategy to Meet
the Challenges of a Changing Ocean (Ocean Acidification), 2011 Report on Climate
Stabilization Targets: Emissions, Concentrations, and Impacts over Decades to Millennia
(Climate Stabilization Targets),  2011 National Security Implications for U.S. Naval Forces
(National Security Implications), 2011 Understanding Earth's Deep Past: Lessons for Our
Climate Future (Understanding Earth's Deep Past), 2012 Sea Level Rise for the Coasts of
California, Oregon, and Washington: Past, Present, and Future, 2012 Climate and Social Stress:
Implications for Security Analysis (Climate and Social Stress), and 2013 Abrupt Impacts of
Climate Change (Abrupt Impacts) assessments.

   The findings of the recent scientific assessments confirm  and strengthen the science that
supported the 2009 Endangerment Finding.  The NCA3 indicates that climate change "threatens
human health and well-being in many ways, including impacts from increased extreme weather
events, wildfire, decreased air quality, threats to mental health, and illnesses transmitted by food,
water, and disease-carriers such as mosquitoes and ticks."22  Most recently, the USGCRP
released a new assessment, "The Impacts of Climate Change on Human Health in the United
                                             1-13

-------
                                                                            Introduction
States: A Scientific Assessment" (also known as the USGCRP Climate and Health Assessment).
This assessment finds that "climate change impacts endanger our health" and that in the United
States we have "observed climate-related increases in our exposure to elevated temperatures;
more frequent, severe, or longer lasting extreme events; diseases transmitted through food, water,
or disease vectors such as ticks and mosquitoes; and stresses to mental health and well-being."
The assessment determines that "[e]very American is vulnerable to the health impacts associated
with climate change." Climate warming will also likely "make it harder for any given regulatory
approach to reduce ground-level ozone pollution," and, unless offset by reductions of ozone
precursors, it is likely that "climate-driven increases in ozone will cause premature deaths,
hospital visits, lost school days, and acute respiratory symptoms."23

   Assessments state that certain populations are particularly vulnerable to climate change.  The
USGCRP Climate and Health Assessment assesses several disproportionately vulnerable
populations, including those with low income, some communities of color, immigrant groups,
indigenous peoples, pregnant women, vulnerable occupational groups, persons with disabilities,
and persons with preexisting or chronic medical conditions. The Climate and Health Assessment
also concludes that children's  unique physiology and developing bodies contribute to making
them particularly vulnerable to climate change. Children also have unique behaviors and
exposure pathways that could  increase their exposure to environmental stressors, like
contaminants in dust or extreme heat events. Impacts from climate change on children are likely
expected from heat waves, air pollution, infectious and waterborne illnesses, disruptions in food
safely and security, and mental health effects resulting from extreme weather events. For
example, climate change can disrupt food safety and security by significantly reducing food
quality, availability and access.  Children are more susceptible to this disruption because
nutrition is important during critical windows of development and growth.  Older people with
pre-existing chronic heart or lung disease are at higher risk of mortality and morbidity both as a
result of climate warming and during extreme heat events. Pre-existing chronic disease also
increases susceptibility to adverse cardiac and respiratory impacts of air pollution and to more
severe consequences from infectious and waterborne diseases.  Limited mobility among older
adults can also increase health risks associated with extreme weather and floods.

   The new assessments also confirm and strengthen the science that supported the 2009
Endangerment Finding.  The NRC assessment Understanding Earth's Deep Past stated that "[b]y
the end of this century, without a reduction in emissions, atmospheric CCh is projected to
increase to  levels that Earth has not experienced for more than 30 million years."  In fact, that
assessment stated that "the magnitude and rate of the present GHG increase place the climate
system in what could be one of the most severe increases in radiative forcing of the global
climate system in Earth history."24  Because of these unprecedented changes in atmospheric
concentrations, several assessments state that we may be approaching critical, poorly understood
thresholds.  The NRC Abrupt  Impacts report analyzed the potential for abrupt climate change in
the physical climate system and abrupt impacts of ongoing changes that, when thresholds are
crossed, could cause abrupt impacts for society and ecosystems.  The report considered
destabilization of the West Antarctic Ice Sheet (which could cause 3-4 m of potential sea level
rise) as an abrupt climate impact with unknown but probably low probability of occurring this
century. The report categorized a decrease in ocean oxygen content (with attendant threats to
aerobic marine life); increase in intensity, frequency, and duration of heat waves;  and increase in
frequency and intensity of extreme precipitation events (droughts, floods, hurricanes, and major
                                              1-14

-------
                                                                           Introduction
storms) as climate impacts with moderate risk of an abrupt change within this century.  The NRC
Abrupt Impacts report also analyzed the threat of rapid state changes in ecosystems and species
extinctions as examples of an irreversible impact that is expected to be exacerbated by climate
change. Species at most risk include those whose migration potential is limited, whether because
they live on mountaintops or fragmented habitats with barriers to movement, or because climatic
conditions are changing more rapidly than the species can move or adapt. While some of these
abrupt impacts may be of low or moderate probability in this century, the probability for a
significant change in many of these processes after 2100 was judged to be higher, with severe
impacts likely should the abrupt change occur. Future temperature changes will be influenced by
what emissions path the world follows. In its high emission scenario, the IPCC AR5 projects
that global temperatures by the end of the century will likely be 2.6°C to 4.8°C (4.7 to 8.6°F)
warmer than today.  There is very high confidence that temperatures on land and in the Arctic
will warm even faster than the global average.  However, according to the NCA3, significant
reductions in emissions would lead to noticeably less future warming beyond mid-century, and
therefore less impact to public health and welfare.  According to the NCA3, regions closer to the
poles are projected to receive more precipitation, while the dry subtropics expand (colloquially,
this has been summarized as wet areas getting wet and dry regions getting drier), while "[t]he
widespread trend of increasing heavy downpours is expected to continue, with precipitation
becoming less frequent but more intense." Meanwhile, the NRC Climate Stabilization Targets
assessment found that the area burned by wildfire in parts of western North America is expected
to grow by 2 to 4 times for 1°C (1.8°F) of warming.  The NCA also found that "[extrapolation
of the present observed trend suggests an essentially ice-free Arctic in summer before mid-
century." Retreating snow and ice, and  emissions of carbon dioxide and methane released from
thawing permafrost, are very likely to amplify future warming.

   Since the 2009 Endangerment Finding, the IPCC AR5, the USGCRP NCA3, and three of the
new NRC assessments provide estimates of projected global average sea level rise. These
estimates, while not always  directly comparable as they assume different emissions scenarios
and baselines, are at least 40 percent larger than, and in some cases more than twice as large as,
the projected rise estimated  in the IPCC AR4 assessment,  which was referred to in the 2009
Endangerment Finding.  The NRC Sea Level Rise  assessment projects a global average sea level
rise of 0.5 to 1.4 meters by 2100.  The NRC National Security Implications assessment suggests
that "the Department of the Navy should expect roughly 0.4 to 2 meters global average sea-level
rise by 2100." The NRC Climate Stabilization Targets assessment states that a global average
temperature increase of 3°C will lead to a global average sea level rise of 0.5 to 1  meter by 2100.
These NRC and IPCC assessments continue to recognize and characterize the uncertainty
inherent in accounting for melting ice sheets in sea level rise projections.

   Carbon dioxide in particular has unique impacts on ocean ecosystems.  The NRC Climate
Stabilization Targets assessment found that coral bleaching will likely increase due both to
warming and ocean acidification. Ocean surface waters have already become 30 percent more
acidic over the past 250 years due to absorption of CCh from the atmosphere. According to the
NCA3, this "ocean acidification makes water more corrosive,  reducing the capacity of marine
organisms with shells or skeletons made of calcium carbonate (such as corals, krill, oysters,
clams, and crabs) to survive, grow, and reproduce, which in turn will affect the marine food
chain." The NRC Understanding Earth's Deep Past assessment notes four of the five major coral
reef crises of the past 500 million years appear to have been driven by acidification and warming
                                              1-15

-------
                                                                           Introduction
that followed GHG increases of similar magnitude to the emissions increases expected over the
next hundred years.  The NRC Abrupt Impacts assessment specifically highlighted similarities
between the projections for future acidification and warming and the extinction at the end of the
Permian which resulted in the loss of an estimated 90 percent of known species.

   In addition to future impacts, the NCA3 emphasizes that climate change driven by human
emissions of GHGs is already happening now and it is happening in the U.S. According to the
IPCC AR5 and the NCA3, there are a number of climate-related changes that have been
observed recently, and these changes are projected to accelerate in the future:

       •   The planet warmed about 0.85°C (1.5°F) from 1880 to 2012.  It is extremely likely
          (>95 percent probability) that human influence was the dominant cause of the
          observed warming since the mid-20th century, and likely (>66 percent probability)
          that human influence has more than doubled the probability of occurrence of heat
          waves in some locations.  In the Northern Hemisphere, the last 30 years were likely
          the warmest 30 year period of the last 1400 years.
       •   Global sea levels rose 0.19 m (7.5 inches) from 1901 to 2010.  Contributing to this
          rise was the warming of the oceans and melting of land ice. It is likely that 275
          gigatons per year of ice melted from land glaciers  (not including ice sheets) since
           1993, and that the rate of loss of ice from the Greenland and Antarctic ice sheets
          increased substantially in recent years, to 215 gigatons per year and 147 gigatons per
          year respectively since 2002. For context, 360 gigatons of ice melt is sufficient to
          cause global sea levels to rise 1 mm.
       •   Annual mean Arctic sea ice has been declining at 3.5 to 4.1 percent per decade, and
          Northern Hemisphere snow cover extent has decreased at about 1.6 percent per
          decade for March and 11.7 percent per decade for  June.
       •   Permafrost temperatures have increased in most regions since the 1980s, by up to 3°C
          (5.4°F) in parts of Northern Alaska.
       •   Winter storm frequency and intensity have both increased in the Northern
          Hemisphere.  The NCA3 states that the increases in the severity or frequency of some
          types of extreme weather and climate events in recent decades can affect energy
          production and delivery, causing supply disruptions, and compromise other essential
          infrastructure such as water and transportation systems.
   In addition to the changes documented in the assessment literature, there have been other
climate milestones of note. In 2009, the year of the Endangerment Finding, the average
concentration of CCh as measured on top of Mauna Loa was 387 parts per million, far above
preindustrial concentrations of about 280 parts per million.25  The average concentration in 2015
was 401 parts per million, the first time an annual average concentration has exceeded 400 parts
per million since record keeping began at Mauna Loa in 1958, and likely for at least the past
800,000 years.26 Arctic sea ice has continued to decline, with September of 2012 marking the
record  low in terms of Arctic sea ice extent, 40 percent below the 1979-2000 median. Sea level
has continued to rise at a rate of 3.2 mm per year (1.3 inches/decade) since satellite observations
started in 1993,  more than twice the average rate of rise in the 20th century prior to 1993.27  And
2015 was the warmest year globally in the modern global surface temperature record, going back
to 1880, breaking the record previously held by 2014; this now means that the last 15 years have
been 15 of the 16 warmest years on record.28
                                              1-16

-------
                                                                            Introduction
   These assessments and observed changes raise concerns that reducing emissions of GHGs
across the globe is necessary in order to avoid the worst impacts of climate change, and
underscore the urgency of reducing emissions now.  The NRC Committee on America's Climate
Choices listed a number of reasons "why it is imprudent to delay actions that at least begin the
process of substantially reducing emissions."29 For example:

       •   "The faster emissions are reduced, the lower the risks posed by climate change.
          Delays in reducing emissions could commit the planet to a wide range of adverse
          impacts,  especially if the sensitivity of the climate to GHGs is on the higher end of
          the estimated range.
       •   Waiting for unacceptable impacts to occur before taking action is imprudent because
          the effects of GHG emissions do  not fully manifest themselves for decades and, once
          manifested, many of these changes will persist for hundreds or even thousands of
          years.
       •   In the committee's judgment, the risks associated with doing business as usual are a
          much greater concern than the risks associated with engaging in strong response
          efforts."
1.5.1.2 Overview of Climate Change Impacts in the United States

   The NCA3 assessed the climate impacts in eight regions of the U.S., noting that changes in
physical climate parameters such as temperatures, precipitation,  and sea ice retreat were already
having impacts on forests, water supplies, ecosystems, flooding, heat waves, and air quality. The
U.S. average temperatures have similarly increased by 1.3 to 1.9°Fs F since 1895, with most of
that increase occurring since 1970, and the most recent decade was the U.S.'s hottest as well as
the world's hottest. Moreover, the NCA3 found that future warming is projected to be much
larger than recent observed variations in temperature, with 2 to 4°Fs F warming expected in most
areas of the U.S.  over the next few decades,  and up to  10°Fs F possible by the end of the century
assuming continued increases in emissions.  Extreme heat events will continue to become more
common, and extreme cold less common. Additionally, precipitation is considered likely  to
increase in the northern states, decrease in the southern states, and with the heaviest precipitation
events projected to increase everywhere.

   In the Northeast,  temperatures increased almost 2°F from 1895 to 2011, precipitation
increased by about 5 inches (10 percent), and sea level rise of about a foot has led to an increase
in coastal flooding.  In the future, if emissions continue to increase, the Northeast is projected to
experience 4.5 to  10°F of warming by the 2080s. This is expected to lead to more heat waves,
coastal and river flooding, and intense precipitation events.  Sea levels in the Northeast are
expected to increase faster than the global average because of subsidence, and changing ocean
currents may further increase the rate of sea  level rise.

   In the Southeast,  average annual temperature during the last century cycled between warm
and cool periods.  A warm peak occurred during the 1930s and 1940s followed by a cool period
and temperatures then increased again from  1970 to the present by an average of 2°F. Louisiana
has already lost 1,880 square miles of land in the last 80 years  due to sea level rise and other
contributing factors. The Southeast is exceptionally vulnerable to sea level rise, extreme heat
events, hurricanes, and decreased water availability. Major risks of further warming include
significant increases in the number of hot days (95°F or above) and decreases in freezing events,
as well as exacerbated ground level ozone in urban areas. Projections suggest that there  may be
                                              1-17

-------
                                                                            Introduction
fewer hurricanes in the Atlantic in the future, but they will be more intense, with more Category
4 and 5 storms. The NCA identified New Orleans, Miami, Tampa, Charleston, and Virginia
Beach as cities at particular risk of flooding.

   In the Northwest, temperatures increased by about 1.3°F between 1895 and 2011. Snowpack
in the Northwest is an important freshwater source for the region.  More precipitation falling as
rain instead of snow has reduced the snowpack, and warmer springs have corresponded to earlier
snowpack melting and reduced stream flows during summer months.  Drier conditions have
increased the extent of wildfires in the region.  Average annual temperatures are projected to
increase by 3.3°F to 9.7°F by the end of the century (depending on future global GHG
emissions), with the greatest warming is expected during the summer.  Continued increases in
global GHG emissions are projected to result in up to a 30 percent decrease in summer
precipitation. Warmer waters are expected to increase disease and mortality in important fish
species, including Chinook and sockeye salmon. Ocean acidification also threatens species such
as oysters, with the Northwest coastal waters already being some of the most acidified
worldwide due to coastal upwelling and other local factors.

   In Alaska, temperatures have changed faster than anywhere  else in the U.S. Annual
temperatures increased by about 3°F in the past 60 years.  Warming in the winter has been even
greater, rising by an average of 6°F.  Glaciers in Alaska are melting at some of the fastest rates
on Earth. Permafrost soils are also warming and beginning to thaw. Drier conditions had already
contributed to more large wildfires in the 10 years prior to the NCA3 than in any previous
decade since the 1940s, when recordkeeping began, and subsequent years have seen even more
wildfires. By the end of this century, continued increases in GHG emissions are expected to
increase temperatures by 10 to 12°F in the northernmost parts of Alaska, by 8 to 10°F in the
interior, and by 6 to 8°F across the rest of the state.  These increases will exacerbate ongoing
arctic sea ice loss, glacial melt, permafrost thaw and increased wildfire, and threaten humans,
ecosystems, and infrastructure.

   In the Southwest, temperatures are now about 2°F higher than the past century, and are
already the warmest that region has experienced in at least 600 years. The NCA notes that there
is evidence that climate-change induced warming on top of recent drought has influenced tree
mortality, wildfire frequency and area, and forest insect outbreaks.  At the time of publication of
the NCA, even before the last 2 years of extreme drought in California, tree ring data was
already indicating that the region might be experiencing its driest period in 800 years.  The
Southwest is projected to warm an additional 5.5 to 9.5°F over the next century if emissions
continue to increase.  Winter snowpack in the Southwest is projected to decline (consistent with
recent record lows), reducing the reliability of surface water supplies for cities, agriculture,
cooling for power plants,  and ecosystems. Sea level  rise along the California coast is projected
to worsen coastal erosion, increase flooding risk for coastal highways, bridges, and low-lying
airports, and pose a threat to groundwater supplies in coastal cities. Also, "The combination of a
longer frost-free season, less frequent cold air outbreaks, and more frequent heat waves
accelerates crop ripening and maturity, reduces yields of corn, tree fruit, and wine grapes,
stresses livestock, and increases agricultural water  consumption."  Increased drought, higher
temperatures, and bark beetle outbreaks are likely to  contribute to continued increases in
wildfires.
                                              1-18

-------
                                                                            Introduction
   The rate of warming in the Midwest has markedly accelerated over the past few decades.
Temperatures rose by more than 1.5°F from 1900 to 2010, but between 1980 and 2010 the rate of
warming was three times faster than from 1900 through 2010. Precipitation generally increased
over the last century, with much of the increase driven by intensification of the heaviest rainfalls.
Several types of extreme weather events in the Midwest (e.g., heat waves and flooding) have
already increased in frequency and/or intensity due to climate change. In the future, if emissions
continue increasing, the Midwest is expected to experience 5.6 to 8.5°F of warming by the
2080s, leading to more heat waves. Specific vulnerabilities highlighted by the NCA include
long-term decreases in agricultural productivity, changes in the composition of the region's
forests, increased public health threats from heat waves and degraded air and water quality,
negative impacts on transportation and other infrastructure associated with extreme rainfall
events and flooding, and risks to the Great Lakes including shifts in invasive species, increases in
harmful algal blooms, and declining beach health.

   High temperatures (more than 100°F in the Southern Plains and more than 95°F in the
Northern Plains) are projected to occur much more frequently by mid-century. Increases in
extreme heat will increase heat stress for residents, energy demand for air conditioning, and
water losses. In Hawaii, other Pacific islands, and the Caribbean, rising air and ocean
temperatures, shifting rainfall patterns, changing frequencies  and intensities of storms and
drought,  decreasing base flow in streams, rising sea levels, and changing ocean chemistry will
affect ecosystems on land and in the oceans, as well as local communities, livelihoods, and
cultures.  Low islands are particularly at risk.

   In Hawaii and the Pacific islands, "Warmer oceans are leading to increased coral bleaching
events and disease outbreaks in coral reefs, as well as changed distribution patterns of tuna
fisheries.  Ocean acidification will reduce coral growth and health. Warming  and acidification,
combined with existing stresses, will strongly affect coral reef fish communities." For Hawaii
and the Pacific islands, future sea surface temperatures are projected to increase 2.3°F by 2055
and 4.7°F by 2090 under a scenario that assumes continued increases in emissions.

1.5.1.3 Recent U.S. Commitments on Climate Change Mitigation

   In 2009, President Obama adopted a goal of reducing U.S. GHG emissions by approximately
17 percent below 2005 levels by 2020.30 The Administration subsequently took several major
actions towards this goal under its Climate Action Plan, most notably the historic National
Program standards to reduce new car and light truck GHG emissions levels by 50 percent by
2025 (see above for the history of the National Program),  promulgating the first standards to
reduce GHGs and improve fuel efficiency for medium- and heavy-duty vehicles for model years
2014-2018 (Phase 1) and proposing further Phase 2 standards for this segment, the investment of
more than $80 billion in clean energy  technologies under the  economic recovery program,
implementing various energy efficiency measures,  and promulgating the Clean Power Plan (i.e.
the standards of performance for new  and existing electric power plant stationary sources under
sections 111 (b) and (d) of the Clean Air Act) to reduce CCh emissions from the electric power
sector.

   In December 2015,  the U.S. was one of over 190 signatories to the Paris Climate Agreement,
widely regarded as the most ambitious climate change agreement in history.  In the Paris
agreement, individual countries agreed to commit to putting forward successive and ambitious
nationally determined contributions (NDCs) for greenhouse gas emissions reductions to the
                                              1-19

-------
                                                                            Introduction
United Nations Framework Convention on Climate Change. Further, the countries agreed to
revise their NDCs every five years, with the expectation that they will strengthen over time.  The
Paris agreement reaffirms the goal of limiting global temperature increase to well below 2°Fs
Celsius, and for the first time urged efforts to limit the temperature increase to 1.5°Fs Celsius.
The U.S. submitted a non-binding intended NDC target of reducing economy-wide GHG
emissions by 26-28 percent below its 2005 level in 2025 and to make best efforts to reduce
emissions by 28 percent.31 This pace would keep the U.S. on a trajectory to achieve deep
economy-wide reductions on the order of 80 percent by 2050.

1.5.1.4 Recent California Commitments on Climate Change

   With climate change threatening California's resources, economy, and quality of life, the
State is squarely focused on addressing it and protecting our natural and built environments.
Over the past several decades, California has taken a number of innovative actions to cut
emissions from the transportation sector.  Collectively, the State's set of vehicle, fuels, and land
use policies will cut in half emissions from passenger transportation and drivers' fuel costs over
the next 20 years. California's Low Carbon Fuel Standard (LCFS) is beginning to drive the
production of a broad array of cleaner fuels.  Since its launch in 2011, the regulation has
generated a multitude of unique approaches for cleaner fuels. The cars on California's roads are
also undergoing a transformation. California's vehicle GHG standards-authorized by AB  1493
(Pavley) in 2002, first approved in 2004, and extended in 2012- are delivering both carbon
dioxide reductions and savings at the pump. The transition to a fleet of lower-emitting, more-
efficient vehicles in  California will continue beyond 2020, as these rules cover model years
through 2025,  and turnover of the fleet will deliver additional benefits from these rules for many
more years. California (CARB) is also working with EPA and NHTSA on national GHG
standards and corresponding fuel  efficiency standards for medium- and heavy-duty trucks.
Furthermore, California is making major strides toward reducing the number of miles people
drive, through more sustainable local and regional housing, land use, and transportation
planning. However, California has recognized these actions will not be sufficient to address
deep GHG emission reductions. To begin laying the foundation for further actions, the Governor
issued an Executive Order in 2015 establishing new  2030 targets and a revised statewide climate
plan is being developed this year. The Governor's 2030 targets include a 40 percent reduction in
GHG emissions below 1990 levels, a 50 percent renewable portfolio standard for electricity (now
established as law with legislation in late 2015), and a 50 percent reduction in petroleum usage
from the state's cars and trucks.

   Additionally, reducing emissions  of short-lived climate pollutants (SLCPs), such as black
carbon (BC), CH4, and some fluorinated gases (such as a number of hydrofluoroethers and
hydrofluorocarbons) may help slow the near-term rate of climate change.  This may be
particularly important in regions such as the Arctic, where the climate is changing most rapidly,
and where BC has additional impacts due to its ability to darken snow and ice. The majority of
BC emissions come from  mobile sources (predominantly diesel) and open biomass burning.  In
April 2016, California released a Proposed SLCP Reduction Strategy which is designed to meet
planning targets of reducing CH4  and HFC emissions by 40 percent below 2013 levels by 2030,
and reducing BC emissions by 50 percent below 2013 levels by 2030.

1.5.1.5 Contribution of Cars and Light Trucks to the U.S. Greenhouse Gas Emissions
Inventory
                                              1-20

-------
                                                                           Introduction
   The most recent U.S. GHG emission inventory32 includes seven greenhouse gases: carbon
dioxide (CCh), methane (CH4), nitrous oxide (N2O), perfluorocarbons (PFCs),
hydrofluorocarbons (HFCs), sulfur hexafluoride (SFe), and nitrogen trifluoride (NFs).

   Mobile sources, which include cars, light trucks and medium-duty passenger vehicles (the
largest sport utility vehicles and full-size passenger vans), heavy-duty trucks and buses,
airplanes, railroads, marine vessels, and a variety of smaller sources, are significant contributors
of four of the seven GHGs listed above. CCh, CFLt, and N2O emissions are present in vehicle
tailpipe emissions, and HFCs are used in automotive air conditioning systems. In recent years,
the annual GHG emissions inventory due to light-duty vehicles has been slightly more than 1
billion metric tons per year. Currently, HFCs are a small fraction of the total climate forcing
emissions, but they are the fastest growing source of GHG emissions in California. Across the
US, emissions of HFCs are increasing more quickly than those of any other GHGs, and globally
they are increasing 10-15 percent annually.33 At that rate, emissions are projected to double by
2020 and triple by 2030.34  The growth is driven both by increased demand for refrigeration and
air-conditioning, especially for stationary applications, and because these substances were
developed and are being implemented as alternatives to ozone-depleting substances (ODS) under
the Montreal Protocol.35'36

   In 2013, mobile sources emitted 30 percent of all U.S. GHG emissions, the second largest
contribution after  power plants. Transportation sources, which are largely synonymous with
mobile sources but which exclude certain off-highway sources such as farm and construction
equipment, account for 27 percent of U.S. GHG emissions. Motor vehicles alone, which include
cars, light trucks and medium-duty passenger vehicles, heavy-duty trucks and buses, and
motorcycles, are responsible for 23 percent of U.S. GHG emissions.  CCh emissions represent 96
percent of total mobile source GHG emissions.

   Cars, light trucks, and medium-duty passenger vehicles, the motor vehicles covered by the
Light-Duty GHG/CAFE National Program, alone account for 16 percent of all U.S. GHG
emissions.

1.5.1.6 Importance of the National Program in the U.S. Climate Change Program

   The Light-Duty GHG/CAFE National  Program is a centerpiece of the U.S. climate change
program. The GHG standards that took effect with model year 2012 cars, light trucks, and
medium-duty passenger vehicles, promulgated  under the Clean Air Act, were the first-ever
national GHG emissions standards in the U.S.

   The Light-Duty GHG/CAFE National  Program is projected to achieve very large GHG
emissions reductions. In the analysis for the 2025 rulemaking, EPA projected that the
cumulative GHG emissions savings over the lifetimes of the new light duty vehicles sold in
model years 2012 through 2025 would be 6 billion metric tons (these reductions would begin in
calendar year 2012 and would end in the calendar year when the last model year 2025 vehicles
would be retired from the fleet).37

   Because EPA GHG emissions standards will remain in effect unless and until they are
changed, GHG emissions savings will continue to accrue for vehicles sold after model year 2025,
and these longer-term GHG emissions (CChe) savings are not reflected in the 6 billion metric ton
value above.  In terms of on-the-ground reductions in specific calendar years, EPA projected, in
                                              1-21

-------
                                                                           Introduction
the 2012 Final Rule analysis, that the National Program would yield GHG (CChe) emissions
reductions of 180 million metric tons (MMT) in calendar year 2020, 380 MMT in 2025, 580
MMT in 2030, 860 MMT in 2040, and 1100 MMT in calendar year 2050.  The cumulative GHG
emissions savings over calendar years 2012 through 2050 were projected to be 22 billion metric
tons.38

   Comparing GHG emissions reductions across various countries and policies is complicated,
involving many assumptions in order to yield "apples-to-apples" comparisons. In 2014, The
Economist published a comparison of global programs that yielded large GHG emissions
reductions.39 In terms of annual GHG emissions reductions, the article concluded that the Light-
Duty GHG/CAFE National Program yielded the sixth-greatest rate of GHG emissions reductions
among all of the programs evaluated, worldwide.

1.5.2   Petroleum Consumption and Energy Security

1.5.2.1 Overview of Petroleum Consumption and Energy Security

   In 1975, Congress enacted the Energy Policy and Conservation Act (EPCA) mandating that
NHTSA establish and implement a regulatory program for motor vehicle fuel economy to
address "the need of the United States to conserve energy." While the U.S. has plentiful
resources for most energy feedstocks, the one source of energy for which the U.S. has been
dependent upon imports for many decades is petroleum.  Accordingly, NHTSA concluded that
the EPCA goal of "the need of the United States to conserve energy" means "the consumer cost,
national balance of payments, environmental, and foreign policy implications of our need for
large quantities of petroleum, especially imported petroleum."40 NHTSA first implemented the
corporate average fuel  economy (CAFE) program in  1978.  Congress reaffirmed the CAFE
program with the Energy Independence and Security Act (EISA) of 2007.

   Dependence on imported petroleum leads to many risks: the potential for oil suppliers to
manipulate market mechanisms and thereby raise prices, the threat of supply disruptions which
can have significant economic and national security ramifications, and the export of domestic
capital to pay for imported petroleum which can have a wide variety of deleterious impacts on
domestic economic growth and trade balances. For these reasons,  reducing excessive reliance on
imported oil has been a national priority since the first oil embargo in  1973-1974. Despite these
concerns, net imports of petroleum grew fairly consistently for three decades from around 5
million barrels per day (MBPD) in the early 1970s to over 12 MBPD in 2004-2007, and the
import share of U.S. oil consumption over the same period doubled from about 30 percent to
about 60 percent.41 The direct costs of U.S. net oil imports fluctuate with world oil  prices, of
course, ranging in this  century from a little over $100 billion in 2000 to an all-time high of nearly
$400 billion in 2008.42 The U.S. reliance on imported petroleum has decreased significantly in
recent years as domestic oil and natural gas liquids production reversed its historical decline and
increased from 6.8 MBPD in 2008 to 11.7 MBPD in 2014, at a time when  total domestic
petroleum demand decreased slightly.43 Accordingly, net oil imports have declined from a peak
of over 12 MBPD a decade ago to 5.0 MBPD in 2014, representing 27 percent of total U.S. oil
consumption, with the  latter value similar to that in the early 1970s.44

   While oil imports have declined in recent years, oil prices rose from $15-30 per barrel in the
late 1980s through the  early 2000s to $50-100 per barrel since, which yields national average
gasoline prices of $2.50 to $4.00 per gallon. Accordingly, while payments for imported oil have
                                             1-22

-------
                                                                            Introduction
decreased, payments for total U.S. oil consumption remained at about $600 billion in 2014.
These higher oil prices have yielded national average gasoline prices on the order of $3-4 per
gallon over much of the last few years, which significantly increased the cost-of-living for
American families.  Gasoline prices have fallen since late 2014 and averaged about $2.50 per
gallon during most of 2015. As of February 2016, the Short-Term Energy Outlook from EIA
forecasts the U.S. retail regular gasoline price to average $1.98 per gallon in 2016 and $2.21 per
gallon in 2017.45 U.S. drivers have benefited considerably from these low prices. Nevertheless,
DOT must set fuel economy standards considering estimates of future fuel prices.

   The history of the oil market over the last few decades has been longer periods of relative
stability interrupted by shorter periods of high market volatility.  Oil prices dropped significantly
in late 2014, and so U.S. payments for both imported oil and total oil are lower today than in the
recent past. The Energy Information Administration's AEO 2015 projected a wide range of
possible oil prices out to 2040, ranging from a low of $76 per barrel under its Low Oil Price
scenario to a high of $252 per barrel in its High Oil Price scenario, with a reference case price of
$141 per barrel  (all Brent Spot Prices in 2013 dollars).46 The uncertainty and volatility
associated with  world oil prices are another risk associated with our dependence on petroleum.

1.5.2.2 Recent U.S.  Commitments on Petroleum and Energy Security

   Dependence  on imported oil has been identified as an important challenge since the first oil
embargo in 1973-74.

   On March 30, 2011, the U.S. pledged to reduce oil imports by one-third by 2025, or by about
3.6 MBPD.47 The long-term strategy advanced for achieving this historic reduction in oil
imports included several elements: fuel economy/GHG  standards for both light-duty and heavy-
duty vehicles, expanding domestic oil development, and developing alternative fuels.  Due to a
combination of factors, primarily increased domestic oil production, but also higher oil prices
and the first few years of the CAFE/GHG standards, the one-third reduction in oil imports, or 3.6
MBPD, has already been achieved well in advance of 2025.  The broader challenge will be to
retain, or even build on, this successful reduction in oil imports over the next decade given the
history of volatility in oil markets.

1.5.2.3 Contribution of Cars and Light Trucks to U.S.  Petroleum Consumption

   In 2014, transportation sources accounted for 70 percent of U.S. petroleum consumption.
Cars, light trucks, and medium-duty passenger vehicles, the  motor vehicles covered by the
National  Program, account for about 60 percent of all U.S. transportation oil consumption, about
8 million barrels per day, or about 42 percent of total U.S. petroleum consumption.48

1.5.2.4 Importance of National Program to Petroleum Consumption and Energy Security

   The CAFE standards have long been regarded as a major reason for the significant increase in
average light vehicle fuel economy from the late 1970s  through the mid-1980s, and therefore
reduced petroleum consumption and improved energy security relative to what would likely have
been the case without the CAFE standards.49  While the CAFE standards were relatively
unchanged from the mid-1980s through the mid-2000s,  the standards began to be raised for
MY2005 for light trucks and then for both cars and light trucks in MY2011.50  The National
Program, which covers new cars, light trucks, and medium-duty passenger vehicles beginning in
MY2012, represent the most significant increases in fuel economy standards in over 30 years.
                                              1-23

-------
                                                                          Introduction
The projected oil savings from the Light-Duty GHG/CAFE National Program are very
significant. Fuel economy improvements under U.S. CAFE standards have already helped the
Nation to reduce its fuel consumption by more than a trillion gallons of fuel.  New standards
have the potential to help the Nation to reduce its fuel consumption by a similar amount between
now and 2050.

   These very large reductions in fuel consumption should dampen world oil prices (see further
discussion in Chapter 10) which would further increase consumer fuel savings that are not
directly included in our projections.
                                             1-24

-------
                                                                                       Introduction
References
1 75 FR 25324, May 7, 2010.
2 77 FR 62624, October 15, 2012.
3See40CFR86.1818-12(h).
4 See 40 CFR 86.1818-12(h)(2)(i).
5 77 FR 62784, October 15, 2012.
6 77 FR 62639, October 15, 2012.
749U.S.C. 32902.
8 Center for Biological Diversity v. NHTSA, 508 F. 3d 508 (9th Cir. 2007).
9 549 U.S. 497 (2007).
10 68 FR 52922 (Sept. 8, 2003).
11 549 U.S. at 531-32.
12 For further information on Massachusetts v. EPA see the July 30, 2008 Advance Notice of Proposed Rulemaking,
"Regulating Greenhouse Gas Emissions under the Clean Air Act", 73 FR 44354 at 44397. There is a
comprehensive discussion of the litigation's history, the Supreme Court's findings, and subsequent actions
undertaken by the Bush Administration and the EPA from 2007-2008 in response to the Supreme Court remand.
Also see 74 FR 18886, at 1888-90 (April 24, 2009).
13 75 FR 25324, May 7, 2010.
14 77 FR, 62624, October 15, 2012.
15 Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2013, EPA Publication number EPA 430-R-15-
004, April 15, 2015.
16 49 CFR 523.2.
17 See 49 CFR 531.5 and 49 CFR 533.5 for the CAFE standards for passenger cars and light trucks, respectively, and
40 CFR 86.1818-12 for the GHG standards.
18 78 FR 2112, January 9, 2013.
19http://www.arb.ca.gov/msprog/consumer_info/advanced_clean_cars/consumer_acc.htm.
20 National Research Council (NRC), Climate Stabilization Targets, p.3.
21 "Endangerment and Cause or Contribute Findings for Greenhouse Gases Under Section 202(a) of the Clean Air
Act," 74 FR 66496 (Dec. 15, 2009) ("Endangerment Finding").
22 USGCRP, Third National Climate Assessment, p. 221.
23 See also Kleeman, M. J., S.H. Chen, and R. A. Harley 2010 Climate change impact on air quality in California:
Report to the California Air Resources Board http://www.arb.ca.gov/research/apr/past/04-349.pdf
4 National Research Council, Understanding Earth's Deep Past, p. 138.
5ftp://aftp.cmdl.noaa.gov/products/trends/co2/co2_annmean_mlo.txt.
6 http://www.esrl.noaa.gov/gmd/ccgg/trends/.
7 Blunden, J., and D.  S. Arndt, Eds., 2015: State of the Climate in 2014. Bull. Amer. Meteor. Soc., 96  (7), S1-S267.
8 http://www.ncdc.noaa.gov/sotc/global/201513.
9 NRC, 2011: America's Climate Choices, The National Academies Press, p. 2.
30 Fact Sheet: U.S.-China Joint Announcement on Climate Change and Clean Energy Cooperation, The White
House, Office of the Press Secretary, November 11, 2014, available at http://www.whitehouse.gov/tahe-press-
office/2014/11/11/fact-sheet-us-china.
31 United States of America, Intended Nationally Determined Contribution, March 31, 2015,
http://www4.unfccc.int/submissions/INDC/Published%20Documents/United%20States%20of%20America/l/U.S.%
20Cover%20Note%20INDC%20and%20Accompanying%20Information.pdf.
32 Inventory of U.S. Greenhouse Gas Emissions and Sinks, 1990-2013, U.S. Environmental Protection  Agency,
2015, available at epa.gov/climatechange/ghgemissions/usinventoryreport.html.
33 UNEP 2011. HFCs: A Critical Link in Protecting Climate and the Ozone Layer. United Nations Environment
Programme.
34 Akerman, Nancy H., Hydrofluorocarbons and Climate Change: Summaries of Recent Scientific and Papers, 2013.
35 Rigby, M., R. G. Prinn, S. O'Doherty, B. R. Miller, D. Ivy, J. Mtihle, C. M. Harth, P. K. Salameh, T. Arnold, R. F.
Weiss, P. B. Krummel, L. P. Steele, P. J. Fraser, D. Young, and P. G. Simmonds, Recent and future trends in
synthetic greenhouse gas radiative forcing, Geophys. Res. Lett, 41, 2623-2630, doi: 10.1002/2013GL059099, 2014.
                                                     1-25

-------
                                                                                       Introduction
36 Velders, G. J. M., A. R. Ravishankara, M. K. Miller, M. J. Molina, J. Alcamo, J. S. Daniel, D. W. Fahey, S. A.
Montzka, and S. Reimann (2012), Preserving Montreal Protocol climate benefits by limiting HFCs., Science,
335(6071), 922-923, doi:10.1126/science.l216414. [Available at http://www.ncbi.nlm.nih.gov/pubmed/22362993.
37 Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average
Fuel Economy  Standards, Regulatory Impact Analysis, U.S. Environmental Protection Agency, EPA-420-R-12-016,
August 2012, page 7-32.
38 Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average
Fuel Economy  Standards, Regulatory Impact Analysis, U.S. Environmental Protection Agency, EPA-420-R-12-016,
August 2012, page 7-35.
39 The Deepest Cuts: Our guide to the actions that have done the most to slow global warming, The Economist,
September 20, 2014.
40 42 FR 63184, 63188 (1977).
41 Monthly Energy Review, Department of Energy, Energy Information Administration, May 2015.
42 EPA projections based on world oil price and net oil import data from Department of Energy.
43 Monthly Energy Review, Department of Energy, Energy Information Administration, June 2015.
44 Monthly Energy Review, Department of Energy, Energy Information Administration, May 2015.
45  Short-Term Energy Outlook (STEO), February 2016, available at http://www.eia.gov/forecasts/steo.
46 Annual Energy Outlook 2015, Executive Summary, U.S. Energy Information Administration, April 14, 2015.
47 Blueprint for a Secure Energy Future, The White House Office of the Press Secretary, March 30, 2011.
48 Transportation Energy Data Book: Edition 34, Oak Ridge National Laboratory, August 2015, available at
cta.ornl.gov/data.
49 Effectiveness and Impact of Corporate Average Fuel Economy Standards, National Research Council/National
Academy of Sciences, 2002.
50 Summary of Fuel Economy Performance, U.S. Department of Transportation, National Highway Traffic Safety
Administration, June 26, 2014.
                                                     1-26

-------
                          Overview of the Agencies' Approach to the Draft TAR Analysis

Table of Contents

Chapter 2:  Overview of the Agencies' Approach to the Draft TAR Analysis	2-1
  2.1    Factors Considered in this Report	2-1
  2.2    Gathering Updated Information since the 2012 Final Rule	2-2
    2.2.1   Research Projects Initiated by the Agencies	2-2
    2.2.2   Input from Stakeholders	2-6
       2.2.2.1   Automobile Manufacturers	2-6
       2.2.2.2   Automotive Suppliers	2-6
       2.2.2.3   Environmental Non-governmental Organizations (NGOs) and Consumer Groups
               2-7
       2.2.2.4   State and Local Governments	2-7
    2.2.3   Other Key Data Sources	2-8
  2.3    Agencies' Approach to Independent GHG and CAFE Analyses	2-9

-------
                           Overview of the Agencies' Approach to the Draft TAR Analysis

Chapter 2:  Overview of the Agencies' Approach to the Draft TAR Analysis

2.1    Factors Considered in this Report

   The Midterm Evaluation (MTE) is a comprehensive assessment of all of the factors
considered by the agencies in setting the MY 2022-2025 standards.  The 2017-2025 final rule
(FRM) preamble stated that "bothNHTSA and EPA will develop and compile up-to-date
information for the midterm evaluation through a collaborative, robust and transparent process,
including public notice and comment. The evaluation will be based on (1) a holistic assessment
of all the factors considered by the agencies in setting standards, including those set forth in this
final rule and other relevant factors and (2) the expected impact of those factors on the
manufacturers' ability to comply, without placing decisive weight on any particular factor or
projection"^

   The 2017-2025 final rule preamble further provided an outline of what the agencies would
consider in the Draft TAR, stating that the  "TAR will examine the same issues and underlying
analyses and projections considered in the original rulemaking, including technical and other
analyses and projections relevant to each agency's authority to set standards as well as any
relevant new issues that may present themselves. "B For EPA's part, the EPA regulations state
that  in making the determination required, the Administrator "shall consider the information
available on the factors relevant to setting greenhouse gas emission standards under section
202(a) of the Clean Air Act for model years 2022 through 2025, including but not limited to:

       •  The availability and effectiveness of technology, and the appropriate lead time for
          introduction of technology;
       •  The cost on the producers or purchasers of new motor vehicles or new motor vehicle
          engines;
       •  The feasibility and practicability of the standards;
       •  The impact of the standards on reduction of emissions, oil conservation, energy
          security, and fuel savings by consumers;
       •  The impact of the standards on the automobile industry;
       •  The impacts of the standards on automobile safety;
       •  The impact of the greenhouse gas emission standards on the Corporate Average Fuel
          Economy standards and a national harmonized program; and
       •  The impact of the standards on other relevant factors. "c
   The preamble to the final rule further listed ten  relevant factors that  the agencies will consider
at a minimum during the MTE.D These factors are:

       •  Development of powertrain improvements to gasoline and diesel powered vehicles
          (Chapter 5)
       •  Impacts on employment, including the  auto sector (Chapter 7)
A 77 FR 62652, October 15, 2012.
B 77 FR 62784, October 15, 2012.
c40CFR86.1818-12(h).
D 77 FR 62784, October 15, 2012.
                                              2-1

-------
                           Overview of the Agencies' Approach to the Draft TAR Analysis

       •  Availability and implementation of methods to reduce weight, including any impacts
          on safety (Chapter 5 and 8)
       •  Actual and projected availability of public and private charging infrastructure for
          electric vehicles, and fueling infrastructure for alternative fueled vehicles (Chapter 9)
       •  Costs, availability, and consumer acceptance of technologies to ensure compliance
          with the standards, such as vehicle batteries and power electronics, mass reduction,
          and anticipated trends in these costs (Chapters 5, 6, 12, and 13)
       •  Payback periods for any incremental vehicle costs associated with meeting the
          standards (Chapter 12 and 13)
       •  Costs for gasoline, diesel  fuel, and alternative fuels (Chapters 10, 12 and 13)
       •  Total light-duty vehicle sales and projected fleet mix (Chapter 4)
       •  Market penetration across the fleet of fuel efficient technologies (Chapter 3, 4, 12,
          and 13)
       •  Any other factors that may be deemed relevant to the review
   Each of the factors listed above is addressed in this Draft TAR, primarily in the chapters
indicated above. Among the other factors deemed relevant, EPA's analysis for the Draft TAR
examines the potential impact of the  California Zero Emission Vehicle (ZEV) program which
California has revised since the final rule (Chapter 4) and both EPA and NHTSA also examined
the availability and use of credits, including credits for emission reductions from air conditioning
improvements and off-cycle technologies (Chapters 5 and 11).

2.2    Gathering Updated Information since the 2012 Final Rule

   The agencies' goal is that the midterm evaluation will be conducted through a collaborative,
data-driven, and transparent process. In gathering data and information for this Draft TAR, the
agencies pulled from a wide range of sources.  These sources included research projects initiated
by the agencies, input from stakeholders, and information from technical conferences, published
literature, and studies published by various organizations. Each of these sources is described
further below.  The agencies  will continue to gather and evaluate more up-to-date information to
inform our analyses as we  move forward with our respective actions.

2.2.1   Research Projects Initiated by the Agencies

   EPA, NHTSA, and CARB have each initiated new research since the 2012 final rule to inform
the MTE.  This research has been coordinated across the three agencies and, where possible,
each agency has made the results of a variety of projects available to the public (e.g., through
published papers, presentations at public forums and on agency web sites).E This section
summarizes each agency's  research projects in more detail.

   EPA has research projects underway in a wide range of areas. Through the National Vehicle
and Fuel Emissions Laboratory (NVFEL) in Ann Arbor, Michigan, starting in 2013 EPA has
been conducting a major research benchmarking program for advanced engine and transmission
technologies. To date, more  than 20 currently  available production vehicles have been tested to
E For EPA projects, see Msi/f^MS^^Ei^SQyJMiS/^lMM^IlMlsMM, for NHTSA projects, see
  iMJBl^wwOll]^
  http://www.nhtsa.gov/Laws+&+Regulations/CAFE+-+Fuel+Economy/nhtsa-epa-carb-workshop-03012016.
                                              2-2

-------
                          Overview of the Agencies' Approach to the Draft TAR Analysis

assess their engine and/or transmission efficiencies. These data provide inputs and validation for
EPA's vehicle simulation model, the Advanced Light-Duty Powertrain and Hybrid Analysis
(ALPHA) model (described further in Chapter 5.3). Thus far, EPA has published more than 15
papers for SAE International describing various aspects of the benchmarking program and
ALPHA model validation work.1

   EPA has continued studies of the costs of fuel economy technologies through state-of-the art
cost teardown studies with the engineering firm FEV.  EPA has built upon the cost teardown
work supporting the FRM with new technologies, including mild hybrid systems, advanced
boosted engines, naturally aspirated high compression ratio engines, and diesel engines.2 In
addition, the previous teardown studies have been updated to reflect current costs. In other
research related to the costs of the program,  EPA commissioned a literature review of the effects
of manufacturer "learning by doing."3

   EPA has built upon previous studies of mass reduction feasibility and costs with the addition
of a new study examining the mass reduction potential of full-size light-duty pickup trucks. This
study builds upon the mass reduction studies done previously by EPA and NHTSA, respectively,
for a mid-size crossover vehicle and mid-size sedan.

   EPA has initiated research on consumer issues, including a project exploring automotive
reviews of fuel economy technologies,4 an assessment of consumer satisfaction of new vehicle
purchases, a review of literature on consumers' willingness to pay for vehicle attributes,  and an
updated assessment of vehicle affordability that examines potential impacts on low-income
households, low-priced vehicle segments, and the automotive loan market.5

   In continuing to explore economic impacts of the standards,  EPA has completed new research
on the vehicle miles travelled (VMT) rebound effect6, and is currently conducting a literature
review of the research on the light-duty vehicle VMT rebound effect.

   Finally, EPA has continued the development of modeling tools, including the ALPHA full
vehicle simulation model,7 the Optimization Model for reducing Emissions of Greenhouse gases
from Automobiles (OMEGA), and the Lumped Parameter Model (LPM) for assessing vehicle
technology package efficiencies. EPA also has continued to explore the potential use of
consumer choice modeling by attempting to validate EPA's current working model with actual
market impacts.8

   NHTSA has also sponsored new studies and research to inform the midterm evaluation:

   The National Academy of Sciences (NAS) has long had a role in helping to inform NHTSA
on issues related to fuel economy. Section 107 of EISA 2007 instructed NHTSA to contract with
the NAS to "develop a report evaluating vehicle fuel economy standards, including an
assessment of automotive technologies and costs to reflect developments since the [NAS]'s 2002
report (NAS 2002) evaluating the corporate  average fuel economy standards was conducted and
an assessment of how such technologies may be used to meet the new fuel economy standards."
Section 107 also noted that the report should be updated at 5-year intervals through 2025.
                                             2-3

-------
                           Overview of the Agencies' Approach to the Draft TAR Analysis

   In 2011, the first such report in response to this mandate was released, "Assessment of Fuel
Economy Technologies for Light-Duty Vehicles" (NAS 2011).F This is referred to as the Phase
1 report, which examined categories of near-term technologies important for reducing fuel
consumption, their costs, issues associated with estimating costs and price impacts of these
technologies, and approaches for estimating the fuel consumption benefits from combinations of
these technologies.

   In 2015, NAS issued the second report (NAS 2015) in this series titled "Cost, Effectiveness
and Deployment of Fuel Economy Technologies for Light-Duty Vehicles."0 The Phase 2 report
was purposely timed to inform the mid-term evaluation by considering technologies applicable in
the 2020 to 2030 timeframe. In particular, the committee was asked to include the following in
its assessment:

       •  Methodologies and programs used to develop  standards for passenger cars and light
          trucks under current and proposed CAFE programs;
       •  Potential for reducing mass by up to 20 percent, including materials substitution and
          downsizing of existing vehicle designs, systems or components;
       •  Other vehicle technologies whose benefits may not be captured fully through the
          federal test procedure, including aerodynamic drag reduction and improved efficiency
          of accessories;
       •  Electric powertrain technologies, including the capabilities of hybrids, plug-in
          hybrids, battery electric vehicles, and fuel cell vehicles;
       •  Advanced gasoline and diesel engine technologies that will increase fuel economy;
       •  Assumptions, concepts, and methods used in estimating the costs of fuel economy
          improvements, including the degree to which time-based cost learning for well-
          developed existing technologies and/or volume-based cost learning for newer
          technologies  should apply, and the differences between Retail Price Equivalent and
          Indirect Cost Multipliers;
       •  Analysis of how fuel economy technologies may be practically integrated into
          automotive manufacturing processes and how such technologies are likely to be
          applied;
       •  Costs and benefits in vehicle value that could accompany the introduction of
          advanced vehicle technologies;
       •  Test procedures and calculations used to determine fuel economy values for purposes
          of determining compliance with CAFE standards; and,
       •  Assessment of the consumer impacts of factors that may affect changes in vehicle
          use.
   The overall report estimates the cost, potential efficiency improvements, and barriers to
commercial deployment of technologies that might be employed from 2020 to 2030. The report
describes these promising technologies and makes recommendations for their inclusion on the
list of technologies applicable for the 2022-2025 CAFE standards.
F Available at http://www.nap.edu/catalog/12924/assessment-of-fuel-economv-technologies-for-light-dutv-vehicles
  (last accessed Feb. 26, 2016).
G Available at http://www.nap.edu/catalog/21744/cost-effectiveness-and-deplovment-of-fuel-economv-technologies-
  for-light-dutv-vehicles (last accessed Feb. 26, 2016).
                                              2-4

-------
                           Overview of the Agencies' Approach to the Draft TAR Analysis

   NHTSA has funded new work at Argonne National Laboratory (ANL) to conduct large-scale
simulation using DOE's Autonomie vehicle simulation tool to estimate the effects of
combinations of technologies on fuel economy.  Simulation of feasible technology combinations
will yield databases that are flexible, account for all technology interactions, and can be fed
directly into the CAFE model, which NHTSA uses for fleet-level analysis.  Numerous
presentations and papers on the new work have been presented at conferences.9'10'11'12'13'14

   NHTSA conducted a mass reduction and feasibility cost study on a passenger car to determine
the maximum feasible weight reduction while maintaining the same vehicle functionalities, such
as performance, safety, crash rating etc., as the baseline vehicle. Furthermore, another objective
was to maintain retail price of the light-weighted vehicle(s) within +10 percent of the original
vehicle.  The original report, cost, Computer-Aided Engineering (CAE) models, and peer review
report are all publicly available on the NHTSA website.15'16'17 The mass reduction study is
discussed in detail in Chapter 8.

   NHTSA has funded a similar mass reduction feasibility and cost study for a full-size pickup
(MY 2014 Chevrolet Silverado) that is ongoing. A related study is ongoing on the production
costs of changing vehicle attributes (e.g., track width, wheelbase) and determining the effect of
these changes on other vehicle characteristics that affect fuel economy.

   The FRM also relied on statistical analysis of historical crash data to assess the effects of
vehicle mass reduction and size on  safety.18  In addition, Volpe is working to update a 2012
report on the relationship between vehicle mass (represented as curb weight) and societal fatality
risk.19 The updated analysis incorporates data from multiple sources required to represent
fatalities, baseline driving risk (i.e., induced exposure), and VMT across distributions of driver-,
crash- and vehicle-specific factors.  The primary sources applied within the analysis are: the
Fatality Analysis Reporting System (FARS), State crash records, R.L. Folk's National Vehicle
Population Profiles (NVPP) and odometer readings, and a range of sources of values for curb
weight, footprint, track width, wheelbase and other vehicle attributes.

   Certain studies used to inform the 2012 final rule continue to inform the safety analysis for
the Draft TAR:

       •  Systems modeling to assess the effects of future lightweight vehicle designs on
          overall fleet safety.  The approach includes estimating the real-world level of safety in
          a vehicle for its own occupants (self-protection) and for the occupants in vehicles
          with which it collides (partner protection.20
   Fuel economy and GHG emissions standards benefit society by reducing fuel and emissions
resulting from the operation of motor vehicles, so estimates of the extent to which vehicles will
be driven annually are central to the agencies' evaluation of the benefits of new standards. Based
on an analysis of more than 70 million odometer readings reported  by IHS Automotive (formerly
R.L. Polk), NHTSA has developed updated estimates of annual mileage accumulation over
vehicles' useful lives. We note that there are many factors that influence how much  people drive
aside from fuel efficiency.

   CARB has also undertaken research since the 2012 rule was finalized. To meet fuel economy
and greenhouse gas standards, it is expected that the vast majority of reductions will come from
improvements to the vehicle powertrain—specifically, the engine and the transmission.
                                              2-5

-------
                           Overview of the Agencies' Approach to the Draft TAR Analysis

However, there are other improvements that can increase efficiency and the agencies did assume
some reductions from these areas. Notably, items like vehicle aerodynamics, low rolling
resistance tires, and making vehicles lighter can have an appreciable contribution by making it
easier for the vehicle to overcome resistance from wind and road friction, and thus go farther on
the same amount of fuel.  To better understand some of the possibilities for these other
technologies, CARB  commissioned a study with Novation Analytics (formerly known as
ControlTec). The study analyzed all available vehicles in the 2014 model year, identified the
better performers in class-specific road load characteristics,  and then upgraded the entire vehicle
fleet to nominally have best-in-class aerodynamics, tire rolling resistance, and mass efficiency.
The road load reduction study is discussed in further detail in Appendix A.

2.2.2  Input from Stakeholders

   In developing this Draft TAR, the agencies gathered input,  data, and information from a wide
range of stakeholders. The agencies conducted outreach with numerous stakeholders, including
auto manufacturers, automotive suppliers, environmental and other non-governmental
organizations (NGOs), consumer groups, labor unions, automobile dealers, state and local
governments, fuels and energy providers, and others. Below we characterize the nature of the
dialogs conducted with various stakeholders and the kinds of information shared with the
agencies.

2.2.2.1 Automobile Manufacturers

   The agencies met with nearly all automobile manufacturers individually as well as through
their trade associations on numerous occasions. We met with automakers including BMW, Fiat-
Chrysler, Ford, General Motors, HondaH Jaguar Land Rover, Mazda, Mercedes-Benz, Nissan,
Porsche, Subaru, Tesla, Toyota, Volkswagen, and Volvo. Individually, each auto manufacturer
generally provided the agencies with information on the company's overall strategy for meeting
the 2022-2025 GHG/CAFE standards, the technologies and  products they planned to bring to
market and the sequence of that product plan, input on the effectiveness, costs, and
implementation of those technologies, and challenges in meeting the standards.  Several
companies also provided feedback on credit provisions contained in the existing GHG  and CAFE
programs, and offered ideas on additional  flexibilities that the  companies believed could ease
implementation of the program. By its nature, most of the information provided to the agencies
was claimed to be confidential business information.

   The automobile manufacturer  trade associations, the Alliance of Automobile Manufacturers
and the Global Automakers, provided the agencies with information on several technical projects
they initiated.  This work included an assessment of the penetration of GHG/fuel economy
technologies in model year 2012-2014 vehicles, an assessment of technology effectiveness, and
an assessment of vehicle footprint.

2.2.2.2 Automotive Suppliers
H Per Honda's request, EPA has placed in the docket a public version of the company's presentation materials, from a
  meeting on October 7, 2015. The presentation materials for other auto manufacturers were designated as
  confidential business information by the manufacturers.
                                              2-6

-------
                           Overview of the Agencies' Approach to the Draft TAR Analysis

   The agencies met with numerous automotive suppliers on several occasions, including Aisin,
Borg-Warner, Bosch, Continental, Dana, Delphi, Denso, Eaton, Getrag, Honeywell, Jatco,
Mahle, Ricardo, Roechling Automotive, Schaeffler, Tennaco, Valeo, and many others.
Automotive suppliers provided the agencies with detailed information on the effectiveness, costs,
lead-time and implementation issues surrounding various GHG/fuel economy technologies
including powertrain systems, engines, transmissions, accessories, tires, valve trains, axles,
active aerodynamics, braking systems, and electrification (stop-start, mild hybrids, 48-volt
systems). Much of this information was used directly to inform the agencies' inputs for
technology costs, effectiveness, and lead-time, which are described in detail in Chapter 5.

   In addition, the agencies met with many trade organizations of various materials used in
automotive manufacturing, including the Aluminum Association, American Plastics Council,
American Iron and Steel Institute, and others. Much of this discussion related to the potential for
various materials, such as high-strength steel, aluminum, and plastics, to contribute to vehicle
mass reduction, which is described further in Chapter 5.

2.2.2.3 Environmental Non-governmental Organizations (NGOs) and Consumer Groups

   The agencies met with a broad coalition of organizations representing both environmental and
consumer advocacy, including the Union of Concerned Scientists, Natural Resources Defense
Council, Environmental Defense Fund, Sierra Club, American Council for an Energy Efficient
Economy, International Council on Clean Transportation, Environment America, Safe Climate
Campaign, Blue Green Alliance, Ceres, Consumer Federation of America, Consumers Union,
Pew Charitable Trusts, Better World Group, and Cater Communications. The groups stressed
the need to ensure that the environmental benefits expected when the National Program was
finalized are actually realized, noting that the Paris international climate agreements will require
continued substantial further reductions in GHG emissions across all sectors, including
transportation. The organizations pointed to the rapid pace of automotive technology
advancements in the marketplace and the important role of the standards in setting long-term
targets and stimulating innovation, and encouraged the agencies to ensure the Draft TAR
analyses are based on the latest data and projections for technology developments out to the 2025
timeframe.  Consumer groups relayed  survey information showing that consumers continue to
want fuel economy improvements, since they expect gas prices will rise. Consumer groups also
noted that gasoline costs are a significant portion of consumers' pocketbook spending,  even more
so for lower income families. Several NGOs noted research projects they're initiating to address
issues relevant to the MTE.  The groups also stressed the need for additional GHG reductions
beyond 2025, and encouraged the agencies to begin exploring a framework for post-2025
standards.

2.2.2.4 State and Local Governments

   The National Association of Clean Air Agencies (NACAA), including many of their state and
local government members, met with the agencies to express their support for strong GHG and
fuel  economy standards. NACAA members expressed their perspective that they are seeing
many fuel saving technologies already in today's vehicles and at greater levels than expected
when the standards were first set. The state/local government agency representatives believe that
the public is concerned about potential rising fuel prices and that, regardless of pump prices,
consumers value the fuel savings that come from improved efficiency.  NACAA members urged
the agencies to conduct a forward-leaning analysis, believing that technologies will develop even
                                              2-7

-------
                           Overview of the Agencies' Approach to the Draft TAR Analysis

faster than the agencies project. The state/local governments want to ensure not only the
significant GHG reductions from these standards, but also the co-pollutant benefits that come
from reduced fuel consumption. NACAA also encouraged the agencies to begin working toward
strong standards for post-2025.

2.2.3   Other Key Data Sources

   In addition to relying on research from the agencies' studies and gathering input from
stakeholders, the agencies also reviewed relevant studies published by other organizations.  One
key study informing the agencies'  assessment is the National Research Council (NRC) of the
National Academies of Sciences (NAS) report, "Cost, Effectiveness and Deployment of Fuel
Economy Technologies for Light-Duty Vehicles" issued in June 2015, as discussed above.21
Throughout this Draft TAR, the agencies discuss specific information provided in the NAS
report, as well as address many of the report's recommendations.

   The agencies have relied on studies published by other federal government organizations,
including the Department of Energy (DOE) studies in areas such as vehicle mass reduction,
impacts of mass reduction on vehicle safety, and battery cost modeling. The Energy Information
Administration's (EIA) 2015 Annual Energy Outlook formed the basis for the agencies'
assumptions about full production, future fuel  prices, and the sizes of the future passenger car
and light truck markets. Market forecast information from IHS Automotive informed
assumptions regarding brand and segment shares of the future light vehicle market.

   Beyond our partners in the U.S. government, the Canadian government, including
Environment and Climate Change Canada (ECCC) and Transport Canada, has supported
significant research in the areas of vehicle light-weighting, aerodynamics, tire efficiency, the
effect of mass reduction on vehicle dynamics performance (e.g., braking and handling), and all-
wheel drive vehicle technology. These reports are described in more detail in Chapter 5. This
work is part of a collaboration under the framework of the Canada-U.S. Air Quality Agreement
which includes a commitment for ECCC and EPA to work together toward the alignment of
vehicle and engine emission regulations and coordinated implementation. The Canadian
government has established light-duty GHG standards aligned with the U.S. standards through
2025, and Canada plans to collaborate with the U.S. on a midterm evaluation of the model year
2022-2025 standards.

   The agencies stayed abreast of technology and economic developments by reviewing
published literature and attending  technical/scientific conferences.22 For  example, since late
2012, there have been hundreds of papers published in the literature (e.g., SAE International)
related to GHG/fuel economy technologies, as well as numerous publications presented in other
forums.  Collectively  the agencies' staff attended more than 60 technical conferences.  Data
gathered from these papers and conferences directly informed the technology inputs described in
detail in Chapter 5. Agency staff also reviewed relevant literature on the host of other issues
discussed throughout this Draft TAR, including climate science and energy security issues,
economic issues (such as rebound, automotive employment, affordability, consumer willingness
to pay for vehicle attributes), transportation issues (such as travel demand), and others.
                                              2-8

-------
                           Overview of the Agencies' Approach to the Draft TAR Analysis

2.3    Agencies' Approach to Independent GHG and CAFE Analyses

   NHTSA, CARB, and EPA have made significant updates to the assessment of CAFE and
GHG technology readiness, technology effectiveness, and technology costs since the 2012 FRM,
including investigating a number of technologies not considered in 2012.  These efforts are
consistent with the recommendations of the 2015 NAS report1 and reflect the significant rate of
technological progress that has been made in the automotive industry since the FRM.J While all
three agencies have been working collaboratively on an array of issues throughout this initial
phase of the Midterm Evaluation, much of the EPA GHG and DOT CAFE assessments were
done largely independently. The independent analyses were done in some part to recognize
differences in the agencies' statutory authorities  and through independent decisions made in each
agency. The agencies all agree that independent and parallel  analyses can provide
complementary results, and in this Draft TAR the independent NHTSA CAFE assessment and
EPA GHG assessment both show that the 2022-2025 standards can largely be achieved through
the use of advanced gasoline vehicle technologies with modest penetration of lower cost
electrification (like 12-volt start/stop and 48-volt mild hybrids) and low penetrations of higher
cost electrification (like strong hybrids, plug-in hybrids, and all electric vehicles). The CAFE
and GHG assessments show just two of a number of potential pathways for meeting the
MY2022-2025 standards.

   It is clear that the automotive industry is innovating and bringing new technology to market at
a brisk pace and neither of the respective agency analyses reflect all of the latest and emerging
technologies that may be available in the 2022-2025 time frame. For example, the agencies were
not able for this Draft TAR to evaluate the potential for technologies such as  electric turbo-
charging, variable compression ratio, skip-fire cylinder deactivation, and P2-configuration mild-
hybridization.  These technologies may provide  further cost effective reductions in fuel
consumption and the agencies will continue to update their respective analyses throughout the
MTE process as new information becomes available.

   Both agencies have made broad use of the application of full-vehicle simulation.   This is
consistent with the NAS's conclusions in its 2015 report: "Full system simulation is
acknowledged to be the most reliable method for estimating fuel consumption reductions for
technologies before prototype or production hardware becomes available for  testing." In
addition, the NAS also concluded that: "For spark ignition engines, these simulations should be
directed toward the most effective technologies that could be applied in 2025 MY to support the
midterm review of the CAFE standards."  There are many readily available options for full-
vehicle simulation software.  Many vehicle  manufacturers use their own, internally developed
simulation software to estimate the effectiveness of technologies.  In addition, full-vehicle
simulation software packages are also available through engineering consulting firms, such as
Southwest Research Institute, FEV,  Ricardo, AVL, and through academia.

   For the 2012 FRM, both NHTSA and EPA relied on simulation results produced by Ricardo
using Ricardo's proprietary Easy 5 model. Both agencies agreed that greater  transparency would
1 See Chapter 2.2.1 for further discussion of the NAS report.
J See for example Finding 2.4 from the 2015 NAS Study - "Other Technologies by 2025 Not Considered by
  EPA/NHTSA," in which the Committee recommends that NHTSA and EPA consider evaluating a number of
  gasoline engine technologies not evaluated in the 2012 Final Rule.
                                              2-9

-------
                          Overview of the Agencies' Approach to the Draft TAR Analysis

improve the robustness of the regulatory process and both agencies made independent decisions
as to how best to meet this goal.  For this Draft TAR, NHTSA contracted with the Department of
Energy's Argonne National Lab (ANL) to employ the use of the Autonomie model. Autonomie
was developed by ANL and has been largely informed by benchmarking work performed in
ANL's Advanced Powertrain Research Facility and by engine technology analysis performed by
IAV Automotive Engineering. For light-duty, the EPA vehicle simulation model is referred to as
ALPHA - Advanced Light-duty Powertrain and Hybrid Analysis tool.K The supporting
benchmarking and development of ALPHA has been completed by EPA's National Center for
Advanced Technology (NCAT).  In addition, both agencies have applied information regarding
technology  effectiveness from sources other than full-vehicle simulation modeling. These
sources include, for example, stakeholder meetings, the 2015 NAS report, and information from
the technical literature and publications from technical conferences.

  As in past greenhouse gas and fuel economy rulemakings, NHTSA and EPA have utilized
unique program analysis models.  This difference in methodology ensures that the respective
analyses produced by the agencies recognize their respective statutory authorities. EPA has
continued to use its Optimization Model  for Reducing Emissions of Greenhouse Gases from
Automobiles (OMEGA). NHTSA has continued to use its Volpe CAFE Model.

  In addition to the decision to use two different full-vehicle simulation models, NHTSA and
EPA have also made independent decisions regarding some modeling inputs. Many of the
modeling methodologies and inputs are common.L Each of the individual inputs that are
different is  described in its respective section.  The primary differences include engine and
transmission effectiveness,  model year baseline fleet, and mass reduction inputs for both the
baseline assessment and for the overall cost.

The agencies believe that, for this first step of the Draft TAR, it is reasonable to show multiple
pathways for potential compliance with the MY 2022-2025 standards, and to make use of
different data sources and modeling tools.  We welcome public comment on the various sources
of information and analytical approaches. As stated previously, given the rapid pace of
automotive industry innovation, the agencies may consider adding additional technologies as
new information becomes available in the next step of the MTE, in addition to the comments we
receive on this Draft TAR.
K See Chapter 5.3.2 for further discussion of EPA's ALPHA model.
L Where inputs to the analysis are consistent with the FRM, the input has been assessed with respect to the latest
  available information and found to be appropriate.
                                             2-10

-------
                              Overview of the Agencies' Approach to the Draft TAR Analysis

References
1 https://www3 .epa.gov/otaq/climate/mte.htnrfepa-publications.
2 https://www3 .epa.gov/otaq/climate/strategies-vehicle.htm.
3 "Learning Curves used in Developing Technology Costs in the Draft TAR," Memorandum from Todd Sherwood to
Air Docket EPA-HQ-OAR-2015-0827, June 8, 2016.
4 Helfand, G., J. Revelt, L. Reichle, K. Bolon, M. McWilliams, M. Sha, A. Smith, and R. Beach (2015). "Searching
for Hidden Costs: A Technology-Based Approach to the Energy Efficiency Gap in Light-Duty Vehicles." EPA-420-
D-14-010, https://www3.epa.gov/otaq/climate/documents/mte/420dl5010.pdf.
5 Cassidy, A., G. Burmeister, and G. Helfand (2016). "Impacts of the Model Year 2017-25 Light-Duty Vehicle
Greenhouse Gas Emissions Standards on Vehicle Affordability." Draft working paper.
6 Small, K. and K. Hymel. "The Rebound Effect from Fuel Efficiency Standards: Measurement and Projections to
2035." EPA-420-R-15-012. https://www3.epa.gov/otaq/climate/documents/mte/420rl5012.pdf.
7 B. Lee, S. Lee, J. Cherry, A. Neam, J. Sanchez, E. Nam, "Development of Advanced Light-Duty Powertrain and
Hybrid Analysis Tool," SAE 2013-01-0808.
S. Lee, B. Lee, J. McDonald, J. Sanchez, E. Nam, "Modeling and Validation of Power-Split and P2 Parallel Hybrid
Electric Vehicles," SAE 2013-01-1470.
S. Lee, B. Lee, J. McDonald, E. Nam, "Modeling and Validation of Lithium-Ion Automotive Battery Packs," SAE
2013-01-1539.
S. Lee, J. Cherry, B. Lee, J. McDonald, M.  Safoutin "HIL Development and Validation of Lithium Ion Battery
Packs," SAE 2014-01-1863.
M. Stuhldreher, A. Moskalik, C. Schenk, J.  Brakora, D. Hawkins, P. Dekraker, "Downsized boosted engine
benchmarking method and results," SAE 2015-01-1266.
A. Moskalik, P. Dekraker, J. Kargul, D. Barba, "Vehicle Component Benchmarking Using a Chassis
Dynamometer,"  SAE 2015-01-0589.
M. Safoutin,  J. Cherry, J. McDonald, "Effect of Current and SOC on Round-Trip Energy Efficiency of a Lithium-
Iron Phosphate (LiFePO4) Battery Pack," SAE 15PFL-0373.
K. Newman,  J. Kargul, D. Barba, "Benchmarking and Modeling of a Conventional Mid-Size Car Using ALPHA,"
SAE 2015-01-1140.
K. Newman,  J. Kargul, D. Barba, "Development and Testing of an Automatic Transmission Shift Schedule
Algorithm for Vehicle Simulation," SAE 2015-01-1142.
S. Lee, C. Schenk, J. McDonald, "Air Flow Optimization and Calibration in High-compression-ratio Naturally
Aspirated SI  engines with Cooled-EGR," SAE 2016-01-0565.
M. Stuhldreher, "Fuel Efficiency Mapping of a 2014 6-Cylinder GM EcoTec 4.3L Engine with Cylinder
Deactivation," SAE 2016-01-0662.
J. Kargul, K. Newman, P. DeKraker, A. Moskalik, D. Barba, "Estimating GHG Reduction of Combinations of
Current Best-Available and Future Powertrain and Vehicle Technologies for a Midsized Car Using EPA's ALPHA
Model," SAE 2016-01-0910.
B. Ellies, C. Schenk, Paul DeKraker, "Benchmarking and Hardware-in-the-Loop Operation of a 2014 MAZDA
SkyActiv2.0L 13:1 Compression Ratio Engine," SAE 2016-01-1007.
K. Newman,  "EPA ALPHA Modeling of a  Conventional Mid-Size Car with CVT and Comparable Powertrain
Technologies," SAE 2016-01-1141.
A. Moskalik, "Investigating  the Effect of Advanced Automatic Transmissions on Fuel Consumption Using Vehicle
Testing and Modeling," SAE 2016-01-1142.
K. Newman,  "Modeling the Effects of Transmission Type, Gear Count and Ratio Spread on Fuel Economy and
Performance Using ALPHA," SAE 2016-01-1143.
8 Helfand, G., C. Liu, M. Donahue, J. Doremus, A. Kahan, and M. Shelby (2015). "Testing a Model of Consumer
Vehicle Purchases." EPA-420-D-15-011, https://www3.epa.gov/otaq/climate/documents/mte/420d 15011 .pdf.
9 Kim, N., Shidore, N., Rousseau, A. "Sizing Algorithm Validation for Several Vehicle Powertrains" [PowerPoint
slides]. Available at: http://www.nhtsa.gov/staticfiles/rulemaking/pdf/cafe/ANL-Autonomie-Sizing-Algorithm-
Validation-1509.pdf.
10 Rousseau,  A.,  Shidore, N., Karbowski, P., Sharer, P. "Autonomie Vehicle Validation Summary" [PowerPoint
slides]. Available at: http://www.nhtsa.gov/staticfiles/rulemaking/pdf/cafe/ANL-Autonomie-Vehicle-Model-
Validation-1509.pdf.
                                                   2-11

-------
                               Overview of the Agencies' Approach to the Draft TAR Analysis
11 Islam, E., Moawad, A., Rousseau, A. (July 25th, 2015) "Fuel Consumption Improvement Study over
combinations of technologies [Rev. 2] VOLPE vs. Autonomie (ANL)" Presentation for DOT/VOLPE.
12 Moawad, A., Rousseau, A. (June 2, 2015) "Engine Parametric Study /-10kW range with a step of IkW"
Presentation for DOT/VOLPE. [PowerPoint slides]. Available at:
http://www.nhtsa.gov/staticfiles/rulemaking/pdf/cafe/ANL-Engine-Parametric-Study-150602a.pdf
13 Moawad, A., Balaprakash, P., Rousseau, A. EVS28 Conference "Novel Large Scale Simulation Process to
Support DOT's CAFE Modeling System." KINTEX, Korea, May 3-6, 2015. Available at:
http://www.nhtsa.gov/staticfiles/rulemaking/pdf/cafe/ANL-EVS28-Novel-Large-Scale-Simulation-Process-to-
Support-DOT-paper.pdf.
14 Moawad, A., Balaprakash, P., Rousseau, A. EVS28 Conference "Novel Large Scale Simulation Process to
Support DOT's CAFE Modeling System."" KINTEX, Korea, May 3-6, 2015. [PowerPoint slides]. Available at:
http://www.nhtsa.gov/staticfiles/rulemaking/pdf/cafe/ANL-EVS28-Novel-Large-Scale-Simulation-Process-to-
Support-DOT-ppt.pdf.
15 Singh, Harry. (2012, August). Mass Reduction for Light-Duty Vehicles for Model Years 2017-2025. (Report No.
DOTHS 811 666). Program Reference: DOT Contract DTNH22-11-C-00193. Contract Prime: Electricore, Inc.
Available at: ftp://ftp.nhtsa.dot.gov/CAFE/2017-25_Final/811666.pdf.
16 Singh, Harry. (2012, August). Report DOT HS 811 666 Cost and CAE models. Available at:
ftp://ftp.nhtsa.dot.gov/CAFE/2017-25_Final/CostModelsAug2012.zip.
17 NHTSA. (2012, July). Peer Review Report. Available at:
http://www.regulations.gov/#!documentDetail;D=NHTSA-2010-0131-0329.
18 Kahane, C. J. (2012, August). Relationships Between Fatality Risk, Mass, and Footprint in Model Year 2000-
2007 Passenger Cars and LTVs - Final Report. (Report No. DOT HS 811 665). Washington, DC: National Highway
Traffic Safety Administration. Available  at: http://www-nrd.nhtsa.dot.gov/Pubs/811665.pdf.
19 Puckett, S.M. and Kindelberger, J.C. (2016, June). Relationships between Fatality Risk, Mass, and Footprint in
Model Year 2003-2010 Passenger Cars and LTVs - Preliminary Report. (Docket No. NHTSA-2016-0068).
Washington, DC: National Highway Traffic Safety Administration.
20 Samaha, R. R., Prasad, P., Marzougui,  D., Cui, C., Digges, K., Summers, S., Patel S., Zhao, L., & Barsan-Anelli,
A. (2014, August). Methodology for evaluating fleet protection of new vehicle designs: Application to lightweight
vehicle designs. (Report No. DOT HS 812 051 A). Washington, DC: National Highway Traffic Safety
Administration. Available at:
http://www.nhtsa.gov/DOT/NHTSA/NVS/Crashworthiness/Vehicle%20Aggressivity%20and%20Fleet%20Compati
bility%20Research/812051-FleetModeling.zip.
21 National Research Council (NRC) of the National Academies' report, "Cost, Effectiveness and Deployment of
Fuel Economy Technologies for Light-Duty Vehicles"  June 2015.
22 Memo to Docket from Robin Moran, EPA, dated June 27, 2016: List of EPA Stakeholder Meetings and
Conference Participation (November 2012 through June 2016) Informing the Midterm Evaluation.
                                                    2-12

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule


Table of Contents

Chapter 3:   Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule	3-1
   3.1    Changes in the Automotive Market	3-2
     3.1.1   Fuel Economy and GHG Emissions	3-2
     3.1.2   Vehicle Sales	3-3
     3.1.3   Gasoline Prices	3-4
     3.1.4   Car and Truck Mix	3-5
     3.1.5   Vehicle Power, Weight,  and Footprint	3-7
     3.1.6   Technology Penetration	3-11
   3.2    Compliance with the GHG Program	3-14
   3.3    Compliance with the CAFE Program	3-17
   3.4    Emerging Transportation Developments	3-22

Table of Figures

Figure 3.1 Average New Vehicle COa and Fuel Economy for Model Years 1975-2015 (production weighted)4... 3-3
Figure 3.2 Actual and Projected Vehicle Production	3-4
Figure 3.3 Gasoline Prices in the United States	3-5
Figure 3.4 Truck Production Share by Year	3-7
Figure 3.5 Average New Vehicle Fuel Economy, Weight, and Power (production weighted)4	3-8
Figure 3.6 Horsepower by Vehicle Class, MY2008-MY2015	3-9
Figure 3.7 Weight by Vehicle Class, MY2008-MY2015	3-10
Figure 3.8 Footprint by Vehicle Class, MY2008-MY2015	3-10
Figure 3.9 Car and Truck Footprint	3-11
Figure 3.10 Light Duty Vehicle Technology Penetration Share since the 2012 Final Rule	3-12
Figure 3.11 Technology Changes since MY2009	3-14
Figure 3.12 Industry GHG Compliance Values versus Standards in 2012-2014 Model Years	3-15
Figure 3.13 Industry CAFE Compliance Values versus Standards in Model Years 2011-2014	3-18
Figure 3.14 Increase due to Flexible Fuel Vehicles on CAFE Fleet Performance in Model Years 2011-2014	3-20
Figure 3.15 CAFE Credit Flexibilities Used and Civil Penalty Payments for Model Years 2010-2014	3-20


Table of Tables

Table 3.1 Credit Balances at Conclusion of the 2014 Model Year (Mg)	3-16
Table 3.2 Reported Credits  Sold and Purchased as of the 2014 Model Year (Mg)	3-17
Table 3.3 Industry CAFE Compliance Values versus Standards in Model Years 2011-2014	3-18
Table 3.4 CAFE Credit Balances at Conclusion of the 2014 Model Year	3-21

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

Chapter 3:  Recent Trends in the Light-Duty Vehicle Fleet Since the 2012
Final Rule

   In support of the GHG/fuel economy  rules for MY2017-2025 light duty vehicles, EPA and
NHTSAA performed an extensive analysis of the light-duty automobile marketplace and the
projected impacts of the GHG/fuel economy rules. Those analyses were performed in 2012 and
were based on then-available historical data, market forecasts from commercial sources, and
projections based on the work published in the U.S. Energy Information Administration's (EIA)
Annual Energy Outlook 2011 (AEO 2011) and 2012 Early Release (AEO 2012ER) report.1'2

   Since the publication of the 2012 final rule, the agencies have continued to collect and
evaluate  an extensive amount of light-duty automobile data through the GHG, CAFE, and other
regulatory programs. In December 2015, EPA published two reports based on analysis of the
data provided by manufacturers. The first report is "Light-Duty Automotive Technology,
Carbon Dioxide Emissions, and Fuel Economy Trends: 1975-2015"3 which analyzes the GHG
emissions, fuel economy, and technology trends of new vehicles in the United States since  1975.
The second report is "GHG Emission Standards  for Light-Duty Vehicles: Manufacturer
Performance Report for the 2014 Model  Year."4 This report, which is EPA's third annual report,
documents the compliance status of every manufacturer under the GHG program for MY2012-
2014. Combined, these reports provide an extensive review of the current status of the
automotive industry under the light-duty GHG program.

   NHTSA provides information about manufacturer compliance with CAFE on the CAFE
Public Information Center (PIC) website.5  The PIC website was launched in July 2015 as a
public interface for NHTSA's new CAFE database.  This database  was developed to simplify
data submissions between EPA and NHTSA, improve the quality of the agency's data, expedite
public reporting, improve audit verifications and testing, and enable more efficient tracking of
manufacturers' CAFE credits with greater transparency. NHTSA provides the following CAFE
related reporting exclusively available through its PIC: fleet and manufacturers' fuel economy
performance reporting; reporting on manufacturers' CAFE credit balances; reporting on civil
penalties collected; flexed-fuel vehicle reporting; pre and mid-model year early projections of
CAFE data.

   This chapter is intended to give the reader an  overarching summary of the changes in the
light-duty market in the last four years. The reports issued by EPA and NHTSA document the
progress in the industry, and this section will rely heavily on those reports.  In addition to the
updated EPA and NHTSA analysis, this  section will compare industry trends and projections
from the 2012 FRM to updated AEO 2015 projections.6  These data, and continuing updates to
them, will ultimately influence much of the  underlying analysis throughout the midterm
evaluation.  Throughout the midterm evaluation  process, the agencies will continue to rely on the
most up-to-date data.
A EPA finalized GHG standards for model years 2017-2025 under the Clean Air Act. NHTSA finalized Corporate
  Average Fuel Economy (CAFE) standards for model years 2017-2021 and issued augural standards for model
  years 2022-2025 under the Energy Policy and Conservation Act.

                                              3-1

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

3.1    Changes in the Automotive Market

   Since the promulgation of the 2017-2025 final rulemaking (FRM) in 2012, the automotive
marketplace has undergone many changes. New vehicle sales, fuel economy, and horsepower
are all at record highs. Many new technologies have been quickly gaining market share, gasoline
prices have dropped by more than a third, and truck  share has been increasing.

3.1.1  Fuel Economy and GHG Emissions

   Average new vehicle fuel economy has increased in 8 of the last 10 years, and currently
stands at a record high. Over that span, average new vehicle fuel economy has increased 5 mpg
(a 26 percent increase). For MY2014, the average new vehicle fuel economy6 is 30.7 miles per
gallon (35.6 mpg for cars and 25.5 mpg for trucks) as tested on EPA's 2-cycle city and highway
tests. This 2-cycle (or unadjusted) fuel economy is used as the basis for EPA and NHTSA's
regulatory programs, as required by law, and is generally about 25 percent higher than fuel
economy values that are published for new vehicle labels (also referred to as  adjusted fuel
economy).

   In MY2014, average new vehicle fuel economy was unchanged from MY2013, largely due to
an increasing percentage of truck sales.  However, truck fuel economy in MY2014 increased by
0.8 mpg over the previous year, which was the second largest increase in the  last 30 years.
Truck fuel economy has increased for 10 years in a row and is now at a record 25.5 mpg.
Overall, in MY2014 the improved fuel economy in trucks offset the market shift towards trucks
to result in no change to the overall average fuel economy of new vehicles.

   The trends for new vehicle GHG emissions have  also been favorable, with new 2-cycle
vehicle  GHG emissions at a record low of 290 grams of CO2 per mile on average.  Overall GHG
emissions for new light duty vehicles are down 21 percent in the ten years since MY2004.  EPA
projected GHG emissions year-by-year in the 2012 FRM, and although EPA  does not expect that
actual emissions will match projections made in 2012, for MY2014 the actual vehicle GHG
emissions of 290 g/mile did match the level projected in the 2012 FRM. For a detailed year-by-
year comparison  of achieved GHG emissions compared to the FRM projections, see EPA's GHG
Manufacturer Performance Report.

   Projected data for MY2015, provided to EPA by  manufacturers as part of the vehicle labeling
process, suggests that fuel economy and GHG emissions will improve once again.  Average new
vehicle  fuel economy is projected to increase to 31.2 miles per gallon, and GHG emissions are
projected to decrease to 284 grams per mile. However, gas prices dropped significantly at the
beginning of MY2015, after these projections were provided to EPA by manufacturers,  so these
estimates could change. Figure 3.1 shows the trends in fuel economy and GHG emissions from
1975 to 2015.
B "Average vehicle fuel economy" is the production weighted average for all new light-duty vehicles produced for
  sale in the United States for a given model year.

                                             3-2

-------
                   Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
   700-
   600-
E

E
D)
   500-
O
o
"S  400~
to
'•v
(TJ
5  300-
   200-
                                                  40-
                                                  35-
                                               O
                                               CL
                                                >- 30-
                                                o
                                                  25-
                                                T3
                                                S.
                                                in
                                                3
                                                T3
                                                  20-
                                                  15-
                                                  10-
                                                                           Trucks
       1975  1980 1985 1990  1995  2000 2005  2010  2015

                     Model Year
                                                     1975  1980 1985  1990  1995 2000 2005  2010 2015

                                                                   Model Year
Figure 3.1 Average New Vehicle CCh and Fuel Economy for Model Years 1975-2015 (production weighted)4
3.1.2  Vehicle Sales

   Vehicle sales in the United States are currently at record levels. The number of new light-
duty vehicles sold in the United States reached a new all-time high of 17.5 million vehicles in
calendar year 20157 and sales through the first four months of calendar year 2016 are up by
another 3.4 percent.8  The current state of the auto industry is an impressive turnaround from
only a few years ago. Vehicle sales dropped precipitously to 10.4 million vehicles in calendar
year 2009 due to the Great Recession.  The domestic automakers underwent their own well
documented financial turmoil with GM and Chrysler declaring bankruptcy, and the subsequent
purchase of Chrysler by Fiat. Manufacturers have  increased sales to record highs and returned to
profitability while meeting the first three years of the national program CAFE and GHG
standards.

   EPA and NHTSA track vehicle production by model year,c as opposed to vehicles sales in a
calendar year. These two metrics are slightly different, however they are highly correlated and
trend similarly over time.  Figure 3.2 shows historic vehicle production per model year, as
tracked by EPA and NHTSA. It also includes AEO 2015 new vehicle sales projections, which
provide a forecast to 2040. In AEO 2015, EIA projects relatively flat, but slightly increasing
number of vehicle sales per year.  Also included in Figure 3.2 are the projected model year
production values that were used in the 2012 final rulemaking, based on AEO 2011.  Actual
vehicle sales in 2015 exceeded the final rule's projected values for 2017, by about a million
vehicles. However the AEO  2015  projections predict a slower growth rate going into the future,
c Vehicle production data represent production volumes delivered for sale in the U.S. market, rather than actual
  sales data. They include vehicles built overseas imported for sale in the U.S., and exclude vehicles built in the
  U.S. for export.
                                                3-3

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

which is slightly lower than the final rule's projected vehicle sales towards the end of the 2017-
2025 rule timeframe.
          Q_

          -
          y
          -
          HI
          >
                                           CY 2015 record sales
                          Actual Production
                          (EPA Trends)
                      	2012 FRM Projection
                          (AEO 2011)
                         -AEO 2015 Projection
              2000   2005   2010
2015   2020   2025
      Model Year
2030   2035   2040   2045
                        Figure 3.2 Actual and Projected Vehicle Production
3.1.3  Gasoline Prices

   One recent, unexpected, and significant development in the automotive market has been the
volatility in gasoline prices. In October 2012 when the 2017-2025 rule was finalized, U.S.
average gasoline prices were at $3.87 per gallon. The agencies, based on AEO 2011, projected
in the 2012 FRM that gasoline prices would climb slowly over time. Instead, gasoline prices
dropped more than 40 percent in the United States, and ended 2015 at about $2.15 per gallon.9

   Historically, the price of gasoline has been volatile and difficult to predict accurately. The
price of gasoline, which generally reflects crude oil prices, fluctuates based on the world supply
of and demand for oil. Many factors, including growing demand from developing countries,
natural disasters, economic conditions, geo-political events,  and introduction of new technology,
can all have large impacts on the supply and demand for crude oil. In particular, U.S. production
of crude oil increased more than 70 percent between 2010 and 201510 which undoubtedly
affected domestic oil prices. The combination  of many unpredictable factors has led to
sometimes unanticipated shocks in the short-term price of oil and a long-term trend of oscillation
between high and low prices (as seen in Figure 3.3).

   In AEO 2015, the U.S. Energy Information Administration (EIA) provides three projections
for gasoline prices out to the year 2040. The use of reference, high, and low projections is meant
to capture the broad band of uncertainty for key variables that affect gasoline prices to 2040. In
the reference case, AEO 2015 assumes a continuation of the long-term trend of rising gasoline
prices  and estimates gasoline prices to increase to $3.90 by 2040. The primary factor influencing

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

the long term increase in price is increased world oil demand, especially by non-OECD
(Organization for Economic Cooperation and Development) countries like China and India as
they continue to experience strong economic growth, which offsets any decrease in oil and
gasoline prices due to increased production.  In the high oil price scenario, AEO 2015 projects
gasoline prices 62 percent greater than the reference case, due to higher global oil demand, again
driven by non-OECD nations, as well as lower oil production by the Organization of the
Petroleum Exporting Countries (OPEC), and higher costs of production and development from
non-OPEC countries. For the low oil price scenario, just the opposite is projected, and gasoline
prices fall 33 percent below the reference case by 2040.

   The uncertainty in projecting gasoline prices is reflected in the wide range of gasoline prices
projected in the high and low scenarios. In the high scenario, gasoline prices reach $6.33 per
gallon in 2040. In the low scenario, gasoline prices fall through 2017, then increase
incrementally to $2.60 in 2040. AEO 2015 high and low projections vary by a factor of 2.5,
which reinforces the uncertainty of these projections.

   Historical gasoline prices11 and future AEO 2015 projections are shown in Figure 3.3.
Gasoline prices were at an all-time high in 2012, although in terms of constant dollars were only
slightly above gasoline prices in 1981. The gasoline prices used in the 2012 final rulemaking are
also included in Figure 3.3, and show that the prices used in the rule, which were based on AEO
2011, are well above current gasoline prices.  The AEO 2015 reference projections predict lower
gasoline prices than the rule projections through 2040; however, the rule projections for gasoline
prices are well  below the high AEO 2015 scenario. The volatility in oil prices and the wide
range of AEO projections serve to reinforce the problem of predicting future gasoline prices with
any accuracy.
   ^-r^•
   m
   "o
   •J-l
   ro
   HI
   0
   OJ
   ao
   2

   I
      $7.00
      $6.00
$5.00
      $4.00
      $3.00
      $0.00
 Actual Price
 (EIA)


-2012 FRM
 Projection
 (AEO 2011)

•AEO 2015
 Reference
 Projection

•AEO 2015
 High Oil
 Projection

•AEO 2015
 Low Oil
 Projection
          2000   2005   2010  2015   2020  2025   2030   2035  2040   2045
                          Figure 3.3 Gasoline Prices in the United States
3.1.4  Car and Truck Mix
                                               3-5

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

   The 2012 FRM finalized footprint-based standards designed to spur improvements in all types
of vehicles and maintain consumer choice. EPA and NHTSA used AEO 2011 car and truck fleet
mix projections in the 2012 FRM to evaluate the overall impacts of the rule. Since the 2012
FRM, the light duty vehicle market has moved more towards trucks than projected.

   The overall percentage of trucks sold in the United States increased in MY2014 but has been
somewhat volatile in recent years. The percentage of trucks sold increased 4.8 percentage points
to 40.7 percent of all sales in MY2014, the last year for which EPA has final data. This is still
well below the all-time record of 48 percent of all sales, set in MY2004. Truck market share
increased steadily in all but four years between MY1980 and MY2004, then quickly fell 15
percentage points to 33 percent of all sales in MY2009.  Since MY2009, truck market share has
bounced around between 33 percent and 42.2 percent of all sales. Projected sales (based on
preliminary automaker projections) for MY2015 predict a slight drop in the percentage of trucks
sold; however, lower than expected gasoline prices may alter the final sales data.

   In MY2014, pickups captured 12.4 percent of new vehicle sales, while truck SUVs captured
23.9 percent of sales.  Smaller 2WD SUVs and 2WD crossovers are generally considered cars
under the regulations, and those car SUVs captured  10.1 percent of vehicle sales.  Sales of SUVs
(including "crossover" vehicles) are continuing to grow and have increased from 20 percent of
total sales in 2004 to 34 percent in MY2014.  The growth of SUVs looks to continue, especially
as the market for  small SUVs continues to develop.  Vehicles like the Jeep Renegade, Honda
HR-V, and Chevy Trax represent a relatively new market segment of "subcompact SUVs."
These vehicles can be classified as either cars (for the 2WD versions) or as trucks (for 4WD
versions meeting  several requirements, such as ground clearance) and are further blurring the
line between cars and trucks.

   Figure 3.4 shows the recent trend in truck production share by year, the projections from the
2012 FRM, and AEO 2015 projections looking forward. In MY2014, the 2012 FRM projected
38 percent of new vehicles produced would be trucks. The actual percentage of trucks produced
was just under 41 percent, so truck were about 3 percent more of the market than projected. EPA
does not have final data for MY2015 or MY2016 data, but industry reports suggest a strong
demand for trucks.  The AEO 2015 projections account for a significant increase in truck
production share, but also project that truck share will peak in 2015 before slowly drifting back
to lower levels0.  Under the AEO 2015 high oil price scenario, truck production slowly falls to
39 percent of production in MY2025 and in the low oil  price scenario truck production is 53
percent in MY2025. Many factors could influence the future direction of car and truck sales,
most notably the volatile gasoline prices of recent years. For additional analysis of light-duty
vehicle sales by class, see EPA's Light Duty Automotive Technology, Carbon Dioxide
Emissions and Fuel Economy Trends report (Figure 3.4).
D The historical data in AEO 2015 for 2011-2014 show a higher percentage of trucks than what actually occurred.
  The AEO historical data does not impact the analysis in this report, nor does it impact AEOs long term
  projections. The data will be updated in AEO 2016.

                                              3-6

-------
                  Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
1 nno/
on°/
Rrp/n

££
" 7n%
OJ
-c en0/
c
o
t; 50%
^
"o 40%
CL
7^ ^n°/^
3
9n°/

10%
n%
20







_^\. - 	 ~ 	
\/^ ^. 	



00 2005 2010 2015 2020 2025 2030 2035 204C

— Actual % Truck
(EPATrends)



(AEO 2011)
- AEO 2015
Reference
Projection
••AEO 2015
Projection

Low Oil
Projection
1
                         Figure 3.4 Truck Production Share by Year
3.1.5   Vehicle Power, Weight, and Footprint

   The automotive industry is continuously innovating and improving vehicles offered to
consumers. However, innovations in the automotive industry have not always been used for the
same purposes. For example, from the early 1980s to 2004, vehicles grew steadily larger and
more powerful but fuel economy decreased (Chapter 4.1.4.3 discusses the role of innovation and
how it has been applied in the automotive industry).  Vehicle weight, horsepower, and footprint
are correlated to vehicle fuel economy and GHG emissions. The relationship between fuel
economy, weight, and horsepower is shown in Figure 3.5.
                                             3-7

-------
                   Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
           LO
           h~
           CD

            8
            CD
            O)
            c
           O

            g
            0>
           Q_
                120%-
                100%-
                 80%-
60%-
                 40%-
                 20%-
                -20%-
                -40%-
             Unadjusted Fuel Economy
                      1975  1980   1985   1990  1995   2000   2005   2010  2015

                                            Model Year

       Figure 3.5 Average New Vehicle Fuel Economy, Weight, and Power (production weighted)4
   The baseline analysis presented in the 2012 FRM was based on MY2008.  Since then, average
new vehicle sales weighted horsepower has increased 14 horsepower to a projected record high
233 horsepower in MY2015. Horsepower did decrease in MY2009, but that was the first dip in
horsepower in 28 years.  With the exception of MY2009 and MY2012, horsepower has increased
every year since MY 1981. Both cars and trucks are projected to reach record average
horsepower numbers in MY2015. Since MY2008, car horsepower is up 6 horsepower on
average to 200 horsepower, and trucks are up 29 horsepower on average to 283 horsepower.
Increases in horsepower have been a little more volatile the last few years than the very steady
increases seen for more than 25 years, but clearly manufacturers have continued to increase
average vehicle power in the past several years while also significantly reducing GHG emissions
and increasing fuel economy. Examining horsepower by vehicle type clearly shows that pickup
trucks have experienced the largest increase in horsepower, as shown in Figure 3.6.E
E The five vehicle type categories are those used by EPA in the report "Light-Duty Automotive Technology, Carbon
  Dioxide Emissions, and Fuel Economy Trends: 1975 to 2014." Cars are subdivided into Cars and Car SUVs, and
  trucks are subdivided into Pickups, Truck SUVs, and Vans.
                                              3-8

-------
                  Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
         350
         300
         250
       CL
       
-------
                   Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
6000
5500
5000
€. 4500
•M
f 4000
£
3500
3000
20







=r--^.



__
^^^^^^m





• 	 "


•










*-— — .



^

"• —









	 Car
-Car SUV
Trl ._|, C\ \\l

	 Van

07 2008 2009 2010 2011 2012 2013 2014 2015 2016
Model Year
                      Figure 3.7 Weight by Vehicle Class, MY2008-MY2015
   The GHG/fuel economy standards are based on vehicle footprint, where footprint is defined
as the area where the centers of the four tires touch the ground.  EPA began tracking footprint in
MY2008 and since that time, the average new vehicle footprint has increased to the highest level
on record. New vehicle production weighted footprint is projected to be at 49.9 square feet in
MY2015, which is a small increase of one square foot, or about 2 percent,  since MY2008.  The
average new car footprint is up 0.8 square feet since MY2008, and the average new truck
footprint is up 1.5 square feet.  The increase in truck footprint is driven largely by pickup trucks,
which are up almost 3.2 square feet, or 5 percent, since MY2008. In addition, the recent shift
towards trucks is driving up the overall fleet-wide average footprint of new vehicles.  While
pickup truck footprint has increased, other vehicle segments have been relatively constant since
MY2008, as shown in Figure 3.8.
70
65
if 60
cr
1/1 55
c
I 5°
o
45
40
20







•

	 '

•


^— ^-^—

—

	 •


-


^ 	



	 	


•




.






-Car
-Car SUV
-Truck SUV
Pickup
	 Van
07 2008 2009 2010 2011 2012 2013 2014 2015 2016
Model Year
                     Figure 3.8 Footprint by Vehicle Class, MY2008-MY2015
                                              3-10

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

   The average footprint for new cars and trucks is higher than the 2012 FRM projections.  For
MY2014, cars are 1.5 square feet larger, and trucks are 1.1 square feet larger.  Overall, the
average new vehicle in MY2014 had a footprint of 1.6 square feet more than projected, due to
the increasing percentage of trucks sold. The footprint trends for cars and trucks are shown in
Figure 3.9.
            58

            56
                                                                    •ActualTruck Footprint
                                                                    (EPA Trends)
        ^   52
        $                                                      — — — 2012 FRM Truck Projection
        77   50                                                      (AE02011)
        c

        £•   48                                                    — Actual Car Footprint
                                                                    (EPA Trends)
        £   46
                                                               — — — 2012 FRM Car Projection
                                                                    (AE02011)

            42
            40
              2008 2009 2010  2011 2012 2013 2014 2015 2016 2017

                                  Model Year
                              Figure 3.9 Car and Truck Footprint
   Overall, the general trend since the 2012 FRM continues towards slightly larger vehicles with
more power, particularly for pickup trucks. However, overall new vehicle weight has remained
nearly constant even given the continuing trend towards larger vehicles, and overall fuel
economy has improved. For additional analysis of light-duty vehicle footprint, weight, and
horsepower by vehicle class, see EPA's Light Duty Automotive Technology, Carbon Dioxide
Emissions and Fuel Economy Trends report (Figure 3.5).

3.1.6  Technology Penetration

   In the 2012 FRM, the agencies discussed many technologies that were available to the
industry to improve fuel economy and to reduce GHG emissions.  These technologies largely
included continual improvements to the gasoline internal combustion engine, such as more
advanced engines and transmissions, vehicle light-weighting, aerodynamics, and more efficient
accessories. Many of these technologies were already available on vehicles for sale back in
2012, and meeting future standards would require manufacturers to adopt the technologies on a
more widespread basis across their fleets. This is, in fact, exactly what is happening, as
discussed below.

   Based on the technologies discussed in the 2012 FRM, EPA presented a feasible, least cost
pathway to illustrate that manufacturers could comply with the standards. The pathway reflected
in the 2012 FRM was meant to illustrate one possible path that manufacturers could use to meet

                                              3-11

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

the standards, based on the OMEGA model's projection of the least-cost set of technologies to
meet the 2025 standards.  EPA recognized that each manufacturer could chose a pathway based
on many factors, but most manufacturers are beginning to widely use the technologies outlined
in the 2012 FRM.  Several of the major technologies that were discussed in the FRM are tracked
by EPA as part of the GHG compliance program, and are documented in the Fuel Economy
Trends report. For these technologies, EPA can compare the penetration rate of these
technologies at the time of the 2012 FRM and for current models.

   Figure 3.10 shows the change in production for several emerging fuel economy related
technologies between MY2008,  which was the baseline in the 2012 FRM, and MY2015.  The
MY2015 data are based on projected production volumes from the manufacturers and are the
most current data available. All of the technologies in Figure 3.10 are technologies that were
discussed in the FRM as possible options for manufacturers to use to increase fuel economy,
reduce GHG emissions, and comply with the standards. The pathway presented in the 2012
FRM included many of the technologies that are included in Figure 3.10.  Chapter 5 discusses
these technologies in more depth.
    100%

     90%

     80%
     60%

     50%

     40%

     30%

     20%

     10%

     0%
           VVT   Multi-   GDI    Turbo   Non-  Cylinder Diesel  Hybrid  EVand   CVT    Six   Seven +
                 valve               Hybrid  Deact.               PHEVs         Speed  Speed
                                  Stop/Start                                 Trans.  Trans.

         Figure 3.10 Light Duty Vehicle Technology Penetration Share since the 2012 Final Rule
   In particular, vehicles utilizing gasoline direct injection engines (GDI) have been entering the
market at a very rapid pace. In MY2008, GDI engines represented 2.3 percent of production.
That number has grown to just over 45 percent of expected production in MY2015.
Turbocharged engines have also seen a swift increase in market share. These two technologies
are often employed together as a downsized, turbocharged, GDI engine package that many
manufacturers have released to improve fuel economy and reduce GHG emissions.  Stop-start
systems (excluding hybrids) and cylinder deactivation have also increased market share
significantly.
                                              3-12

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

   Transmission technology has also been changing rapidly.  Six-speed transmissions increased
from 19 percent in MY2008 to a projected market share of 57 percent in MY2015. Continuously
variable transmissions (CVTs), which were not projected to increase market share in the 2012
FRM, have increased from 8 percent in MY2008 to capture just over 20 percent of the market for
MY2015. An additional 16 percent of new vehicles expected to be produced in MY2015 will
have transmissions with 7 or more speeds, up from 2 percent of all new vehicles in MY2008.
Transmissions with 5 or less speeds, which made up over 70  percent of the market in MY2008,
now account for only just over 5 percent of vehicle production.

   Hybrid electric vehicles (HEVs) were 2.5 percent of production for MY2008, and reached
their peak market penetration in MY2010 at 3.8 percent of all new vehicles produced. Since then,
they have fallen back slightly, to a projected 2.9 percent in MY2015.  There are several possible
reasons that HEV sales have been flat. First, non-hybrid vehicles continue to improve fuel
economy at a faster rate than hybrids.  Between MY2004 and MY 2014, the difference in fuel
economy between the average hybrid midsize car and the average non-hybrid midsize car has
fallen from 24 mpg to about 13 mpg.  Second, some HEV buyers may also be looking to all
electric vehicles (EVs) and plug-in hybrid electric vehicles (PHEVs) instead of HEVs. Third,
recent low gas prices may make hybrids less appealing to consumers.

   Plug-in hybrids and electric vehicles continue to enter the market.  There are now 12 battery
EVs and 13 PHEVs available, and more are scheduled to be released in the  coming years.  There
are also 2 fuel cell electric vehicles (FCEVs) available to consumers.  Overall,  sales of these
vehicles are still low, but appear to be slowly growing.  Sales of EVs increased 9 percent in 2015
to about 69,000 vehicles, and EV  sales in the first quarter of 2016 are up 22 percent from the first
quarter of 2015. PHEV sales were down 24 percent in 2015 (largely due to limited supply of one
vehicle early in the year), but are up over 80 percent in the first quarter of 2016 compared to the
first quarter of 2015. Both EVs and PHEVs had first quarter sales in 2016 that were higher than
any other year.12 While overall national sales are low, the 2012 FRM assumed only small
numbers of EVs and PHEVs (2 percent of all vehicles) would be needed to meet the  standards in
MY2025. Further, some regions of the nation (most notably  California) already have EV and
PHEV sales in excess of 2 percent of new car sales today.

   Many of the major technologies analyzed in the 2012 FRM appear to be  on trend for reaching
relatively high penetration levels,  similar to what EPA projected for 2021 and 2025 in its
analysis of least cost compliance pathways.  Figure 3.10 shows the technology  penetration for
several major technologies from MY2008 to MY2015.  The MY2021 and the MY2025 projected
technology penetration levels for each of these technologies,  from the 2012 FRM, is  also
included in Figure 3.11 for comparison.  Chapter 5 of this report examines these technologies in
much more detail, and Chapter 12 evaluates and update the projected technology penetrations.
                                             3-13

-------
                  Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
100%
90%
80%
70%
c
O
tt 60%
T3
0
£ 50%
If-
o
| 40%
QJ
0.
30%
20%
10%
0%
20



-•—Direct Injection
-•-8+ Speeds (and CVTs)
—•—Turbo-downsized*
-•-Stop-Start
-•-HEV
-•-Diesel
-•-EV+PHEV



—•==*
»-^


-—-•7^
.


i
/





/
/ S
/









i
>-f^^.
i 	 * — tf -










FRI\

/
•





\ m
/I Compliance /—
Pathways / '
/\J

•





I ;
\





•
\ ^
'. M
• "*t
V






•
• ;
• *

07 2009 2011 2013 2015 2017 2019 2021 2023 2025 2027
Model Year
       * Data through 2015 includes all turbocharged vehicles, not specifically turbo-downsized engines

                        Figure 3.11 Technology Changes since MY2009


3.2    Compliance with the GHG Program

   Three model years, MY2012-2014, have been completed under the new footprint based GHG
regulations.  In all three model years, manufacturers have outperformed the standards by a wide
margin even as the standards have become more stringent.  In MY2014, the industry compliance
was 13 g/mile better than required by the standards. In model years 2012 and 2013, industry
compliance was 11 and  12 g/mile respectively, better than required. This industry-wide
performance means that, across the fleet, consumers continue to buy vehicles with lower GHG
emissions than required by the EPA standards. The standards decreased 12 g/mile from
MY2012 to 2014, and manufacturers more than kept pace by reducing compliance values by 14
g/mile. A summary of industry compliance values versus the standards is shown in Figure 3.12.
                                             3-14

-------
                   Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
   305


   300


   295


^•290

1
                       11 g/mi
                     lower than target
                                                               — Compliance Standard

                                                                 • Compliance Value
                                     7 g/mi
     2
     .00280
     (D
     O 275
        270
        265
        260
                                              12 g/mi
                                             lower than target
                                                             5 g/mi
                                                                 13 g/mi
                                                               lower than target
                      2012
                                         2013
                                     Model Year
2014
       Figure 3.12 Industry GHG Compliance Values versus Standards in 2012-2014 Model YearsF
   The majority of manufacturers, representing more than 99 percent of U.S. production, are in
compliance with the standards for the 2012-2014 model years. In fact, 20 of 24 manufacturers0
are carrying a positive credit balance into the 2015 model year, meaning that these manufacturers
have met the standards in all of the 2012-2014 model years (credits cannot be carried forward if
a deficit exists in a prior model year). The manufacturers currently with deficits in any or all of
the 2012-2014 model years are allowed to carry those deficits forward for three model years,
giving them time to generate or purchase credits to demonstrate compliance with the 2012-2014
model year standards. Thus, a manufacturer with a deficit remaining from the 2012 model year
has until the end of the 2015 model year to offset that deficit. The current status of
manufacturers carrying  a deficit into the 2015 model year is neither compliance nor non-
compliance, rather, they have not yet fully demonstrated compliance.  The makeup of these
credit and deficit balances is tracked by model year.
F The "Compliance Standard" is the effective overall GHG g/mile standard for all light duty vehicles in a given
  model year, based on the production volumes and footprints of the vehicles produced. The "Compliance Value" is
  the effective overall GHG g/mile emission rate actually achieved by the industry in a given model year, based on
  the production volumes and footprints of the vehicles produced.
G Volkswagen is excluded due to an ongoing investigation.
                                               3-15

-------
                   Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

                Table 3.1 Credit Balances at Conclusion of the 2014 Model Year (Mg)
Credit Balances at Conclusion of the 2014 Model
(including credit transfers & trades)
Manufacturer Credits Carried to 2015
Toyota
Honda
GM
Ford
Hyundai
Nissan
Fiat Chrysler
Subaru
Kia
Mazda
BMW
Mitsubishi
All Manufacturers
Note: Volkswagen is not included
81,271,823
39,410,925
30,380,022
27,514,195
19,727,364
17,810,733
13,890,014
10,236,711
9,819,076
7,160,086
1,532,564
1,333,267
Manufacturer
Suzuki*
Mercedes*
Ferrari
Volvo
Fisker*
Coda*
BYD Motors
Tesla
Lotus*
McLaren*
Aston Martin*
Jaguar Land Rover

in this table due to an ongoing investigation. Based on
Year (Mg)
Credits Carried to 2015
428,242
228.172
107,613
74,291
46,694
7,251
4,824
1,965
(2,841)
(6,507)
(35,844)
* (509,745)
265,182,108
the original compliance data,
Volkswagen has a credit balance of 4,751,213 Mg.
fThese companies are using a temporary program for limited-volume manufacturers that allows some vehicles to be subject
to less stringent standards. See Section 3.B.
*Although these companies produced no vehicles for the U.S. in the most recent model year, the credits generated in
previous model years continue to
exist.

   The 2012 FRM also introduced the options for manufacturers to trade credits between
companies. EPA included this provision because it will allow for greater GHG reductions, lower
compliance costs, and greater consumer choice.  Manufacturers have been actively trading
credits, with almost 10 million Megagrams of CCh credits changing hands by the close of the
2014 model year reporting period.

   The credit transactions reported by manufacturers through the 2014 model year are shown in
Table 3.2. Credit distributions are shown as negative values, in that a disbursement represents a
deduction of credits of the specified model year for the selling manufacturer. Credit acquisitions
are indicated as positive values because acquiring credits represents an increase in credits for the
purchasing manufacturer.  The model year represents the "vintage" of the credits that were sold,
i.e., the model year from which the credits originated. Note that each value in the table is simply
an indication of the quantity of credits from a given model year that has been acquired or
disbursed by a manufacturer, and thus may represent multiple transactions with multiple buyers
or sellers. The total credit balances shown in Table 3.1 include the credits transactions reported
in Table 3.2.
                                              3-16

-------
                   Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
            Table 3.2  Reported Credits Sold and Purchased as of the 2014 Model Year (Mg)



in
13
u

Credits
Purchaed

Manufacturer
Honda
Nissan
Tesla
Toyota
Ferrari
Fiat Chrysler
Mercedes

2010
(3,609,383)
(200,000)
(35,580)
(2,507,000)
265,000
5,651,383
435,580
Model
2011

(1,000,000)
(14,192)

500,000
514,192
Year "Vintage"
2012

(250,000)
(177,941)
-
427,941
2013

(1,048,689)
-
1,048,689
2014
-
(1,019,602)
-
1,019,602

Total
(3,609,383)
(1,450,000)
(2,296,004)
(2,507,000)
265,000
8,219,674
1,377,713
   In the first three years of the GHG compliance program, the industry has outperformed the
standards each year, all large manufacturers are carrying forward credits, and there has been
active trading of credits between manufacturers.  The specific details of the compliance program,
including tailpipe emissions, earned credits, credit trading, and comparisons to the 2012 FRM
projections, are all detailed in the EPA report titled, "GHG Emission Standards for Light-Duty
Vehicles: Manufacturer Performance Report for the 2014 Model Year."

3.3    Compliance with the CAFE Program

   An overview of how manufacturers complied  with the CAFE program is provided for model
years 2011 to 2014.H>1 On average, manufacturers showed significant strides complying with the
CAFE program for model years 2011 and later in improving the total fleet fuel economy
performance for passenger cars and trucks despite increasingly more stringent standards over the
period.  Manufacturers were able to successfully  execute compliance strategies for both the
NHTSA and EPA programs that accommodated the differences in compliance flexibilities and
credit balances between the programs.

   As directed by Congress, the total light duty vehicle fleet is divided into three compliance
categories, domestic and  import passenger cars and light trucks, for meeting CAFE standards and
distinct statutory differences exist in the compliance flexibilities for each category. Figure 3.13
and Table 3.3 provide the total fleet standards and actual fleet fuel economy performance for
each vehicle category.  As shown in the figure, for each model year from 2011 through 2014,
manufacturers far exceeded standards for their combined domestic and import passenger cars but
fell short in meeting standards for their combined light truck fleets for model years 2012 and
H Model year 2011 is an important year in the CAFE programs because it signifies the first year EISA amended
  EPCA mandating the first stage of combined footprint-based CAFE standards and established a credit trading
  program that supplemented previous existing credit flexibilities for all passenger cars and light trucks.  EISA and
  EPCA also required CAFE standards that would increase annually and set sufficiently high enough levels to
  ensure that the total fleet average of all new passenger cars and light trucks, combined, was not less than 35 miles
  per gallon by model year 2020.
1 Model year 2014 is the last year manufacturers, NHTSA and EPA have completed production, testing and
  reporting for all vehicles complying with CAFE standards.
                                               3-17

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

2013.  Consumers will save an estimated 16.6 billion gallons of fuel over the lifetime of model
year 2011 to 2014 vehicles due to the manufacturers exceeding the CAFE standards in those
years.
                            Fleet Fuel Economy Performance
2011             2012


DP, Fleet Standard
DP, FE Performance — *- •
                                                   2013
2014
                                         Model Year
                                         IP, Fleet Standard   —•—LT, Fleet Standard
                                         IP, FE Performance  — *--LT, FE Performance
       Figure 3.13 Industry CAFE Compliance Values versus Standards in Model Years 2011-2014
        Table 3.3 Industry CAFE Compliance Values versus Standards in Model Years 2011-2014
Model
Year
2014
2013
2012
2011
Domestic Passenger Car
FE
Performance
(MPG)
36.3
36.1
34.8
32.7
Fleet
Standard
(MPG)
34.0
33.2
32.7
30.0
Sales
Production
Volume
5,563,657
5,566,615
5,260,200
3,986,385
Import Passenger Car
FE
Performance
(MPG)
36.9
36.8
36.0
33.7
Fleet
Standard
(MPG)
34.6
33.9
33.4
30.4
Sales
Production
Volume
3,641,470
4,172,770
3,396,020
2,965,213
Light Truck
FE
Performance
(MPG)
26.5
25.7
25.0
24.7
Fleet
Standard
(MPG)
26.3
25.9
25.3
24.3
Sales
Production
Volume
6,306,647
5,457,777
4,788,574
5,069,696
   The design of the CAFE program, as instructed by Congress, anticipates that not all
manufacturers' compliance fleets will meet CAFE standards for each model year. Fleets not
meeting CAFE standard represented 44 percent of all fleets on average but represented only 33
percent of the total industry production volume for model years 2012 through 2014.  The
majority of these manufacturers failed to meet the standard for their light truck fleets for these
model years but have rebounded for the 2014 compliance period.

   Therefore, to compensate for shifts in production markets and to allow NHTSA to set CAFE
standards  at the maximum feasible levels, the CAFE program was designed to allow
manufacturers to comply by exercising one or more  program flexibilities to leverage compliance
over multiple model years or by eliminating the deficiencies of under complying fleets using the
                                              3-18

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

benefits gained by over performing fleets/ There are three basic flexibilities outlined by
EPCA/EISA that manufacturers can currently use to achieve compliance with CAFE standards
beyond applying fuel economy-improving technologies: (1) building dual- and alternative-fueled
vehiclesK; (2) banking (carry-forward and carry-back), trading, and transferring credits earned
for exceeding fuel economy standards; and (3) paying civil penalties.L

   Using program flexibilities, all manufacturers not beating the standard have either complied
or will be able to comply with CAFE standards through model year 2014.  As the first
compliance pathway, manufacturers are building advanced technology vehicles and eleven
manufacturers are incentivizing the performance of their fleets by building flexible fueled
vehicles. Building flexible fueled vehicles is a major incentive established by Congress for the
CAFE program.M  Figure 3.14 shows the increase in each compliance category for those
manufacturers building flexible fuel vehicles for the applicable model years. On average, these
manufacturers raised the fleet performance of domestic passenger cars by 1.9 percent, import
passenger cars less than 1 percent and light trucks by 3.4 percent over these model years.

   For the remaining compliance pathways, under-complying manufacturers have offset their
compliance shortfall (credit shortfalls) by carrying forward, backward, transferring or trading
credits.  While some manufacturers are also still paying civil penalty payments for
noncompliance, the amount has significantly decreased mainly due to an active credit trading
market.  An overview of the compliance credit flexibilities used by manufacturers from model
year 2011 through 2014 is shown in Figure 3.15.

   NHTSA anticipates that credit trading will continue to be a major incentive for manufacturers
in the upcoming model years as credit trading was the primary flexibility in model year 2014.
NHTSA predicts that the CAFE credit market moving into model year 2015 for each compliance
fleet is robust enough to allow manufacturers not meeting  standards to continue to comply for
the next several model years. A summary of the CAFE credits carrying into model year 2015 is
shown in Table 3.4.
1EPCA, as amended by EISA, is very prescriptive with regard to the number of flexibilities that are available to
   manufacturers to help them comply with the CAFE standards but intentionally placed some limits on certain
   flexibilities and incentives for the purpose of balancing energy-savings.
K Incentives are allowed for building advanced technology vehicles such as hybrids and electric vehicles,
   compressed natural gas vehicles and building vehicles able to run on dual fuels such as E85 and gasoline.
L We note that while these flexibility mechanisms will reduce compliance costs to some degree for most
   manufacturers, although 49 U.S.C. 32902(h) expressly prohibits NHTSA from considering the availability of
   statutorily-established credits (either for building dual- or alternative-fueled vehicles or from accumulated
   transfers or trades) in determining the level of the standards. Thus, NHTSA may not raise CAFE standards
   because manufacturers have enough of those credits to meet higher standards. This is an important difference
   from EPA's authority under the CAA, which does not contain such a restriction, and which would allow EPA to
   set more stringent standards as a result.
M Congress established the  Alternative Motor Fuels Act (AMFA) which allows manufacturers to increase fleet Fuel
   Economy Performance values by producing dual fueled vehicles. For model years 1993 through 2014, the
   maximum increase in CAFE performance for a manufacturer attributable to dual fueled vehicles is 1.2 miles per
   gallon for each model year and thereafter decreases by 0.2 miles per gallon each model year until ending in 2019
   (see 49 U.S.C. 32906).

                                                 3-19

-------
                  Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule
                                                                          Without AM FA
                                                                         ' (no FFVcredits)

                                                                         I With AM FA
                                                                          (FFV credits)
            5.0
                   10,0
                          15.0
                                 20.0
                                         25.0
                                                30.0
                                                       35.0
                                                              40.0
                                                                     45.0
Figure 3.14 Increase due to Flexible Fuel Vehicles on CAFE Fleet Performance in Model Years 2011-2014
                         CAFE Compliance by Flexibility
50,000,000
45,000,000
40,000,000
35,000,000
30,000,000
25,000.000
20,000,000
15,000,000
10,000,000
 5,000,000
        0
                                                                             I Credit Shortfall
                                                                             I Carry-forward
                                                                             i Trade
                                                                             I Carryback
                                                                             i Civil Penalty
                                                                             (Transfer
                      2010
                          2011
    2012
Model Year
2013
2014
   Figure 3.15 CAFE Credit Flexibilities Used and Civil Penalty Payments for Model Years 2010-2014
                                              3-20

-------
                       Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

                   Table 3.4 CAFE Credit Balances at Conclusion of the 2014 Model Year
Manufacturer
Aston Martin
BMW
BYD Motors
Coda
Daimler
Fiat Chrysler
Ford
General Motors
Honda
Hyundai 2
Jaguar Land Rover
Kia3
Lotus
Mazda
McLaren 4
Mitsubishi
Nissan 5
Pasani6
Spvker7
Subaru
Suzuki
o
Tesla
Toyota
Volkswasen
Volvo
All Manufacturers
Credits
DP




0
99,987,234
157,373,701
65,229,249
133,012,923




15,526


116,007,703


2,256,442

8,020,132
167,007,230
8,756,755
37,435
757,704,330
Carried to 2015
IP
0
6,030,713
177,951
331,750
0
284,321
1,175,577
10,617,792
35,237,193
127,023,114
0
54,652,961
0
49,341,062
0
6,067,098
3,014,623
0
0
4,528,333
2,016,752

342,032,536
24,505,396
-247,890
666,789,282
LT

235,952


0
-4,174,892
17,818,347
38,007,715
32,427,500
7,060,784
0
3,838,194

6,525,997

2,574,682
5,399,372


50,901,342
244,384

29,446,815
2,921,482
-315,044
192,912,630
 Aston Martin has submitted a petition for an alternative standard for MYs 2008 - 2014. This petition for an alternate standard is
pending.
2 MY2014 EPA report is pending
3 MY2014 EPA report is pending
 McLaren has submitted a petition for an alternative standard for MYs 2012 - 2014. This petition for an alternate standard is
pending.
 Nissan IP and DP fleets were exempt from two-fleet rule for model years 2006 - 2010
 Pagani has submitted a petition for an alternative standard for MY2014.
7 Spyker has submitted a petition for an alternative standard for MYs 2008 - 2010. This petition for an alternate standard is
pending.
8 Priorto MY2012, per 40 CFR600.001(b)(l), manufacturers that produced only electric vehicles were exempt from submitting
CAFE information . EPA did not test vehicles and confirm compliance values of manufacturers who produce only electric vehicles
from this time period.
9 Volkswagen is included in an ongoing investigation. Data provided is based on original compliance data.
                                                          3-21

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

3.4    Emerging Transportation Developments

   The automotive industry of today is rapidly evolving, and the pace of change is only
increasing. Major automotive CEOs are not just talking about horsepower, but about becoming
"mobility" companies13 and "disrupting" the industry.14 Technology companies that have not
previously been associated with the automotive industry are further challenging and changing the
industry as connectivity, autonomous  driving, and infotainment systems continue to become a
more prominent part of automotive design.

   Autonomous vehicle developments are regularly in the headlines, with most manufacturers,
many suppliers, and several technology companies actively developing and testing autonomous
systems.  Semi-autonomous systems are already available in some luxury vehicles today, and
many more are promised in the next several years.  The race to develop fully autonomous
vehicles is clearly a high priority across the industry.  Emerging in parallel with vehicle
automation is vehicle connectivity. Vehicle connectivity, in conjunction with automated
systems, has the potential benefit of allowing vehicles to communicate with each other and with
infrastructure to optimize vehicle driving behavior to current conditions, and to interact with
other vehicles and infrastructure to reduce congestion. And of course, connectivity can also
mean more access to high speed data,  entertainment, and productivity applications.

   In addition to connected and  automated vehicles, new companies based on the idea of the
sharing economy are already upending how some people think about transportation and mobility
in general. Ride hailing services continue to grow quickly and are already disrupting rental car
and taxi business models. The largest ride hailing service is already valued more than several
major OEMs after only a few years of existence.15'16  The rapidly expanding list of transportation
related apps for everything from finding a parking spot more efficiently, sharing rides,  or finding
public transit options also point  to  an  industry that is facing rapid change.

   Autonomous vehicles, shared mobility, parking apps, and other innovations were not
considered by the agencies in the GHG rules, but their net impact on GHGs and fuel economy is
yet unknown.  They could ultimately have a very profound impact on the efficiency of our future
transportation system. Preliminary research suggests that connected and automated vehicles
could lead to dramatically reduced GHG emissions through more efficient driving, better traffic
flow, shared mobility, and by enabling greater use of electrification. However, the  same research
acknowledges that the technology  could also lead to increased vehicle miles traveled, higher
speeds, and more vehicle content which could result in a large increase in emissions instead.17'18'
19,20  1^^ emerging technologies and transportation changes will pose considerable future
challenges and uncertainties. At the present time and probably even over the next several years,
there will  continue to be much uncertainty around the impacts of these changes on the
transportation system. It is likely that many of these transformational changes will  have impacts
for the longer-term, and it will be difficult to assess any specific impacts in the 2022-2025
timeframe. While the agencies will continue to keep  abreast of data and analyses surrounding
transportation impacts of these transformations, it is likely that such uncertainty will remain
throughout the timeframe of the midterm evaluation.  EPA, NHTSA, and CARB are beginning to
explore research on the potential emissions and fuel economy impacts of emerging
transformational technologies and  transportation trends.  The agencies will continue to stay
abreast of future research and partner  with stakeholders to evaluate these emerging technologies
                                              3-22

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

and transportation trends, which may help to inform any regulatory development beyond model
year 2025.
                                            3-23

-------
	Recent Trends in the Light-Duty Vehicle Fleet Since the 2012 Final Rule

References
1 U.S. Energy Information Administration, Annual Energy Outlook 2011, DOE/EIA-0383(2011) (Washington, DC,
April 2011), www.eia.gov/forecasts/aeo.
2 U.S. Energy Information Administration, Annual Energy Outlook 2012 Early Release, DOE/EIA-0383(2012)
(Washington, DC, January 2012), www.eia.gov/forecasts/aeo.
3 (EPA) Environmental Protection Agency 2015. Greenhouse Gas Emission Standards for Light-Duty Vehicles—
Manufacturer Performance Report for the 2014 Model Year, EPA Office of Transportation and Air Quality, EPA-
420-R-14-023a, December 2015.
4 (EPA) Environmental Protection Agency 2015. Light-Duty Automotive Technology, Carbon Dioxide Emissions,
and Fuel Economy Trends: 1975 through 2015. U.S. EPA-420-R-15-001, Office of Transportation and Air Quality,
December 2015.
5 National Highway Traffic Safety Administration, CAFE Public Information Center.
http://www.nhtsa.gov/CAFE PIC/CAFE PIC home.htm. Accessed 3/4/2016.
6 U.S. Energy Information Administration, Annual Energy Outlook 2015, DOE/EIA-0383(2015) (Washington, DC,
April 2015), www.eia.gov/forecasts/aeo.
7 Specter, M.; Bennett, I, U.S. Car Sales Set Record in 2015. The Wall Street Journal, 1/5/2016.
8 Market Data Center, Auto Sales. Wall Street Journal, http://online.wsj.eom/mdc/public/page/2  3022-
autosales.html. Accessed 5/11/2016.
9 U.S. Energy Information Administration, Weekly Retail Gasoline and Diesel Prices.
http://www.eia.gov/dnav/pet/pet_pri gnd dcus  nus w.htm. Accessed 5/11/2015.
10 U.S. Energy Information Administration, U.S. Field Production of Crude Oil.
http://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=pet&s=mcrfpus2&f=a. Accessed 5/11/2016.
11 U.S. Energy Information Administration, Short Term Energy Outlook: Real Prices Viewer.
http://www.eia.gov/forecasts/steo/realprices/. Accessed 5/11/2015.
12 Cobb, J. March 2016 Dashboard,  http://www.hvbridcars.com/market-dashboard/. Accessed 5/11/2016.
13 Kiley, D. Ford's Mark Fields is all about mobility and tech, links with Google and Amazon. Forbes. January 5,
2016.
14 DeBord, M., General Motors CEO Mary Barra: 'We are disrupting ourselves, we're not trying to preserve a model
of yesterday. Business Insider. November 16, 2015.
15 LaMonica, P. Is Uber really worth more than Ford and GM? CNN Money. October 27, 2015.
16 Hirsh, J. Major auto industry disruption will lead to robotic taxis, Morgan Stanley says. Los Angeles Times. April
7,2015.
17 Brown, A.; Gonder J.; Repac, B. An analysis of possible energy impacts of automated vehicles. In Road Vehicle
Automation; Meyer, G., Beiker S. Eds.; Springer International Publishing: 2014; pp 137-153.
18 Greenblatt, J. B.; Saxena, S. Autonomous taxis could reduce greenhouse gas emissions of light-duty vehicles by
more than 90 percent. Nature Climate Change. 2015, 5 (9), 860-863; DOI 10.1038/nclimate2685.
19 Wadud, Z.; MacKenzie, D.; Leiby, P. Help or Hindrance? The travel, energy, and carbon impacts  of highly
automated vehicles. Transportation Research Part A, 2016, 86, pp 1-18.
20 Simon K.; Alson, J; Snapp, L; Hula, A. Can Transportation Emission Reductions be achieved autonomously?
Environ. Sci. Technol., 2015, 49 (24), pp 13910-13911.
                                                     3-24

-------
                                                  Baseline and Reference Vehicle Fleets
Table of Contents

Chapter 4:  Baseline and Reference Vehicle Fleets	4-1
  4.1    EPA's Baseline and Reference Vehicle Fleets	4-1
    4.1.1   Why does the EPA Establish Baseline and Reference Vehicle Fleets?	4-1
    4.1.2   EPA's 2014 MY Baseline Fleet	4-2
       4.1.2.1   EPA's MY2014 Based MY2022-2025 Reference Fleet	4-9
         4.1.2.1.1   On What Data Are EPA's Reference Vehicle Fleet Volumes Based?	4-10
         4.1.2.1.2   How did the EPA develop the 2014 Baseline and 2022-2025 Reference
         Vehicle Fleet Volumes?	4-11
         4.1.2.1.3   How was the 2014 Baseline Data Merged with the IHS-Polk Data?	4-11
         4.1.2.1.4   How were the IHS-Polk Forecast and the Unforced AEO 2015 Forecast
         Used to Project the Future Fleet Volumes?	4-12
       4.1.2.2   What Are the Sales Volumes and Characteristics of the MY2014 Based
       Reference Fleet?	4-19
       4.1.2.3   What Are the Differences in the Sales Volumes and Characteristics of the
       MY2008 Based and the MY2014 Based Reference Fleets?	4-22
    4.1.3   Relationship Between Fuel Economy and Other Vehicle Attributes	4-26
       4.1.3.1   Recent Studies of the Engineering Tradeoffs between Power and Fuel Economy,
       and Increases in Innovation	4-29
       4.1.3.2   The Role of the Standards in Promoting Innovation	4-32
       4.1.3.3   Potential Ancillary Benefits of GHG-Reducing Technologies	4-34
       4.1.3.4   Estimating Potential Opportunity Costs and Ancillary Benefits	4-36
    4.1.4   Incorporation of the California Zero Emissions Vehicle (ZEV) Program into the
    EPA Reference Fleet	4-37
       4.1.4.1   The ZEV Regulation in OMEGA	4-37
       4.1.4.2   The ZEV Program Requirements	4-43
         4.1.4.2.1   Overview	4-43
         4.1.4.2.2   ZEV Credit Requirement	4-44
         4.1.4.2.3   Proj ected Representative of PHEV and BEV Characteristics for MY2021 -
         2025      4-45
         4.1.4.2.4   Calculation of Incremental ZEVs Needed for ZEV Program Compliance.. 4-
         49
  4.2    Development of the CAFE Light Duty Analysis Fleet	4-53
    4.2.1   Why did NHTSA Develop the Analysis Fleet?	4-53
    4.2.2   How the MY2015 Analysis Fleet Was Developed	4-53
       4.2.2.1   Background	4-53
    4.2.3   NHTSA Decision to use 2015 Foundation for Analysis Fleet	4-54
    4.2.4   Developments in 2015	4-55
    4.2.5   Manufacturer-Provided Information for 2015	4-56
    4.2.6   Other Data	4-57
       4.2.6.1   Redesign/Refresh Schedules	4-57
       4.2.6.2   Technologies	4-58
       4.2.6.3   Engine Utilization	4-58
    4.2.7   Estimated Technology Prevalence in the MY2015 Fleet	4-59
    4.2.8   Engine and Platform Sharing	4-62
       4.2.8.1   Platform Sharing	4-62

-------
                                                         Baseline and Reference Vehicle Fleets
       4.2.8.2  Engine Sharing & Inheritance	4-63
     4.2.9  Class Types and Assignment	4-64
       4.2.9.1  Regulatory Class	4-64
       4.2.9.2  Safety Class	4-64
       4.2.9.3  Technology Class	4-64
       4.2.9.4  Technology Cost Class	4-65
     4.2.10  Mass Reduction and Aero Application	4-65
       4.2.10.1   Mass Reduction	4-65
          4.2.10.1.1   Mass Reduction Residual Analysis for Footprint	4-73
          4.2.10.1.2   Mass Reduction Residual Analysis for Low and High Price Platforms 4-76
          4.2.10.1.3   Mass Reduction Residual Trends for Company Heritage	4-78
       4.2.10.2   Aerodynamic Application	4-80
     4.2.11  Projecting Future Volumes for the Analysis Fleet	4-82

Table of Figures

Figure 4.1  The Verify Process for the Data EPA's MY2014 Baseline Vehicle Fleet is Based	4-2
Figure 4.2  Process Flow for Creating the Baseline and Reference Fleet	4-4
Figure 4.3  Process Flow for Determining where Segment Volume Should Move	4-15
Figure 4.4  Relative Cost of ZEV Credits for Different Ranges and Battery Costs	4-46
Figure 4.5  Mass Reduction Regression Residual Plot by Body Style	4-69
Figure 4.6  Mass Reduction Assignments by Platform	4-70
Figure 4.7  Mass Reduction Residual Histogram for All MY2015 Platforms	4-73
Figure 4.8  Mass Reduction Platform Residuals vs. Footprint	4-75
Figure 4.9  Mass Reduction Residual Distribution of Platforms with Base Price of $30k or Less	4-76
Figure 4.10 Mass Reduction Residual Distribution of Platforms with Base Price between $30k-$50k	4-77
Figure 4.11 Mass Reduction Residual Distribution of Platforms with Base Price of $50k and Above	4-77
Figure 4.12 Mass Reduction Residuals for Platforms with North American Heritage	4-78
Figure 4.13 Mass Reduction Residuals for Platforms with European Heritage	4-79
Figure 4.14 Mass Reduction Residuals for Platforms with Asian Heritage	4-79
Figure 4.15 Distribution of Aerodynamic Drag Coefficients by Vehicle Body Style	4-81
Figure 4.16 Data Sources and Construction of the Production Forecast	4-83
Figure 4.17 MY2015 Market Shares by Manufacturer	4-85
Figure 4.18 MY2025 Market Shares by Manufacturer	4-85


Table of Tables

Table 4.1  MY2014 Engine Technology Penetration	4-6
Table 4.2  Change (2014-2008) in Engine Technology Penetration	4-8
Table 4.3  MY2015 Ford Engine Technology Penetration	4-9
Table 4.4  AEO 2015 Unforced Reference Case Values used in the 2014 Market Fleet Projection	4-10
Table 4.5  AEO 2015 Reference Case Values	4-10
Table 4.6  List of IHS-Polk Segments	4-13
Table 4.7  Example of Honda Vehicles Being Mapped to Segments Based On the IHS-Polk Forecast	4-13
Table 4.8  Example Honda 2014 Volumes by Segment from the IHS-Polk Forecast	4-16
Table 4.9  Example Values Used to Determine the MDPV Multiplier for FCA	4-17
Table 4.10 Example Values Used to Determine FCA's 2025 Van Volume	4-17
Table 4.11 Example Values Used to Determine FCA 2025 Individual Full-Size Non-Premium Van Multiplier.. 4-18
Table 4.12 Example Applying the Individual Full-Size Non-Premium Van Multiplier for FCA	4-18
Table 4.13 Example Unforced AEO 2015 Truck and Car Multipliers in MY2025	4-18
Table 4.14 Example Applying the Unforced AEO Truck Multiplier to FCA Full-Size Non-Premium Vans	4-19
Table 4.15 Vehicle Segment Volumes	4-19

-------
                                                           Baseline and Reference Vehicle Fleets
Table 4.16 Car and Truck Volumes	4-19
Table 4.17 Car and Truck Definition Manufacturer Volumes	4-20
Table 4.18 Production Weighted Foot Print Mean	4-21
Table 4.19 Percentages of 4, 6, and 8 Cylinder Engines by Model Year	4-21
Table 4.20 Vehicle Segment Volumes Differences	4-22
Table 4.21 2014 Projection-2008 Projection Total Fleet Volumes Differences	4-23
Table 4.22 2014 Projection - 2008 Projection Manufacturer Volumes Differences	4-24
Table 4.23 2014 Projection - 2008 Projection Production Weighted Foot Print Mean Difference	4-25
Table 4.24 Differences in Percentages of 4, 6 and 8 Cylinder Engines by Model Year	4-26
Table 4.25 OMEGA MY2021 Car Fleet using the AEO 2015 Reference Fuel Price Case	4-38
Table 4.26 OMEGA MY2021 Truck Fleet using the AEO 2015 Reference Fuel Price Case	4-38
Table 4.27 OMEGA MY2025 Car Fleet using the AEO 2015 Reference Fuel Price Case	4-39
Table 4.28 OMEGA MY2025 Truck Fleet using the AEO 2015 Reference Fuel Price Case	4-39
Table 4.29 Vehicle Types Considered for Conversion to ZEVs	4-41
Table 4.30 Example Manufacturer Fleet from which ZEVs are to be Created	4-41
Table 4.31 Number of Additional ZEV Program Sales from each Platform	4-42
Table 4.32 Percentage of Additional ZEV Program Sales from Each Vehicle Model	4-42
Table 4.33 Example Manufacturer's OMEGA Fleet including ZEV Program Sales	4-42
Table 4.34 ZEV Regulation Credit Requirements	4-44
Table 4.35 Range Characteristics of BEVs for MY2014	4-47
Table 4.36 Range Characteristics of PHEVs for MY2014	4-47
Table 4.37 Projected Sales Weighted BEV Range for MY2021-2025	4-48
Table 4.38 Projected Sales Weighted PHEV Range andUS06 Capability for MY2021-2025	4-49
Table 4.39 Incremental PHEV40s and BEV200s needed in MY2021	4-51
Table 4.40 Incremental PHEV40s and BEV200s needed in MY2025	4-52
Table 4.41 Summary of Portfolio Revisions by Manufacturer	4-55
Table 4.42 Estimated Average Production Life For Freshly Redesigned Vehicle, By Manufacturer, By Segment... 4-
            57
Table 4.43 Engine Technologies by Manufacturer	4-60
Table 4.44 Electrification Technologies by Manufacturer	4-61
Table 4.45 Transmission Technology by Manufacturer	4-62
Table 4.46 Mass Reduction Body Style Sets	4-66
Table 4.47 Regression Statistics for Curb Weight (Ibs.)	4-67
Table 4.48 Mass Reduction Levels by Residual Error	4-69
Table 4.49 Vehicle Platforms with Highest Estimated Levels of Mass Reduction Technology	4-71
Table 4.50 2015MY Mass Reduction Level by Manufacturer as a Percent of Vehicle Sales	4-72
Table 4.51 Mass Reduction Platform Residuals for Platforms with the Smallest Footprint	4-74
Table 4.52 Criteria for Limiting Additional Application of Mass Reduction Technology in the CAFE Analysis. 4-75
Table 4.53 Mass Reduction Average Residual by Average Platform Base Price	4-76
Table 4.54 Mass Reduction Average Residual by Parent Company Heritage	4-78
Table 4.55 Aerodynamic Drag Coefficients by Body Style	4-80
Table 4.56 Aerodynamic Application by Manufacturer as a Percent of MY2015 Sales	4-81

-------
                                                   Baseline and Reference Vehicle Fleets
Chapter 4: Baseline and Reference Vehicle Fleets

4.1    EPA's Baseline and Reference Vehicle Fleets

   The passenger cars and light trucks sold currently in the United States, and those that are
anticipated to be sold in the MYs 2021-2025 timeframe, are highly varied and satisfy a wide
range of consumer needs. From two-seater miniature cars to 11-seater passenger vans to large
extended cab pickup trucks, American consumers have a great number of vehicle options to
accommodate their needs and preferences. The recent decline in oil prices and the improved
state of the economy have demonstrated that consumer demand and choice of vehicles within
this wide range can be sensitive to these factors. Although it is impossible to precisely predict
the future, the agencies need to characterize and quantify the future fleet in order to assess the
impacts of the 2022-2025 GHG standards that would affect that future fleet.  The EPA has
examined various publicly-available sources (some require purchase), and then used inputs from
those sources in a series of models to project the composition of baseline and reference fleets for
purposes of this analysis. This chapter describes this process, and the characteristics of the
baseline and reference fleets.

   The EPA has made every effort to make this analysis transparent and duplicable.  Because
both the input and output sheets from our modeling are public,1 stakeholders can verify and
check EPA's modeling results, and perform their own analyses with these.

4.1.1   Why does the EPA Establish Baseline and Reference Vehicle Fleets?

   In order to calculate the impacts of the final 2022-2025 GHG standards, it is necessary to
estimate the composition of the future vehicle fleet absent the 2022-2025 standards.  EPA has
developed a baseline/reference fleet in two parts.  The first step was to develop a "baseline" fleet.
The baseline fleet represents data from a single model year of actual vehicles sales. The EPA
creates a baseline fleet in order to track the volumes and types of fuel  economy-improving and
CCh-reducing technologies that are already present in the existing vehicle fleet.  Creating a
baseline fleet prevents the OMEGA model from adding technologies to vehicles that already
have these technologies, which would result in "double counting"  of technologies' costs and
benefits.  The second step was to project the baseline fleet sales into MYs 2022-2025. This is
called the "reference" fleet volumes, and it represents the fleet volumes (but, until later steps, not
additional levels of technology) that the EPA believes would exist in MYs 2022-2025 absent the
application of the 2022-2025 GHG standards.  For this Draft TAR, the EPA also projected the
fleet from MYs 2026-2030 though we are only showing the result out to 2025.

   After determining the reference fleet volumes, the third step is to account for technologies
(and corresponding increases in cost and reductions in CCh emissions) that could be added to the
baseline technology vehicles in the future, taking into account previously-promulgated standards,
and assuming MY2021 standards apply at the  same levels through MY2025. This step uses the
OMEGA model to add technologies to vehicles in each of the baseline market forecasts such that
each manufacturer's car and truck average CO2 levels reflect MY2021 standards. The models'
output, the "reference case," is the light-duty fleet estimated to exist in MYs 2022-2025 without
new GHG standards.  All of the EPA's estimates of emission reductions improvements, costs,
and societal impacts for purposes of this Draft TAR are developed in relation to the EPA
reference case. This chapter describes the first two steps of the development of the baseline and
                                             4-1

-------
                                                       Baseline and Reference Vehicle Fleets
reference fleets volumes. The third step of technology addition is developed as the outputs of the
OMEGA model (see Chapter 12 for an explanation of how the models apply technologies to
vehicles in order to evaluate potential paths to compliance).

4.1.2  EPA's  2014 MY Baseline Fleet

   EPA has chosen to use the final 2014 MY fleet GHG data as the basis for the baseline fleet
used in its analysis.  The 2014 MY fleet GHG data is the most recent complete set of final  U.S.
vehicle data that has actual manufacturer volumes and CCh values that is available to use in this
Draft TAR. The 2014 MY volumes and CCh values comes from the EPA VerifyA database.  The
data contained in the Verify  system is quite robust since it under goes a complex number of
quality checks  done by the manufacturer, the Verify database software, and finally EPA's
certification staff. Figure 4.1 shows the quality steps that are completed before data is available
for use in the Verify system.  The finalized 2014 GHG certification data is an accurate
representation  of vehicle and technology mix for the 2014 model year.  Estimated volumes are
also available for the 2015 model  year (CAFE midyear report data), however, EPA chose to use
the final 2014 MY data in lieu of the 2015 MY midyear estimates because the final 2014 MY
was the latest data set which had completed the entire Verify quality assurance process.  EPA's
rationale for not using the 2015  MY data is explained in more detail at the end of this section.

   The information used by EPA to develop the 2014  MY baseline  fleet includes final MY2014
GHG certification data for MY2014 model volumes, some valve train information from Wards
Automotive Group B>c, and some technology from a 2014 fleet file that was created for the
California Air Review Board (CARB) by Novation Analytics2 (formerly known as Control Tec).

   EPA will update the baseline fleet for future assessments in the MTE process to the most
recent MY for  which final data is  available for the U.S. fleet.
A manufacturer
must define all
vehicle models



A manufacturer
must define all
engine test groups
and link them to
the vehicle models



Define all test
vehicles that will
be used for
emissions testing.



Submit all test
results

              The Verify database does cross checks against all data submitted at each step.
 Label all vehicles for FE. This ties a
  vehicle, an engine, transmission,
  and driveline with a test. It also
 determines all sub configurations for
        that vehicle.
 Submit final GHG/CAFE data. This step is
  where actual volumes for each vehicle is
submitted. It is also the point where the GHG/
   CAFE standard is calculated for each
          manufacturer.
Final verification is manually done
 to ensure the manufacturers'
 calculations match the Verify
   databases calculation.
        Figure 4.1 The Verify Process for the Data EPA's MY2014 Baseline Vehicle Fleet is Based
A The EPA Verify Database is the electronic system by which vehicle manufacturers provide their compliance data
  to the EPA. There are several built-in quality assurance provisions.
B WardsAuto.com: Used as a source for engine specifications shown in Figure 4.2.
c Note that WardsAuto.com, where this information was obtained, is a fee-based service, but all information is
  public to subscribers.
                                                 4-2

-------
                                                    Baseline and Reference Vehicle Fleets
   Similar to the 2008 baseline used in the 2017-2025 GHG FRM, most of the information about
the vehicles that make up the 2014 fleet was gathered from EPA's emission certification and fuel
economy database, most of which is available to the public. (Note that a 2010 baseline was
created for the 2017-2025 GHG FRM, but it was only used for a sensitivity analysis and will not
be used for analysis in this Draft TAR).3 The 2014 GHG certification data included, by
individual vehicle model  produced in MY2014, vehicle production volume, carbon dioxide
emissions rating for GHG certification, fuel type, fuel injection type, EGR, number of engine
cylinders, displacement, intake valves per cylinder, exhaust valves per cylinder, variable valve
timing, variable valve lift, engine cycle, cylinder deactivation, transmission type, drive (rear-
wheel, all-wheel, etc.), hybrid type (if applicable), and aspiration (naturally-aspirated,
turbocharged, etc.).  In addition, the EPA augmented the 2014 GHG certification and fuel
economy database (the EPA "Verify" database) with publicly-available data which includes
valve information from Ward's Automotive Group, and data from Novation Analytics. Novation
Analytics did an analysis  of the 2014 fleet for CARB. In the process of doing their analysis they
created a detailed fleet file from publicly available sources such as manufacturer's website.
Novation Analytics" source for knowing which vehicles existed in MY2014 is EPA's
certification test car list.0

   The process for creating the 2014 baseline fleet Excel file was more complicated than in the
2012 FRM analysis.  EPA created the baseline using 2014 GHG certification  data from EPA's
Verify database. In the past the data in Verify did not include vehicle footprint data. Verify now
includes a complete set of footprint data for each vehicle, however it is separate from the GHG
information. Manufacturers are required to report the number of each vehicle produced with a
given footprint so the CCh target for that vehicle can be calculated.  Separately, manufacturers
are required to report the  number of each unique combination of vehicle, engine, transmission,
and driveline (2 wheel drive vs.  4 wheel drive) that is produced along with its measured GHG
information. The combination of the two sets of data are used to determine if a manufacturer is
complying with the  GHG standards. These two data sets  along with a data set from  Wards
Automotive, which contains engine cam information, the set from Novation Analytics, and
volume projections from both EIA's  Annual Energy Outlook (AEO) 2015 and IHS-Polk were
used to create the 2014 baseline with the reference fleet volumes. These different sets of data
had to be mapped into a single data set.  Figure 4.2 shows the process for combining the six data
sets with the result being  the completed baseline with reference fleet projections.
D The test car list is available at http:://epa.gov/otaq/tcldata.htm.
                                              4-3

-------
                                                       Baseline and Reference Vehicle Fleets
                      MY2014
                   Baseline Fleet
                 Creation Process
  2014GHG
   Emission
Certification Data
                  2014GHGFoot
                  Print Certification
                      Data
                 Wards Automotive
                   Engine Data
                  Control Tec Data
                       2022-2025
                    Reference Fleet
                       Creation
                 IHS-Polk Forecast
                                          Completed
                                     MY2014 Baseline with
                                     2022-2025 Reference
                                       Fleet Projections
                Figure 4.2 Process Flow for Creating the Baseline and Reference Fleet.
   EPA contracted IHS-Polk to produce an updated long range forecast of volumes for the future
fleet. A detailed discussion of the method used to project the future fleet volumes is in 4.1.2.1.1
of this chapter.
                                                 4-4

-------
                                                   Baseline and Reference Vehicle Fleets
   EPA used the previously mentioned data to populate input files for the OMEGA model.  The
baseline Excel file is available in the docket.4 The Data Definitions tab of the Excel file has a list
of the columns of data page with the units, definition, and source for each item that was
compiled for the baseline data.

   Table 4.1 displays the engine technologies present in the MY2014 baseline fleet. Most of the
information came from certification data with Wards' data only being used for information
regarding utilization of cam technology.
                                              4-5

-------
                             Baseline and Reference Vehicle Fleets
Table 4.1  MY2014 Engine Technology Penetration
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
FCA
FCA
Ferrari
Ferrari
Ford
Ford
GM
GM
Honda
Honda
Hyundai/Kia
Hyundai/Kia
JLR
JLR
Lotus
Lotus
Mazda
Mazda
McLaren
McLaren
Mercedes
Mercedes
Mitsubishi
Mitsubishi
Nissan
Nissan
Subaru
Subaru
Tesla
Tesla
Toyota
Vehicle Type
Both
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Trucks
Cars
Trucks
Cars
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Turbo Charged
15%
18%
10%
0%
0%
93%
100%
7%
2%
0%
0%
29%
34%
24%
2%
0%
0%
3%
6%
9%
17%
0%
0%
0%
0%
100%
0%
46%
42%
7%
0%
4%
0%
11%
3%
0%
0%
0%
Super Charged
1%
1%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
85%
83%
67%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
2%
0%
0%
0%
0%
0%
Single Overhead Cam
6%
5%
7%
0%
0%
2%
0%
6%
0%
0%
0%
0%
7%
0%
0%
57%
43%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
2%
65%
100%
0%
0%
0%
0%
0%
0%
0%
E
to
<_>
T3
to

T3
to

-------
                                                    Baseline and Reference Vehicle Fleets
Toyota
Volkswagen
Volkswagen
Volvo
Volvo
Trucks
Cars
Trucks
Cars
Trucks
0%
73%
54%
79%
45%
0%
6%
28%
0%
0%
0%
9%
0%
0%
0%
100%
91%
100%
100%
100%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
47%
34%
100%
100%
0%
0%
0%
0%
0%
0%
25%
49%
0%
0%
0%
27%
17%
0%
0%
0%
1%
0%
0%
0%
0%
84%
100%
0%
0%
   The data in Table 4.1 indicate that manufacturers have added a significant amount of engine
technology to the vehicles in the baseline (2014) fleet (as also discussed in Chapter 3.1.6). For
example, BMW stands out as having a significant number of gasoline turbocharged direct
injection engines. Most of the fleet's engines are using DOHC (dual overhead cam), and have
discrete variable valve timing (VVT). Over half of Honda's vehicles have engines with cylinder
deactivation.

   The data in Table 4.2 shows the changes between the 2014 engine technology penetrations
and the 2008 engine technology penetrations. To increase fuel economy, manufacturers applied
considerable technology between 2008 and 2014. Manufacturers increased the use of direct
injection 37 percent on cars and 28 percent on trucks. Manufacturers also increased the use of
turbo chargers 14 percent on cars and 9 percent on trucks.
                                              4-7

-------
                                   Baseline and Reference Vehicle Fleets
Table 4.2 Change (2014-2008) in Engine Technology Penetration
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
FCA
FCA
Ferrari
Ferrari
Ford
Ford
GM
GM
Honda
Honda
Hyundai/Kia
Hyundai/Kia
JLR
JLR
Lotus
Lotus
Mazda
Mazda
McLaren
McLaren
Mercedes
Mercedes
Mitsubishi
Mitsubishi
Nissan
Nissan
Subaru
Subaru
Tesla
Tesla
Toyota
Toyota
Vehicle Type
Both
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Trucks
Cars
Trucks
Cars
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Turbo Charged
12%
14%
9%
0%
0%
60%
95%
6%
2%
0%
0%
29%
34%
23%
2%
-4%
0%
3%
6%
9%
17%
0%
0%
-11%
-24%
100%
0%
44%
26%
1%
0%
4%
0%
-4%
0%
0%
0%
0%
0%
Super Charged
1%
0%
1%
0%
0%
-1%
0%
0%
0%
0%
0%
-1%
0%
0%
0%
0%
0%
0%
0%
85%
63%
-11%
0%
0%
0%
0%
0%
0%
-1%
0%
0%
0%
2%
0%
0%
0%
0%
0%
0%
Single Overhead Cam
-14%
-12%
-17%
0%
0%
-12%
0%
-15%
-39%
0%
0%
-15%
-59%
0%
0%
-7%
-14%
0%
0%
0%
0%
0%
0%
0%
-1%
0%
0%
-55%
-35%
-35%
0%
0%
0%
-69%
-70%
0%
0%
0%
0%
Dual Over Head Cam
23%
20%
28%
0%
0%
11%
0%
13%
73%
0%
0%
15%
62%
40%
-1%
7%
14%
0%
0%
0%
0%
0%
0%
1%
1%
100%
0%
53%
35%
35%
0%
-1%
0%
69%
70%
0%
0%
0%
0%
Over Head Cam
-9%
-8%
-11%
0%
0%
0%
0%
2%
-34%
0%
0%
0%
-3%
-40%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Variable Valve Timing
Continuous Intake Only
0%
-8%
12%
0%
0%
-2%
0%
9%
22%
0%
0%
-4%
-28%
-26%
64%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-72%
-35%
-100%
-38%
0%
0%
0%
0%
0%
0%
0%
0%
Variable Valve Timing
Discrete
54%
58%
50%
24%
0%
-84%
-94%
28%
69%
29%
0%
100%
99%
52%
13%
0%
0%
100%
100%
18%
42%
0%
0%
93%
87%
100%
0%
93%
67%
100%
38%
88%
95%
100%
100%
0%
0%
70%
39%
Variable Valve Discrete Lift
Only
-9%
-9%
-9%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-96%
-73%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
11%
0%
0%
-1%
-5%
0%
0%
0%
0%
Variable Valve Lift and
Timing Discrete
13%
14%
11%
-24%
0%
77%
89%
20%
3%
-29%
0%
0%
0%
12%
0%
96%
72%
0%
0%
6%
58%
0%
0%
0%
0%
0%
0%
0%
0%
0%
51%
7%
5%
0%
-23%
0%
0%
1%
0%
Vehicles without Variable
Valve Timing or Lift
-59%
-55%
-64%
0%
0%
8%
5%
-57%
-94%
0%
0%
-96%
-71%
-38%
-76%
0%
0%
-100%
-100%
-24%
-100%
0%
0%
-93%
-87%
0%
0%
-21%
-33%
0%
-62%
-95%
-100%
-99%
-73%
0%
0%
-71%
-39%
Cylinder Deactivation
4%
0%
11%
0%
0%
0%
0%
3%
18%
0%
0%
0%
0%
-1%
28%
55%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
e
_o
u
f
-------
                                                    Baseline and Reference Vehicle Fleets
Volkswagen
Volkswagen
Volvo
Volvo
Cars
Trucks
Cars
Trucks
32%
48%
31%
45%
6%
28%
0%
0%
-70%
0%
0%
0%
70%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-2%
22%
0%
0%
0%
0%
0%
0%
24%
-39%
0%
0%
-22%
17%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
   Finally, the decision to not use the mid-year report data for 2015 MY was based on several
factors.  For mid-year reports manufacturers must estimate volumes for every vehicle they can
produce instead of accounting for the vehicles they actually produce. These include powertrain
and other options that may never ultimately be produced or may be produced at significantly
different volumes than those the manufacturers initially estimated. Manufacturers may certify
these extra configurations (and estimate them as part of certification) in order to ensure they can
continuously produce at plants no matter which components are available. This practice
provides the manufacturers with a high degree of manufacturing flexibility. Table 4.3 shows the
differences between the 2015 midyear estimates and the preliminary final submission (this data
has been entered into the Verify database by Ford and marked final, but has not gone through the
manual verification process) for Ford's vehicles.  The difference between estimated and actual
penetration rates are significant enough to impact EPA's compliance pathway projections. In
addition to estimated vs. actual sales volume projections, mid-year estimates may also affect
individual vehicle fuel economy and greenhouse gas emissions performance.  Label testing is
done based on a manufacturer's high volume seller for a model.  Manufacturers often do
additional emissions testing between the initial labeling for vehicles and when final data is
submitted due to regulatory requirements for meeting CAFE and GHG standards.  Compliance
solutions that are compromised by significant differences in sales volumes can be exacerbated by
changes in individual vehicle emissions.  A different mix of vehicles will end up changing the
reported GHG for a model since GHG is production weighted based on the vehicles within each
model. These differences make using the midyear data a soft basis for projecting the future verse
the solid foundation of exact volumes and exact CCh that final reported data gives.
                      Table 4.3 MY2015 Ford Engine Technology Penetration
                                  Penetration of Turbo's and Supercharged Engines
                                Mid-Year Data
                                 Final Data
 Trucks
49.6%
40.4%
 Cars
33.2%
28.7%
 All
43.1%
35.2%
4.1.2.1 EPA's MY2014 BasedMY2022-2025 Reference Fleet

   This section provides further detail on the projection of the MY2014 baseline volumes into
the MYs 2022-2025 reference fleet. It also describes more of the data contained in the baseline
spreadsheet.

   The reference fleet aims to reflect our latest projections about the market and fleet
characteristics during the model years 2022 to 2025.  Fundamentally, constructing this fleet
involved projecting the MY2014 baseline fleet volumes out to the MYs 2022-2025.  It also
included the assumption that none of the vehicle models changed during this period. As with the
MY2008-based MY2022-2025 reference fleet used in the 2012 FRM, EPA relied on many
                                              4-9

-------
                                                   Baseline and Reference Vehicle Fleets
sources of reputable information to make these projections, yet any future fleet projections are
inherently uncertain.
4.1.2.1.1
On What Data Are EPA 's Reference Vehicle Fleet Volumes Based?
   EPA has based the projection of total car and light truck sales on the Energy Information
Administration (EIA) Annual Energy Outlook (AEO) 2015, which was the most recent
projection available by at the time our Draft TAR analysis was underway. EIA's AEO 2015 also
projects future energy production, consumption and prices.5 EIA issued the AEO 2015 on April
14, 2015.  Similar to the analyses supporting the MYs 2017-2025 rulemaking and for the 2008
based fleet projection, the EPA used the EIA's National Energy Modeling System (NEMS) to
estimate the future relative market shares of passenger cars and light trucks. However, in
NEMS, EIA models the light-duty fleet to comply with CAFE and GHG standards from 2012
through 2025 (along with the car/truck mix). In order to create a reference fleet absent the 2022-
2025 standards, we only wanted NEMS to modify the fleet up to MY2021. Therefore, for the
current analysis, EPA and NHTSA developed a new projection of passenger car and light truck
sales shares by using NEMS to run scenarios from AEO 2015 cases (reference, high, low),
holding post-2021 CAFE and GHG standards constant at MY2021 levels. The output of the
NEMS model is consistent with AEO 2015 since it has the same inputs as AEO 2015.  As with
the comparable exercise for the 2012 FRM baseline fleet, this case is referred to as the
"Unforced Reference Case," and the values are shown below in Table 4.4.  The "unforced
reference case" will be referred to as "unforced AEO 2015" for the rest of Chapter 4.1. Table 4.5
shows the originally published AEO 2015 fleet projections.
      Table 4.4 AEO 2015 Unforced Reference Case Values used in the 2014 Market Fleet Projection
Model Year
2021
2022
2023
2024
2025
Cars
8,136,376
8,143,641
8,269,894
8,410,497
8,597,413
Trucks
7,960,213
7,884,714
7,820,048
7,798,752
7,827,599
Total Vehicles
16,096,589
16,028,354
16,089,941
16,209,249
16,425,012
                          Table 4.5 AEO 2015 Reference Case Values
Model Year
2021
2022
2023
2024
2025
Cars
8,132,575
8,140,457
8,224,600
8,323,431
8,517,159
Trucks
7,964,258
7,889,725
7,864,634
7,886,273
7,911,763
Total
Vehicles
16,096,833
16,030,182
16,089,233
16,209,704
16,428,922
   In 2021, car and light truck sales are projected to be 8.1 and 7.9 million units, respectively.
While the total level of sales of 16 million units is similar to pre-2008 levels, the fraction of car
sales in 2021 and beyond is projected to be lower than some of the previous AEO projections.
                                             4-10

-------
                                                   Baseline and Reference Vehicle Fleets
   In addition, sales of segments within both the car and truck markets have also been changing
and are expected to continue to change in the future.  In order to reflect these changes in fleet
makeup, EPA used a custom long range forecast purchased from IHS-Polk Automotive (IHS
bought CSM from whom we previously purchased a long range forecast).  IHS also purchased
Polk automotive which has registration data for all the vehicles in the United States.  IHS-Polk is
a well-known industry analysis source for forecasting casting and other data. EPA decided to
use the forecast from IHS-Polk for MY2014-based market forecast for several reasons. First,
IHS-Polk Automotive  continues to use CSM's bottom-up approach (e.g., looking at the number
of plants and capacity for specific engines, transmissions, vehicles, and now registration data
from Polk) for their forecast, which we believe is a robust forecasting approach. Second, IHS-
Polk agreed to allow EPA to publish their entire forecast in the public domain.  Third, the IHS-
Polk forecast covered the timeframe of greatest relevance to this analysis (2022-2025 model
years). Fourth, it provided projections of vehicle sales both by manufacturer and by  market
segment.  Fifth, it utilized market segments similar to those used in the EPA emission
certification  program and fuel economy guide, such that the EPA could include only the
segments types covered by the light-duty  vehicle  standards.

   IHS-Polk created a  custom forecast for EPA that covered model years 2012-2030. Since the
EPA is using this forecast to generate the  reference fleet volumes for this Draft TAR (i.e., the
fleet expected to be sold absent any increases in the stringency regulations after the 2021 model
year), it is important for the forecast to be independent of increases during 2022-2025 in the
stringency of CAFE/ GHG standards.  IHS-Polk does not normally use the CAFE or GHG
standards as  an input to their model, and EPA specified that they assume that the stringencies of
the two programs would stay constant at 2021 levels in the 2022-2025 time frame for our
forecast.  This was done to eliminate the effects of the current EPA standards on the  2022 to
2025 MY fleet. In addition, EPA specified that the IHS-Polk forecast use EIAs AEO 2015 fuel
prices and economic indicators to create the forecast. IHS-Polk uses many additional inputs in
their model including GDP growth, interest rates, the unemployment Rate, and crude oil prices to
determine overall demand. They then use vehicle size, price, and function to forecast with
enough resolution to predict brand and fleet segmentation. Additional details regarding the IHS-
Polk forecast can be found in a methodology description provided  by IHS-Polk to EPA is
available in the docket EPA-HQ-OAR-2015-0827.

   The EPA combined the IHS-Polk forecast with data from other sources to create the 2014
baseline reference fleet projections.  This  process is discussed in sections that follow.

4.1.2.1.2       How did the EPA develop the 2014 Baseline and 2022-2025 Reference Vehicle
Fleet Volumes?

   The process of producing the MY2014 baseline 2022-2025 reference fleet volumes involved
combining the baseline fleet with the projection data described above.  This was a complex
multistep procedure, which is described in this section. The procedure is new and some of the
steps are different than those used with the MY2008 baseline fleet projection used in the FRM.

4.1.2.1.3       How was the 2014 Baseline Data Merged with the IHS-Polk Data?

   EPA employed a method similar to the method used in the FRM for mapping certification
vehicles to IHS-Polk vehicles.  Merging the 2014 baseline data with the 2022-2025 IHS-Polk
data required a thorough mapping of certification vehicles to IHS-Polk vehicles by individual
                                             4-11

-------
                                                   Baseline and Reference Vehicle Fleets
make and model. One challenge that the EPA faced when determining a reference case fleet was
that the sales data projected by IHS-Polk has similar but different market segmentation than the
data contained in EPA's internal database.  In order to create a common segmentation between
the two databases, the EPA performed a side-by-side comparison of each vehicle model in both
datasets, and created an additional "IHS-Polk Class" modifier in the baseline spreadsheet to map
the two datasets. The reference fleet volumes based on the "IHS-Polk Class" was then projected.

   The baseline data and reference fleet volumes are available to the public. The baseline Excel
spreadsheet in the docket is the result of the merged files.6  The spreadsheet provides specific
details on the sources and definitions for the data. The Excel file contains several tabs. They
are: "Final Data," "Data Tech Definitions, "Platforms," "VehType," "VehType(2)," "Lookups,"
"Metrics," "Machine," "MarketFile2021," and "MarketFile2025,"  "Final Data" is the tab with
the raw data.  "Data Tech Definitions" is the tab where each column is defined and its data
source named.

   In the combined EPA certification and IHS-Polk data, all 2014 vehicle models were assumed
to continue out to 2025, though their volumes changed in proportion to IHS-Polk projections.
This methodology is used to provide surrogate greenhouse gas performance data for new
emerging models. As a result, new models expected to be introduced within the 2015-2025
timeframe are mapped to existing models. Remapping the volumes from these new vehicles to
the existing models via manufacturer segments preserves the overall fleet volume.  All MYs
2022-2025 vehicles are mapped from the existing vehicles to the manufacturer's future segment
volumes. The mappings are discussed in the next section. Further discussion of this limitation is
discussed below in Section 4.1.2.1.4. The statistics of this fleet will be presented below since
further volume modifications were required.

4.1.2.1.4     How were the IHS-Polk Forecast and the Unforced AEO 2015 Forecast Used to
Project the Future Fleet Volumes?

   As with the comparable step in the MY2008 baseline 2022-2025 reference fleet process, the
next step in the EPA's generation of the reference fleet is one of the more complicated steps to
explain. First, each vehicle in the 2014 data had an IHS-Polk segment mapped to it.  Second, the
breakdown of segment volumes by manufacturer was compared between the IHS-Polk and 2014
data set.  Third, a correction was applied for Class 2B vehicles in the IHS-Polk data.  Fourth, the
individual manufacturer segment multipliers were created by year.  And finally, the absolute
volumes of cars and trucks were  normalized (set equal) to the total  sales estimates  of the
unforced AEO 2015.  This final step is required to create a fleet forecast that reflects the official
government forecast for future vehicle sales. The unforced AEO 2015 forecast alone does not
have the necessary resolution, down to the vehicle segment level, for EPA to perform its
analysis.  Therefore EPA applies both a purchased forecast from IHS-Polk and the unforced
AEO 2105 forecast to create a  complete fleet forecast.

   The process started with mapping the IHS-Polk segments to each vehicle in the baseline data.
The mapping required determination of the IHS-Polk segment by lookup at each of the 2,160
baseline vehicles in the IHS-Polk forecast (which has only 617 vehicles since they do not
forecast powertrain or footprint differences), and labeling it in the "IHS-Polk Class" column of
the baseline data. The IHS-Polk data has 52 segments. Table 4.6 has the IHS-Polk segments for
reference. Table 4.7 shows some of the Honda vehicles in the CAFE data with their "IHS-Polk
Segment" identified.
                                             4-12

-------
                                                      Baseline and Reference Vehicle Fleets
                               Table 4.6 List of IHS-Polk Segments
                                      IHS-Polk Segments
 Micro Non-premium Car
                                     Compact Non-premium Car
Mid-Size Premium Van
 Micro Non-premium Sporty
                                     Compact Non-premium MPV
Mid-Size Super Premium Car
 Mini Non-premium Car
                                     Compact Non-premium Sporty
                                      Compact Non-premium SUV
Mid-Size Super Premium Sporty
Mini Non-premium MPV
                                      Compact Non-premium Van
Mid-Size Super Premium SUV
Mini Non-premium Sporty
Full-Size Non-premium Car
 Mini Non-premium SUV
                                     Compact Premium Car
Full-Size Non-premium Pickup
 Mini Premium Car
                                     Compact Premium Sporty
Full-Size Non-premium Sporty
 Mini Premium Sporty
                                     Compact Premium SUV
Full-Size Non-premium SUV
 Subcompact Non-premium Car
                                     Compact Super Premium Sporty
Full-Size Non-premium Van
 Subcompact Non-premium MPV
                                     Compact Super Premium SUV
Full-Size Premium Car
 Subcompact Non-premium Pickup
                                     Mid-Size Non-premium Car
Full-Size Premium Sporty
 Subcompact Non-premium Sporty
                                     Mid-Size Non-premium MPV
Full-Size Premium SUV
 Subcompact Non-premium SUV
                                     Mid-Size Non-premium Pickup
Full-Size Premium Van
 Subcompact Premium Car
                                     Mid-Size Non-premium Sporty
Full-Size Super Premium Car
 Subcompact Premium MPV
                                     Mid-Size Non-premium SUV
Full-Size Super Premium Sporty
 Subcompact Premium Sporty
                                     Mid-Size Premium Car
Full-Size Super Premium SUV
 Subcompact Premium SUV
                                     Mid-Size Premium Sporty
 Subcompact Super Premium Sporty
                                     Mid-Size Premium SUV
    Table 4.7 Example of Honda Vehicles Being Mapped to Segments Based On the IHS-Polk Forecast
Manufacturer
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Name Plate
Acura
Acura
Acura
Acura
Acura
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Honda
Model
ILX
MDX
RDX
RLX
TSX
ACCORD
ACCORD
CIVIC
CIVIC
FCX
CR-V
CR-Z
CROSSTOUR
FIT
INSIGHT
ODYSSEY
PILOT
RIDGELINE
IHS-Polk Segment
Compact Premium Car
Mid-Size Premium SUV
Compact Premium SUV
Mid-Size Premium Car
Mid-Size Premium Car
Mid-Size Non-Premium Sporty
Mid-Size Non-Premium Car
Compact Non-Premium Car
Compact Non-Premium Sporty
Compact Non-Premium Car
Compact Non-Premium SUV
Mini Non-Premium Sporty
Mid-Size Non-Premium SUV
Subcompact Non-Premium Car
Compact Non-Premium Car
Mid-Size Non-Premium MPV
Mid-Size Non-Premium SUV
Mid-Size Non-Premium Pickup Truck
   In the next step, segment volume by manufacturer was compared between the baseline and
IHS-Polk data sets. This is necessary to determine if all of the segments a manufacturer will
produce in the future are currently represented by the 2014 certification data. The forecasts used
                                                4-13

-------
                                                   Baseline and Reference Vehicle Fleets
in past rulemakings predicted very few new segments for manufactures.  The new forecast from
IHS-Polk projects that manufacturers will be entering more new segments (i.e., segments they
currently do not participate in) than in previous forecasts.  This requires making sure a
manufacturers volume in the new segment be added to the volume of a manufacturers closest
existing segment.  The flow chart below (Figure 4.3) shows the process for determining this
"closest class,"  This process worked well for the majority of manufacturers with the exception
being Tesla and Aston Martin who will be entering the SUV segment in the future but in
MY2014 were currently only in the car segment.  We believe that this process of establishing
closest class surrogates provides the best estimate of the potential current performance of a given
vehicle type and the technology that will be required to meet the 2025 standards.
                                             4-14

-------
                                                              Baseline and Reference Vehicle Fleets
                                                               -Is there a Non-Premium
                                                               Segment in the same size
                                                                  and category?
                                                               /Is there a Premium
                                                                 Segment in same
                                                               category in the next siz
                                                                     arge?    ,
1s there a Premium Segment^
   in the same size and
      category?      /
  Is there a Non-Premiu
   Segment in same
  category in the next size
       arge?   /
                                                           _/ Is there a Premium Segment in
                                                            same size category is there a similar
                                                             segment? (SUV, MPV, VAN are
                                                             Similar; Car, Sporty are Similar)/
                                  /Is there a Premium Segment in
                                  same size category is there a similar
                                     gment? (SUV, MPV, VAN are
                                   Similar; Car, Sporty are Similar)/"
                                                            s there a Premium Segment in the
                                                            same category in the next smaller
                                                                     size?
                                   s there a Premium Segment in the\
                                   same category in the next smaller
                                           size?        /
                                       Try the processes again assuming that the
                                       original segment was one larger. If that does
                                        not work try the process with one smaller.
                                        Continue until starting from one larger or
                                             smaller gives a segment.
               Figure 4.3  Process Flow for Determining where Segment Volume Should Move
   Table 4.8 shows Honda's segments with their volumes for both the baseline data and IHS-
Polk.   Note that Compact Premium Sporty, Subcompact Non-Premium SUV, and Subcompact
Premium SUV segments don't exist in the baseline data. The closest classes to those are
Compact Non-Premium Car, Compact Non-Premium SUV, and Compact Premium SUV.
                                                       4-15

-------
                                                    Baseline and Reference Vehicle Fleets
   It is also important to note the difference between Model Year (MY) and Calendar Year (CY)
sales. Model Year sales can be shorter or longer than a full calendar year due to product launch
and change decisions made by a manufacturer. As a result the Model Year (MYE) sales can be
less than or greater than a respective Calendar Year sales. In Table 4.8 below a manufacturer
example is provided. For 2014 MY, Honda produced 26,689 vehicles that fell into the Compact
Non-Premium Car class.  The IHS data shows 276,287 vehicles in their CY forecast.  This is
because the baseline data represents what was built for 2014 model year in both calendar years
2013 and 2014; and, IHS-Polk data is showing the total volume for 2014 calendar year which has
both 2014 and 2015 model year vehicles represented. In this case Honda was introducing a new
Civic.  It started 2014 calendar year building 2015 model year Civics instead of continuing to
build 2014 model year Civics till June as is the usual practice. As a result, the 2014 MY vehicles
were most likely built in 2013 CY and the 2014 CY volumes reflect a large volume 2015 MY
Civics.  In years that are close to the baseline year this can be a source of error, but as years
progress, calendar year and model year volumes become the same in a forecast since models are
not added or deleted in the forecast. This allows EPA to use a calendar year forecast since we
are concerned with vehicles being built far enough in the future that calendar year and model
year volumes are approximately the same.
           Table 4.8 Example Honda 2014 Volumes by Segment from the IHS-Polk Forecast
Honda-Baseline Data
Compact Non-Premium Car
Compact Non-Premium Sporty
Compact Non-Premium SUV
Compact Premium Car

Compact Premium SUV
Mid-Size Non-Premium Car
Mid-Size Non-Premium MPV
Mid-Size Non-Premium Pickup
Truck
Mid-Size Non-Premium Sporty
Mid-Size Non-Premium SUV
Mid-Size Premium Car
Mid-Size Premium SUV
Mini Non-Premium Sporty
Subcompact Non-Premium Car


2014
MY
28,689
239,044
383,890
16,349

43,179
327,677
138,203
13,790
62,019
93,652
27,055
68,547
3,473
599


Honda-IHS-Polk Data
Compact Non-Premium Car
Compact Non-Premium Sporty
Compact Non-Premium SUV
Compact Premium Car
Compact Premium Sporty
Compact Premium SUV
Mid-Size Non-Premium Car
Mid-Size Non-Premium MPV
Mid-Size Non-Premium Pickup
Mid-Size Non-Premium Sporty
Mid-Size Non-Premium SUV
Mid-Size Premium Car
Mid-Size Premium SUV
Mini Non-Premium Sporty
Subcompact Non-Premium Car
Subcompact Non-Premium SUV
Subcompact Premium SUV
2014
CYF
276,287
49,696
335,019
17,854
0
44,865
353,508
122,738
13,389
34,866
120,659
39,447
65,681
3,562
63,305
0
0
2018 CY
327,993
30,053
299,644
15,379
797
40,642
338,848
106,887
52,244
0
144,182
44,876
53,249
10,915
54,988
73,855
23,977
Action




Move Volume to Compact
Non-Premium Sporty









Move Volume to Compact
Non-Premium Car
Move Volume to Compact
Non-Premium SUV
Move Volume to Compact
Premium
E Model Year sales may begin as early as January 1 of the previous calendaryear (MY -1).
F 2014 Calendar Year can include both 2014 and 2015 Model Year vehicle sales if both are built in the calendar
  year.
                                             4-16

-------
                                                   Baseline and Reference Vehicle Fleets
   A step that is related to the comparison step is the filtering of Class 3 vehicles from the IHS-
Polk forecast. IHS-Polk includes Class 2b and Class 3 vehicles (vans and large pickup trucks) in
its light-duty forecast.  Class 2b vans are all appropriately classified as MDPVs (Medium Duty
Passenger Vehicles) and must be included in the forecast since they are regulated under the light-
duty GHG program. Class 2b large pickup trucks, however, are not regulated under the light-
duty GHG program (rather under the medium-duty and heavy-duty fuel efficiency and GHG
programs, see 76 FR 57120), and must therefore be removed from the forecast.  Since, IHS-Polk
labels the Class 2b/3 pickup trucks with an HD, it was readily apparent which Class 2b pickup
trucks to filter from the forecast. Vans in the IHS-Polk forecast on the other hand have both
Class 2b and 3 and MDPVs in their totals and must have a correction factor applied. This is
accomplished by creating a multiplier for each manufacturer's Full-Size Non-Premium Vans and
applying it to each manufacturer's Full-Size Non-Premium Van volume every model year in the
IHS-Polk forecast; specifically, by taking a manufacturer's 2014 model year Full-Size Non-
Premium Van baseline volume and dividing its 2014 calendar year Full-Size Non-Premium Van
IHS-Polk volume. Table 4.9 shows the volumes and the resulting multiplier for FCA, while

   Table 4.10 shows the 2025 IHS-Polk volume, the multiplier and the result of applying the
multiplier to the original volume for FCA.
             Table 4.9 Example Values Used to Determine the MDPV Multiplier for FCA
Manufacturer


FCA
NEW SEGMENT


Full-Size Non-Premium Van
IHS-Polk
2014
Volume
24,840
2014 CAFE
Volume

10,485
MDPV
Multiplier

0.42
               Table 4.10 Example Values Used to Determine FCA's 2025 Van Volume
Manufacturer



FCA
NEW SEGMENT



Full-Size Non-Premium Van
Original
2025
Volume

15,074
MDPV
Multiplier


0.42
2025
Volume
after
Multiplier
6,331
   EPA next created individual manufacturer segment multipliers to be used with the individual
2014 vehicle volumes to create projections for the future fleet.  The individual manufacturer
segment multipliers are created by dividing each year of the IHS-Polk forecast's individual
manufacturer segment volume by the manufacturer's individual segment volume determined
using 2014 data.  Table 4.11 has the 2014 Volume, the 2025 IHS-Polk Full-Size Non-Premium
Van volume after Class 2b vehicles were removed, and the individual manufacturer volume for
Full-Size Non-Premium Van.  The multiplier is the result of dividing the 2025 volume by the
2014 volume.
                                             4-17

-------
                                                   Baseline and Reference Vehicle Fleets
Table 4.11 Example Values Used to Determine FCA 2025 Individual Full-Size Non-Premium Van Multiplier
Manufacturer
FCA
IHS-Polk
Segment
Full-Size
Non-
Premium
Van
2014 Cafe Volume
10,485
2025 Volume after
Multiplier
6,331
Fiat/Chrysler Individual Full-
Size Non-Premium Van
Multiplier for 2025
60.4%
   Now that the individual manufacturer segment multipliers are calculated, they can be applied
to each vehicle in the 2014 data. The segment multipliers are applied by multiplying the 2014
volume for a vehicle by the multiplier for its manufacturer and segment. Table 4.12 shows the
2014 volumes, the individual manufacturer segment multipliers, and the result of multiplying the
multiplier and the volume for 2025 project volumes for many of FCA's Full-Size Non-Premium
Van
      Table 4.12  Example Applying the Individual Full-Size Non-Premium Van Multiplier for FCA

Manufacturer




FCA
FCA
Model





Cargo Van A
Cargo Van B
IHS-Polk Segment





Full-Size Non-Premium Van
Full-Size Non-Premium Van
2014 CAFE
Volume




10,428
57
Fiat/Chrysler
Individual Full-
Size Non-
Premium Van
Multiplier for
2025
60.4%
60.4%
2025 Project
Volume Before
AEO
Normalization


6,374
34
   Normalizing to unforced AEO 2015 forecast for cars and trucks must be done once the
individual manufacturer segment multipliers have been applied to all vehicles across every year
(2011-2025) of the IHS-Polk forecast. In order to normalize a year, the number of trucks and the
number of cars produced must be determined. Then, the truck and car totals from the unforced
AEO 2015 are used to determine a normalizing multiplier.  Table 4.13 has the 2025 car and truck
totals before normalization, the unforced AEO 2015 car and truck totals in 2025, and the
multipliers which are the result of dividing the unforced AEO 2015 totals by totals before
normalization.
            Table 4.13 Example Unforced AEO 2015 Truck and Car Multipliers in MY2025
Vehicle Type
Cars
Trucks
2025 Total Before
Normalization
10,317,314
6,588,526
2025 Total from AEO 2015
8,597,413
7,827,599
2025
Normalizing
Multiplier
83%
119%
   The final step in creating the reference volumes is applying the unforced AEO multipliers.
The AEO multipliers are applied C/T type. Table 4.14 shows the normalized volume, the
unforced AEO 2015 truck multiplier for MY2025, and the final resulting volume for a number of
FCA Full-Size Non-Premium Vans.
                                             4-18

-------
                                                   Baseline and Reference Vehicle Fleets
  Table 4.14 Example Applying the Unforced AEO Truck Multiplier to FCA Full-Size Non-Premium Vans
Manufacturer




FCA
FCA
Model




Cargo Van A
Cargo Van B
C/T Type




Truck
Truck
2025 Project
Volume Before
Unforced AEO
2015
Normalization
6,374
34
Unforced AEO
2015 Truck
Multiplier for
2025

119%
119%
2025 Project
Volume with
Unforced AEO
2015
Normalization
7,585
41
4.1.2.2 What Are the Sales Volumes and Characteristics of the MY2014 Based Reference
Fleet?

   Table 4.15 and Table 4.16 below contain the sales volumes that result from the process above
for MY2014 and 2021-2025.
                             Table 4.15 Vehicle Segment Volumes
Segment

SubCmpctAuto
CompactAuto
MidSizeAuto
LargeAuto

SmallPickup
LargePickup
SmallSuv
MidSizeSuv
LargeSuv
ExtraLargeSuv
MiniVan
CargoVan
Actual and Projected Sales Volume
2014
1,031,572
2,545,441
3,538,186
479,217

12,143
1,917,061
2,012,400
1,547,977
1,053,497
664,625
602,694
68,613
2021
748,954
2,463,368
2,753,505
412,879

15,227
2,110,946
2,607,502
2,018,262
1,447,471
769,029
553,890
80,731
2022
765,720
2,433,865
2,780,716
423,053

14,222
2,061,737
2,566,936
2,005,227
1,416,403
786,535
579,944
80,598
2023
813,046
2,470,343
2,792,830
420,770

16,067
2,048,645
2,562,497
2,032,018
1,404,005
736,815
582,605
86,960
2024
813,046
2,470,343
2,792,830
420,770

16,067
2,048,645
2,562,497
2,032,018
1,404,005
736,815
582,605
86,960
2025
837,044
2,590,597
2,914,865
430,890

16,123
2,089,897
2,602,465
2,027,569
1,394,281
717,962
576,009
92,852
                             Table 4.16 Car and Truck Volumes
Vehicle Type
Cars
Trucks
Cars and Trucks
Actual and Projected Sales Volume
2014
9,206,786
6,311,548
15,518,335
2021
8,136,376
7,960,213
16,096,589
2022
8,143,641
7,884,714
16,028,354
2023
8,269,894
7,820,048
16,089,941
2024
8,410,497
7,798,752
16,209,249
2025
8,597,413
7,827,599
16,425,012
   Table 4.17 below contains the sales volumes by manufacturer and C/T type for MY2014 and
MY2021-2025.
                                             4-19

-------
                                 Baseline and Reference Vehicle Fleets
Table 4.17  Car and Truck Definition Manufacturer Volumes
Manufacturers
All
All
All
Aston Martin*
Aston Martin*
BMW
BMW
FCA
FCA
Ferrari*
Ferrari*
Ford
Ford
GM
GM
Honda
Honda
Hyundai/Kia
Hyundai/Kia
JLR
JLR
Lotus*
Lotus*
Mazda
Mazda
McLaren*
McLaren*
Mercedes
Mercedes
Mitsubishi
Mitsubishi
Nissan
Nissan
Subaru
Subaru
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
Volvo
Volvo
C/T
Type
Both
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
2014
Baseline
Sales
15,517,776
9,206,227
6,311,548
1,272
-
297,388
81,938
648,377
1,446,365
2,301
-
1,258,732
1,075,502
1,556,701
1,164,610
868,337
577,828
1,017,541
67,198
12,323
55,233
280
-
217,333
78,826
279
-
278,126
92,312
60,679
29,828
935,995
389,639
109,078
356,818
17,791
-
1,420,641
772,809
487,086
107,580
16,526
15,063
2021
Projected
Volume
16,096,589
8,136,376
7,960,213
1,324
-
298,980
110,369
607,666
1,444,140
2,255
-
935,011
1,359,683
1,211,835
1,324,550
794,566
751,770
1,109,815
159,409
24,161
103,489
234
-
249,017
108,003
900
-
226,604
159,880
47,096
29,325
767,876
559,691
134,897
473,112
86,636
-
1,132,086
1,026,564
464,804
303,810
40,612
46,418
2022
Projected
Volume
16,028,354
8,143,641
7,884,714
1,252
-
310,188
106,188
622,729
1,436,314
2,234
-
923,142
1,354,424
1,210,542
1,336,118
805,183
753,442
1,108,568
151,953
25,231
101,072
232
-
247,556
113,502
991
-
230,007
155,589
49,341
28,931
758,406
545,463
141,558
452,946
84,235
-
1,123,827
1,008,534
459,367
292,272
39,052
47,964
2023
Projected
Volume
16,089,941
8,269,894
7,820,048
1,238
-
322,601
103,272
610,278
1,442,585
2,361
-
899,877
1,329,699
1,271,586
1,279,587
817,840
761,501
1,115,024
153,506
26,015
96,894
231
-
240,049
116,282
1,120
-
240,403
152,041
53,787
30,024
786,515
529,810
138,204
482,833
92,841
-
1,132,703
1,011,496
479,608
285,503
37,614
45,013
2024
Projected
Volume
16,209,249
8,410,497
7,798,752
1,213
-
330,953
101,755
607,979
1,437,882
2,605
-
884,594
1,310,402
1,275,810
1,272,362
851,073
751,782
1,131,799
154,656
25,855
96,194
232
-
248,180
113,869
1,290
-
243,482
151,376
58,324
29,533
794,964
529,675
139,851
483,575
96,530
-
1,183,829
1,018,822
494,474
303,415
37,462
43,454
2025
Projected
Volume
16,425,012
8,597,413
7,827,599
1,345
-
324,223
101,636
622,911
1,470,099
2,735
-
929,684
1,289,230
1,287,730
1,280,168
844,715
738,106
1,154,680
157,166
25,245
95,454
233
-
259,477
114,518
1,263
-
245,341
151,199
59,327
33,126
827,952
542,008
144,187
499,218
103,502
-
1,207,430
997,624
512,191
311,139
43,244
46,908
                           4-20

-------
                                                     Baseline and Reference Vehicle Fleets
*Note: These manufacturers are shown here for reference but are not in the analysis in Chapter 12 or considered in
the ZEV sales that are part of the analysis fleet as discussed in Section 4.1.4.
   Table 4.18 also shows how the change in fleet make-up may affect the footprint distributions
over time.  The resulting data indicate that footprint will not change significantly between 2014
and 2025.
                         Table 4.18 Production Weighted Foot Print Mean
Model Year
2014
2017
2018
2019
2020
2021
2022
2023
2024
2025
Average Footprint of all Vehicles
49.7
50.0
50.1
50.1
50.0
50.0
50.0
49.9
49.9
49.8
Average Footprint Cars
46.0
46.0
46.1
46.1
46.1
46.1
46.1
46.0
46.0
46.1
Average Footprint Trucks
55.0
54.0
54.0
54.1
54.0
54.1
54.1
54.0
54.0
54.0
   Table 4.19 below shows the changes in engine cylinders over the model years. The current
assumptions show that engines shrink slightly between 2014 and 2017 and then remain relatively
constant over the 2018-2025 time frame with only a very slight shift to 4 cylinders in trucks
(may be due to an increase in small SUVs).
                 Table 4.19 Percentages of 4,6, and 8 Cylinder Engines by Model Year

Model
Year
2014
2017
2018
2019
2020
2021
2022
2023
2024
2025
Trucks
4
Cylinders
24.4%
26.7%
27.7%
28.0%
28.2%
28.2%
27.9%
28.4%
28.5%
28.7%
6
Cylinders
50.4%
51.4%
50.2%
49.9%
49.9%
50.1%
50.7%
50.4%
50.3%
49.9%
8
Cylinders
25.3%
21.9%
22.1%
22.1%
21.9%
21.7%
21.5%
21.1%
21.2%
21.4%
Cars
4
Cylinders
78.1%
78.8%
78.3%
78.4%
78.6%
78.6%
78.3%
78.4%
78.6%
78.7%
6
Cylinders
19.1%
18.4%
18.9%
18.9%
18.7%
18.6%
18.8%
18.8%
18.6%
18.5%
8
Cylinders
2.8%
2.7%
2.8%
2.7%
2.7%
2.8%
2.9%
2.8%
2.8%
2.8%
                                               4-21

-------
                                                   Baseline and Reference Vehicle Fleets
4.1.2.3 What Are the Differences in the Sales Volumes and Characteristics of the MY2008
Based and the MY2014 Based Reference Fleets?

   This section compares some of the differences between the fleet based on MY2008 data and
the fleet based on MY2014 data. The 2008 fleet projection is based on MY2008 data, a long
range forecast provided by CSM, and interim unforced AEO 2011.  The 2014 fleet projection is
based on MY2014 data, a long range forecast provided by IHS-Polk Automotive, and the
unforced AEO 2015. All tables in this section show the differences using the two fleets (2008
and 2014).

   Table 4.20, Table 4.21, and Table 4.22 below contain the sales volume differences between
the two fleets, calculated by subtracting the 2008 MY based fleet projection from the 2014 MY
based fleet projection.  The sales in MY2014 were  significantly higher (by 1,077,263 vehicles)
than in MY2008. This shows a recovery from the recession that is higher than was forecasted.

   For 2014, there is an increase in the number of compact and midsize autos, large trucks, and
all SUVs except the largest.  For 2025, one of the biggest difference between the two forecasts is
the number of cars, which in part seem to be replaced by small and midsize SUVs. The shift
from cars to trucks is due to application of the unforced AEO 2015 data while the shifts within
segments are due to the data from the IHS-Polk forecast.
                        Table 4.20 Vehicle Segment Volumes Differences
Reference Class Segment
SubCmpctAuto
CompactAuto
MidSizeAuto
LargeAuto

SmallPickup
LargePickup
SmallSuv
MidSizeSuv
LargeSuv
ExtraLargeSuv
MiniVan
CargoVan
Actual Sales
Volume
2014-2008
-265,541
584,624
446,370
-86,859

-165,354
352,618
403,602
256,647
402,787
-84,450
-116,835
35,229
Difference in Projected Sales Volume
2021
-1,787,930
-39,361
-679,286
27,223

-134,896
758,085
1,055,347
580,907
383,384
76,700
-292,166
-12,829
2022
-1,830,082
-139,737
-757,981
59,207

-132,916
726,969
1,009,846
565,592
334,685
65,578
-269,726
-11,526
2023
-1,815,047
-143,393
-889,726
52,386

-135,248
761,061
1,019,382
594,599
298,936
-11,772
-266,845
-5,960
2024
-1,888,930
-254,771
-948,257
58,527

-138,560
790,453
989,326
578,514
241,424
2,028
-248,133
78
2025
-1,944,516
-232,670
-889,366
55,876

-138,714
843,144
1,013,801
564,831
202,637
-23,134
-263,443
4,280
                                             4-22

-------
                                                    Baseline and Reference Vehicle Fleets
             Table 4.21 2014 Projection - 2008 Projection Total Fleet Volumes Differences
C/T Type

Cars
Trucks
Cars and Trucks
Difference in Actual Sales Volume
2014 - 2008
1,077,263
729,925
1,807,188
Difference in Projected Sales Volume
2021
48,868
-2,251,770
48,868
2022
-2,251,770
2,300,638
-267,915
2023
-2,472,828
2,204,913
-419,637
2024
-2,576,436
2,156,799
-577,003
2025
-2,724,393
2,147,390
-674,315
   Table 4.22 below contains the differences in sales volumes by manufacturer and C/T type
between the 2008 MY based fleet and the 2014 MY based fleet. The manufacturers with the
next largest increases in sales in 2014 MY (from 2008) are FCA, Ford, Hyundai/Kia, Nissan,
Subaru, and Volkswagen. The manufacturers with a net decrease in sales in 2014 (from 2008)
are Aston Martin, Honda, GM, Mazda, Mitsubishi, Toyota, and Volvo.  The manufacturers with
the next largest increases in sales in 2025 MY are FCA, Subaru, Tesla,  The manufacturers
forecasted to have significant net decrease in sales in 2025 are BMW, Ferrari, GM, Honda,
Hyundai/Kia, Mercedes, Mitsubishi, Nissan, Toyota, and Volvo. Table 4.22 also shows the
market down overall in MY2025 by 674,315 vehicles.
                                             4-23

-------
                                           Baseline and Reference Vehicle Fleets
Table 4.22 2014 Projection - 2008 Projection Manufacturer Volumes Differences
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
FCA
FCA
Ferrari
Ferrari
Ford
Ford
GM
GM
Honda
Honda
Hyundai/Kia
Hyundai/Kia
JLR
JLR
Lotus
Lotus
Mazda
Mazda
McLaren
McLaren
Mercedes
Mercedes
Mitsubishi
Mitsubishi
Nissan
Nissan
Subaru
Subaru
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
Volvo
Segment Type
Both
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
2014-2008
Difference
in Sales
1,807,188
1,077,263
729,925
-98
0
5,592
20,614
-54,781
489,573
851
0
302,033
261,308
-30,690
-343,187
-138,302
72,688
457,692
-45,432
2,727
-351
28
0
-29,328
22,941
279
0
69,931
13,177
-24,679
14,457
218,126
84,093
-6,957
274,272
16,991
0
163,060
-178,327
173,911
61,784
-49,123
2021
Difference
in Volume
48,868
-2,251,770
2,300,638
266
0
-60,118
-18,355
186,653
1,095,527
-4,803
0
-466,606
645,503
-352,442
-205,470
-404,314
215,854
165,141
-92,489
-34,516
45,336
-44
0
-25,723
48,775
900
0
-73,775
60,431
-18,755
-5,984
-144,753
151,662
-95,883
400,339
58,013
0
-765,818
-188,975
-163,081
143,834
-52,114
2022
Difference
in Volume
-267,915
-2,472,828
2,204,913
203
0
-49,847
-22,710
198,556
1,073,306
-4,904
0
-492,079
640,158
-368,014
-171,535
-432,321
214,207
141,503
-100,234
-34,118
42,482
-58
0
-33,595
53,195
991
0
-74,731
54,654
-17,920
-6,295
-179,042
133,579
-97,055
380,210
55,866
0
-856,205
-226,518
-176,598
134,137
-53,460
2023
Difference
in Volume
-419,637
-2,576,436
2,156,799
197
0
-37,960
-24,248
186,395
1,081,521
-4,866
0
-574,919
629,695
-334,909
-217,232
-447,724
224,604
137,970
-103,371
-34,625
38,029
-68
0
-56,861
54,315
1,120
0
-72,104
46,727
-13,893
-5,445
-167,825
112,689
-103,408
409,812
64,691
0
-898,122
-213,484
-160,301
120,206
-59,225
2024
Difference
in Volume
-577,003
-2,724,393
2,147,390
72
0
-57,241
-44,771
181,963
1,092,920
-4,836
0
-619,077
621,548
-360,995
-221,235
-456,778
214,788
122,208
-107,554
-37,873
38,213
-77
0
-52,435
51,899
1,290
0
-88,855
44,292
-12,404
-6,468
-187,807
107,458
-108,432
409,433
65,668
0
-890,407
-189,191
-156,657
135,067
-61,720
2025
Difference
in Volume
-674,315
-2,817,863
2,143,548
163
0
-81,033
-43,772
186,432
1,138,337
-4,924
0
-610,426
604,754
-386,206
-243,840
-495,606
180,409
114,647
-108,623
-40,173
38,648
-83
0
-47,327
53,150
1,263
0
-95,378
50,132
-13,978
-3,260
-186,824
115,554
-112,784
424,496
71,529
0
-894,262
-212,392
-165,029
145,636
-57,863
                                    4-24

-------
                                                     Baseline and Reference Vehicle Fleets
    Volvo
Trucks
-17,685      4,650       6,278       2,982
993      4,319
   Table 4.23 shows the difference in footprint distributions between the 2014 based fleet
projection and the 2008 based fleet projection.  The differences between MYs 2014 and 2008 are
small and are just the result of the manufacturers' product mix in those model years. MY2025
shows an increase in both the average truck and average car footprints. This is due to the
significant decrease in subcompact cars forecast in the 2014 based fleet projection. Because the
total numbers of cars and trucks differs, production weighting can affect the average for the
whole fleet as compared to the averages for cars and trucks. This can cause a counterintuitive
result when taking the difference of the averages.
      Table 4.23 2014 Projection - 2008 Projection Production Weighted Foot Print Mean Difference
Model
Year
2014-2008
2017
2018
2019
2020
2021
2022
2023
2024
2025
Difference in Average Footprint
of all Vehicles
49.7-48.9 = 0.8
50.0-48.3 = 1.7
50.1-48.1 = 2.0
50.1-48.0 = 2.1
50.0 - 48.0 = 2.0
50.0 - 48.0 = 2.0
50.0-47.9 = 2.1
49.9-47.9 = 2.0
49.9-47.7 = 2.2
49.8-47.7 = 2.1
Difference in Average
Footprint Cars
46.0-45.4 = 0.6
46.0-44.9 = 1.1
46.1-44.9 = 1.2
46.1-44.9 = 1.2
46.1-44.9 = 1.2
46.1-44.9 = 1.2
46.1-44.9 = 1.2
46.0-44.9 = 1.1
46.0-44.9 = 1.1
46.1-44.9 = 1.2
Difference in Average Footprint
Trucks
55.0-54.0 = 1.0
54.0-53.8 = 0.2
54.0-53.7 = 0.3
54.1-53.6 = 0.5
54.0-53.7 = 0.3
54.1-53.6 = 0.5
54.1-53.6 = 0.5
54.0-53.5 = 0.5
54.0-53.4 = 0.6
54.0-53.3 = 0.7
   Table 4.24 shows the difference in engine cylinders distribution between the 2014 MY based
fleet and the 2008 MY based fleet. MY2014 has fewer vehicles with 6 and 8 cylinder engines
than MY2008 did. Fewer 6 and 8 cylinders in the baseline fleet along with vehicle mix changes
results in more 4 cylinder engines in trucks and cars by 2025.
                                              4-25

-------
                                                   Baseline and Reference Vehicle Fleets
          Table 4.24 Differences in Percentages of 4,6 and 8 Cylinder Engines by Model Year

Model
Year
2014-2008
2017
2018
2019
2020
2021
2022
2023
2024
2025
Trucks
4 Cylinders
13.8%
15.6%
16.9%
17.4%
17.7%
17.7%
17.5%
18.0%
17.9%
18.0%
6 Cylinders
-5.2%
-11.8%
-13.8%
-15.0%
-15.2%
-15.7%
-15.6%
-16.9%
-17.3%
-17.8%
8 Cylinders
-8.7%
-3.8%
-3.1%
-2.4%
-2.6%
-2.0%
-1.9%
-1.2%
-0.6%
-0.2%
Cars
4 Cylinders
20.4%
16.8%
16.2%
16.3%
16.8%
16.5%
15.8%
16.0%
16.1%
16.1%
6 Cylinders
-17.8%
-14.5%
-14.0%
-14.0%
-14.5%
-14.3%
-13.8%
-14.0%
-14.1%
-14.0%
8 Cylinders
-2.5%
-2.2%
-2.2%
-2.3%
-2.3%
-2.1%
-1.9%
-2.0%
-2.1%
-2.1%
4.1.3   Relationship Between Fuel Economy and Other Vehicle Attributes

   The previous discussion has described the EPA baseline fleet of MY 2014 vehicles, and
development from that baseline fleet to the reference fleet — the projection of the vehicle fleet to
MY 2022-2025 if the standards remained at the MY 2021 standard levels.  Also as discussed
above, EPA's reference fleet assumes that, while relative production volumes will continue to
evolve through 2025, all characteristics of individual vehicle models and configurations (except
GHG emissions and fuel economy driven by the standards) will remain unchanged through 2025.
In other words, for purposes of assessing the regulatory impacts analysis of the MY 2022-2025
standards, and for properly accounting for the cost of the additional technology required to meet
those standards, EPA is making the modeling  assumption that added technology will be used to
reduce greenhouse gas emission and not to improve vehicle performance and utility. EPA used a
similar approach in the 2012-2016 standards rule and the 2017-2025 standards setting rule.
Manufacturers may choose to apply technology to improve vehicle performance in lieu of
efficiency and that could result in higher costs than projected in this analysis. This section
provides a discussion of that assumption.

   For the Draft TAR analysis, EPA is assuming that the MY 2022-2025 reference fleet will
have GHG emissions performance equal to that necessary to meet the MY 2021 standards (in
effect a  "flat" reference fleet). This is consistent with the assumption used in the MY 2017-2025
rulemaking,  where EPA presented a detailed rationale for assuming that there would be no
decrease in fleetwide GHG emissions performance in the reference case fleet for MY 2017-2025
beyond the GHG emissions performance necessary to meet the MY 2016 standards.7 Key
elements of the rationale were:  1) projections that gasoline prices would be relatively stable out
to 2025, 2) historical evidence that during periods of stable gasoline prices and fuel economy
standards, the only companies that typically over-complied with fuel economy standards were
those that produced primarily lighter vehicles  that inherently over-complied with the older
universal (one size fits all) fuel economy standards, 3) that under increasingly stringent footprint-
based GHG and fuel economy standards for the five years from MY 2012-2016, it was likely that
                                             4-26

-------
                                                    Baseline and Reference Vehicle Fleets
most major manufacturers would be constrained by the standards and unlikely to voluntarily
over-comply, and 4) if there were individual manufacturer over-compliance in a reference case
scenario, that manufacturer would likely generate credits that could be sold to other companies,
and therefore not lead to fleetwide over-compliance.

   EPA believes that the case for a flat GHG reference case fleet is even stronger for the MY
2022-2025 timeframe for the following reasons:  1) gasoline prices are about $1 per gallon lower
today than in October 2012 when the MY 2017-2025 final rule was published,  2) AEO 2015
projections for fuel prices in the MY 2022-2025 timeframe are relatively stable and
approximately $1 per gallon lower than the AEO 2012 Early Release projections upon which we
relied in the final rulemaking analysis, 3) another five years of increasingly stringent footprint-
based GHG and fuel economy standards under the National Program (i.e., the MY 2022-2025
reference case fleet must meet the MY 2021 standards, five years later than the MY 2016
standards that were the basis for the MY 2017-2025 reference case fleet) that will have led to
significant commercialization of new technologies, and 4) due to the additional five years of
increasingly stringent standards, credits generated in the MY 2022-2025 timeframe are likely to
be even more valuable, and even more likely to be sold, than in the MY 2017-2021 timeframe.
For all of these reasons, EPA believes that it is very unlikely that there would be any market-
driven decrease in fleetwide GHG emissions performance in a MY 2022-2025 reference case
fleet. In addition, the National Research Council8 in its 2015 report states that assuming
equivalent performance in the fleet "is equivalent to a reference case with no further technical
change in the vehicle market from 2017 to 2025." This, it states, is inconsistent with past trends,
where "the rate of technological progress in vehicle attributes and efficiency has been strong and
continual over the past 30 years." From the 1980s to about 2005, as described in Chapter 3.1.5,
horsepower and weight increased steadily, while fuel economy was either stable or declining.
The NRC suggests developing a reference case that reflects technological progress over time,
and its possible allocation to horsepower  and weight, rather than assuming equivalent
performance. Specifically, the NRC recommends:

   "Recommendation 10.7: The agencies should  consider how to develop a reference case for the
analysis of societal costs and benefits that includes accounting for the potential opportunity costs
of the standards in terms of alternative vehicle attributes forgone."9

   The analysis of the MY 2022-2025 standards would begin with that reference case, containing
vehicles with new and different vehicle characteristics. The cost and effectiveness analysis
would involve adding technologies to those new vehicles, either holding those new vehicles'
characteristics constant or explicitly acknowledging changes in those characteristics to achieve
the standards.

   The technological innovation referred  to by the NRC has been an ongoing process in the auto
industry. Several recent studies,10 discussed in Section 4.1.3.1 below, have sought to estimate
the magnitude of innovation by calculating the relationship between power, fuel economy, and
weight each year.  Over time, if it is possible to have more fuel economy for a constant amount
of power and weight  (or more power or weight for constant fuel economy), those studies define
that increase as innovation.  Similarly to Chapter 3.1.5, these studies argue that most of that
innovation has in the past gone into improvements in vehicle power.  The authors expect that the
vehicle GHG and fuel economy standards are instead directing that innovation toward fuel
economy. As a result, because technological innovation has not been directed toward power,
                                             4-27

-------
                                                     Baseline and Reference Vehicle Fleets
vehicles in the reference case must be less powerful than they would be in the absence of the
standards.  Thus, such studies would suggest that the reference case should be revised to project
that power would have been higher; if vehicles subject to the standards do not achieve that new
reference-case level of power, then the agencies should account for the opportunity cost of the
forgone power.

   In contrast, a working paper from Cooke11 argues that the reference case should not include
these increases in power or other attributes, because the agencies are not required to do more
than preserve the baseline attributes. Cooke argues that increases in power or other vehicle
attributes are optional to manufacturers, and thus not the responsibility of the agencies. If those
technologies were instead applied to vehicle performance or other attributes rather than fuel
economy, and it then becomes more expensive to meet the standards, Cooke argues that that
increase in costs is not attributable to the standards.

   EPA expects that manufacturers will continue to consider ways to improve vehicle utility and
performance, and the potential for tradeoffs between reducing GHG emissions and improving
other vehicle attributes deserves consideration. In principle, methods such as those used in the
studies discussed in Section 4.1.3.1  could be used to develop a reference case that would include
the potential for improvements over time in vehicle attributes or other attributes associated with
improving fuel economy.0 In practice, though, estimating these effects and their magnitudes
involve a number of complexities, including challenges in estimating the tradeoffs and the
innovation likely to occur in the absence of the standards, the role of the standards in promoting
innovation, and the potential for ancillary benefits associated with GHG-reducing technologies.

   The remainder of Chapter 4.1.3 describes these complexities in more detail.  Chapter 4.1.3.1
focuses on the estimation process mentioned above, for the tradeoffs between fuel economy,
power, and weight, and for the measures of innovation.  The magnitudes of both the tradeoff
estimates and the innovation estimates may not yet be known with confidence. The literature
does point to an important aspect of the standards, though: they may increase the amount of
innovation over the reference-case level.  Chapter 4.1.3.2 examines this question more closely.
In particular, it draws on the literature on innovation to distinguish between "incremental,"
small-scale innovation, and "major" innovation. It proposes a thesis that incremental technology
is likely to be what would happen in the absence of the standards, while the standards may
trigger major technology.  If so, both the benefits and the costs of major innovation are
associated with the standards. If incremental innovation can happen irrespective of the standards
- that is, the benefits and costs of incremental innovation are unaffected by the  standards - then
the only tradeoffs important for the  standards are those associated with major innovations.
While Chapter 6.4.1.2 discusses recent EPA research exploring whether there are possible
adverse effects of fuel-saving technologies, Chapter 4.1.3.3 points out that some of these
technologies have ancillary benefits. Finally, Chapter 4.1.3.4 discusses how EPA might evaluate
the impact of the standards on other vehicle characteristics in the benefit-cost analysis.
 As discussed in the Guidelines for Preparing Economic Analyses (U.S. Environmental Protection Agency, 2014,
  MtBi^Aioscmitejiajj^^                                Chapter 5), the baseline (referred to in this
  chapter as the reference case) "is defined as the best assessment of the world absent the proposed regulation or
  policy action." In other words, the analysis should take into account that change is likely to happen even without
  the regulation or action.
                                               4-28

-------
                                                     Baseline and Reference Vehicle Fleets
4.1.3.1 Recent Studies of the Engineering Tradeoffs between Power and Fuel Economy, and
Increases in Innovation

   The recent  studies12 that estimate both technological improvements over time in the auto
industry, as well as the engineering tradeoffs among fuel economy, power, and weight (and
sometimes other characteristics) have much in common with each other.  They all estimate an
equation roughly of the form,

   ln(fuel economy) = PO + pl*ln(horsepower) + P2*ln(weight) + p4*Other Characteristics + s,

   where:

         In refers to the natural logarithm of the term in parentheses,

         PS are coefficients to be estimated in the statistical analysis (and measure elasticities of
      fuel economy with respect to its associated variable)

         s is an error term

   They differ in the additional vehicle characteristics that they include in the regressions, and in
their ways of measuring technological change. Estimates of the elasticities of fuel economy with
respect to horsepower - that is, the engineering tradeoffs between fuel economy and horsepower
- include values from -0.16 (Klier and Linn) to -0.32 (Knittel 2012); the elasticities between fuel
economy and weight include values from -0.336 (Klier and Linn 2016) to -0.521 (MacKenzie
andHeywood2015).H

   Regarding measures of technological change, Knittel (2011) and MacKenzie and Hey wood
(2015) use annual shifts in the tradeoff curves; Klier and Linn (2016)  use engine redesign cycles
for individual  vehicles; and Wang (2016) uses a time trend and the level (stringency) of fuel
economy standards. The papers all find technological innovation, defined as an increase over
time in fuel economy not explained by  changes in horsepower, weight, or other characteristics, to
be ongoing. Knittel (2011) finds truck and car efficiency to have increased about 50 percent
from  1980 to 2006, with innovation higher before 1990 than in subsequent years. MacKenzie
and Heywood find that efficiency measured using horsepower and weight increased about 50
percent from 1975-2009, but nearly 60 percent using acceleration and weight;  using acceleration,
features, and functionality led to an estimate of 70 percent improvement. Klier and Linn (2016)
find that technological innovation varies with the stringency of predicted standards and with the
enactment of new standards but do not provide estimates of the magnitudes of baseline
innovation. Wang (2016) finds that cars innovated 1.19 percent per year, and trucks 0.66 percent
per year; a 1 percent increase in CAFE standards led to an additional increase of 0.32 percent in
innovation for cars, and 0.62 percent for trucks.  These last two studies argue that GHG and fuel
economy standards increase technological innovation above levels without regulation.
H The papers include multiple specifications: they may include different regressions for different vehicle classes, a
  variety of additional covariates, or different functional forms. Some of the studies include torque or zero-to-60
  times instead of or in addition to horsepower. The values given here are from comparing preferred specifications
  specifically using horsepower and weight. The values in different specifications include values within and outside
  these ranges; the ranges cited here thus potentially understate the variation in point estimates.


                                              4-29

-------
                                                    Baseline and Reference Vehicle Fleets
   MacKenzie and Heywood (2015) raise questions with the approach adopted by many of these
studies (focusing on Knittel 2012).  In particular, they argue that horsepower and weight are not
necessarily good proxies for characteristics that consumers want, and that estimates both of the
tradeoffs of these characteristics with fuel economy and of technological change are sensitive to
the additional vehicle characteristics considered in the regressions.

   If horsepower and weight are not themselves of primary  interest to vehicle buyers, then,
according to MacKenzie and Heywood, the measured tradeoffs of horsepower or weight for fuel
economy do not measure changes in metrics important to consumers. Horsepower, for instance,
does not by itself measure the full range of performance-related attributes, which include other
features such as low-end torque, handling, and acceleration.  MacKenzie and Heywood (2012)13
find that acceleration performance in 2010 is 20 to 30 percent faster than comparable vehicles in
the 1970s;1 in other words, horsepower is not directly proportional to acceleration.  Because
acceleration is likely to be of more importance to consumers than horsepower itself, the tradeoff
for horsepower identified in these analyses may not accurately measure impacts important to
consumers.

   Similarly, it is unlikely that consumers care directly about vehicle weight; rather, they are
probably more interested in size,  safety, cargo capacity,  or other characteristics that are
imperfectly correlated with weight. In these studies, a large vehicle with significant mass
reduction and improved fuel economy would show up in the data to have the same attributes as a
small efficient car, though consumers would view them very differently.

   The use of weight and horsepower in these regressions may also bias the estimates of
technological innovation. In these studies, technological innovation is measured as a residual
improvement in fuel economy after other factors that influence fuel economy are considered.
Including a characteristic (including but not limited to horsepower and weight) in the regressions
means that technological innovation will not affect that characteristic; its fuel economy elasticity
is fixed.  MacKenzie and Heywood (2015) show this effect  by using horsepower in their analysis
in one regression, and acceleration (O-to-60 time) in other regressions.  When they use
acceleration instead of horsepower, the amount of technological innovation due to the
relationship between power and acceleration ends up included in their measure of innovation;
that addition increases the estimated level of technological innovation.  They also point out that
technological change to reduce weight will not show up as innovation in these other papers,
because, as mentioned above,  a large vehicle with mass reduction and improved fuel economy
looks in the data like a small, efficient car rather than a vehicle with advanced technology/

   The measures of technological change are also sensitive to the other characteristics used in the
regressions. For instance, Knittel (2011) and Klier and Linn (2016) both include powertrain
types as additional characteristics. By assumption, then, powertrain types are not innovations, or
subject to innovation; a hybrid or diesel will not become more (or less) efficient relative to a
1 They attribute this change to improvements in the transformation of engine power to acceleration.
1 In their paper, MacKenzie and Heywood separately apply an adjustment to account for innovations in weight
  reduction.
                                              4-30

-------
                                                     Baseline and Reference Vehicle Fleets
gasoline vehicle over time.K MacKenzie and Heywood (2015) argue that an analysis should not
include those factors because "shifts toward more inherently efficient powertrain technologies
are themselves a part of the overall process of technology change, so it is desirable to capture
their contributions to overall efficiency in the year fixed effects" that measure innovation (p.
922).

   It is also possible that the estimates for the relationships between fuel economy and other
attributes from these studies may not represent pure technology tradeoffs,  and may therefore be
biased.  Manufacturers do not produce vehicles with all possible combinations of horsepower,
fuel economy, and weight; instead, the vehicles they produce include a mix of those
characteristics that the companies believe consumers prefer. MacKenzie and Heywood (2015)
find that accounting for a vehicle's specific power relative to the specific power of other vehicles
in the fleet (the quintile of specific power) affects fuel economy, as well as the responsiveness of
fuel economy to acceleration or weight.  If these tradeoff curves were purely about technological
relationships, they would not be affected by whether a vehicle was relatively powerful, but only
by its absolute power.  They suggest that "the relative sophistication of a vehicle's engine
(compared to others in the same model year) is correlated with weight and acceleration
performance; new technologies are not applied uniformly across all vehicles" (p. 922).  As a
result, the tradeoff estimates may not represent strictly technological tradeoffs, but also
manufacturer choices that potentially bias tradeoff estimates.

   Based on MacKenzie and Hey wood's (2015) work, then, these other studies may not
accurately measure tradeoffs involving characteristics of interest to vehicle owners.  Weight, for
instance, is unlikely to matter to consumers, except if that weight comes from size or added
features. In other work (MacKenzie and Heywood 2012), in which they focus on the
relationship between horsepower and acceleration, they question whether improvements in
acceleration are going to continue indefinitely; they find that trends in O-to-60 time are consistent
with decay toward an asymptote, and that vehicles in 2010 were within 1 second of the O-to-60
time asymptotic level.L It is not known if this slowdown in acceleration improvements is due to
physical limits or limits in consumer interest.

   Although MacKenzie and Heywood's analysis presents a more detailed discussion of these
issues compared to the other studies examined here, it is not clear that it is suitable for
quantitative development of a new reference case. First,  even O-to-60 time as a measure of
acceleration may be too narrow a criterion for evaluating performance. Performance, as a
consumer experiences it, is a complex combination of multiple characteristics including initial
launch, ability to pass another vehicle at highway speeds, handling, and cornering. Second, Klier
and Linn (2016) and Wang (2016) suggest that the rate of technological innovation is affected by
the level of the standards. MacKenzie and Heywood's analysis does not examine this effect.
Because of the possibility of a downward bias in innovation from those two  studies, their
estimates of innovation are not likely to be sufficient. In addition, the standards for MY2012-
K Interacting the characteristic with a measure of time allows for innovation specifically in that characteristic; for
  instance, Knittel interacts the manual transmission variable with a time trend, which allows the fuel consumption
  of a manual transmission relative to an automatic transmission to vary over time. These papers have few such
  interactions; this is the only one in Knittel (2011).
L They present the analysis, not only for an average vehicle, but also for vehicles in the fifth and ninety-fifth
  percentiles for acceleration. They all show this flattening.


                                               4-31

-------
                                                    Baseline and Reference Vehicle Fleets
2025 are more significant in magnitude than any changes since the introduction of CAFE in the
late 1970s; it is likely that innovation currently underway in the auto industry is of a different
magnitude and kind than in the past. As a result, estimates of innovation from any of these
studies may not be applicable to what is currently happening in the auto industry.

4.1.3.2 The Role of the Standards in Promoting Innovation

   As discussed above, some authors point to the role of standards in promoting innovation.
This section discusses how innovation may be induced by the standards, and how this innovation
should be  viewed differently in accounting for opportunity costs than innovation that may have
occurred in the absence of the standards.

   There is a wide body of literature concerning technological change in general.14  The process
of technological change can be divided into three stages: invention, where a new product or
process is  first developed; innovation, where the product or process is first commercialized; and
diffusion,  where the product or process is widely adopted throughout an industry.  This can be a
challenging process: most inventions never make it to the innovation stage;15 if they are
introduced by a small number of initial adopters, many technologies never diffuse and thus
ultimately fail.16

   It is generally agreed that innovation - the first commercialization of a new product - occurs
on a continuum between two extremes: "major" innovation where product characteristics change,
and "incremental" innovationM which exploits relatively minor changes to the existing product.17
Although accurately and completely categorizing innovation may be more complex than
applying a simple one-dimensional continuum (as Henderson and Clark (1990) claim), the one-
dimensional model does offer some insight into how industries implement innovation.

   A good example of a major innovation, and the role of environmental regulations in spurring
technology diffusion, is gasoline direct injection (GDI).  Mercedes introduced a four-stroke GDI
engine into production in 1955.18 Nonetheless, in 2008, prior to the establishment of the
MY2012-2016 standards, only 2 percent of vehicles used gasoline direct injection.19 By 2014,
this number had risen to 38 percent, with a rate of adoption in 2011 - 2014 of 7 to 8  percentage
points per year. This changeover shows a major innovation, based on previous inventions,
moving from invention to innovation and eventually to diffusion only when stimulated by
emissions  standards.

   As in the GDI example, major innovation does not necessarily proceed immediately (or at all)
to diffusion for all promising technologies. In the absence of a forcing mechanism such as
regulation, risk-averse manufacturers may prefer smaller, incremental innovations.20 There are
multiple reasons why manufacturers may prefer incremental innovation to major innovation,
particularly the risk and uncertainty associated with major innovations.

   When a company implements a major innovation, the development costs may be high and the
market impacts uncertain.  This results in a first-mover disadvantage (see also Chapter 6.3),
where a pioneer company fronts the bill to test out a new technology.  In doing so, it may briefly
capture the market, but this allows all other companies to learn about the true demand for the
M Abernathy and Utterback use "major" and "incremental" Henderson and Clark, with a two-dimensional
  framework, use "radical" and "incremental."
                                              4-32

-------
                                                      Baseline and Reference Vehicle Fleets
technology without themselves facing any risk.21 Consumer response to the first mover may
give the second mover valuable information about market acceptance. There are, therefore,
incentives to delay the development or adoption of a new technology until a competitor has
already proven that the technology is profitable. If all producers wait for another one to
implement the innovation, the innovation will never enter the market at all.

   In addition, Popp et al.22 point out that there could be "dynamic increasing returns" to
adopting some new technologies, wherein the value of a new technology may depend on how
many other companies have adopted the technology.  This could be due to network effects or
learning-by-doing. In a network effects  situation, the usefulness of the technology depends on
adoption of complementary components - for instance, the value of switching to a new fuel
depends on the infrastructure available for providing that fuel, and the value of the infrastructure
depends on the number of vehicles using the new fuel.  Learning by doing (see also Chapter
5.3.2) is the concept that the costs (benefits) of using a particular technology decrease (increase)
with use.  Both of these incentivize firms to pursue a "wait and see" strategy when it comes to
adopting new technologies.

   Finally, fixed costs and switchover disruptions23 delay technology adoption.  Firms often face
major problems in integrating new technologies resulting from major innovations into their
products;  in some cases, they may temporarily reduce  output.

   First-mover disadvantage, dynamic increasing returns, fixed costs, and switchover disruptions
all create barriers to major innovation. Incremental innovations typically face less of these
problems. Thus, in the absence of a driving factor such as regulation, manufacturers are likely to
choose incremental innovations over major innovation.N

   Both scientific research24 and popular press25 suggest that the current CAFE and light duty
GHG standards drive innovation.  The mechanism by which the standards  affect innovation is
the reduction of the barriers to manufacturers for applying major innovation to new vehicles.0

   Since all manufacturers are required to comply with regulations on the same time schedule,
and the technological pace required often outstrips that obtainable by incremental innovation
alone, manufacturers are assured that their competition is likely to implement major
technological innovations simultaneously. Thus, instead of the first-mover disadvantage, there is
a regulation-driven disincentive to "wait and see." It should be noted that companies differ both
in the degree of effort that they face due to the standards, and in the strategies that they choose in
N This discussion is not intended to imply that major innovation will not happen in the absence of regulation. Many
  factors affect the likelihood of a technology proceeding from invention through to widespread dissemination,
  including some degree of luck in having the right invention at the right place at the right time with support from
  key stakeholders.
0 The U.S. Department of Energy's Advanced Technology Vehicles Manufacturing (ATVM) Loan Program
  provides an example of another mechanism to reduce these barriers. The ATVM provides long-term, low-interest
  rate loans to support the domestic manufacturing of advanced technology vehicles and automotive components. It
  can finance a wide range of project costs, including the construction of new manufacturing facilities; retooling,
  reequipping, modernizing, or expanding an existing facility in the U.S; and the engineering integration costs
  necessary to manufacture eligible vehicles and components. It is designed to ensure that rising fuel economy
  standards do not disadvantage domestic manufacturing. With more than $16 billion in remaining loan authority,
  the ATVM program can provide financing to support the manufacturing of fuel-efficient technologies and
  components. See hltj)^Avwv.cnerj>yj>ov/lEO/alv^ for more information.


                                               4-33

-------
                                                    Baseline and Reference Vehicle Fleets
response.  Nevertheless, the benefits of generating (or avoiding the need for) credits suggest that
all companies have incentives to pursue major innovations.  In addition, there can be synergies
from companies (including suppliers) working on the same technologies at the same time.26

   Because of the global nature of the auto industry, it is likely that innovations from U.S.
regulations are likely to affect vehicles in other countries, and regulations from other countries
are likely to affect U.S. vehicles.  Because technologies to reduce GHG emissions do not need to
be reinvented for each country, the fixed costs of innovation can be spread over a global market.
It is even likely that many of these technologies will be used in countries without GHG
standards, due to the use of common manufacturing platforms across countries and to the
ancillary benefits associated with many of these technologies.

   Developing a revised reference case could entail estimating incremental technological change,
and projecting vehicle attributes resulting from that innovation, in the absence of the standards.
Developing the control case - the case with the standards in place - could then entail estimating
major technological change induced by the standards and projections of vehicle characteristics
using that greater innovation.  The discussion above suggests that conducting such  an analysis
may involve inaccurate estimates of the amount of innovation both in the absence of and in the
presence of the  standards, and may provide inaccurate estimates of the consequences of this
innovation for specific vehicle characteristics.

   Rather than assume a control case with "equivalent performance" to the baseline, one
approach could involve assuming a control case with "equivalent performance" to the reference
case. Since innovations in the reference case are incremental, such an approach could define, not
the reference and control case performance specifically, but rather the difference between them.

   In the reference case, it could be assumed that manufacturers would improve vehicle
attributes consistent with historical trends due to the implementation of incremental innovations.
Some of these changes might affect additional implementation of GHG/fuel economy
technologies; in other cases (for example, infotainment systems, automobile connectivity, or
active safety systems), the standards have no technical interaction with those changes.

   In the control case, it could be assumed that the standards induce major technological
improvement used to improve fuel economy.  Incremental technological improvement would still
be used to improve other vehicle attributes at the same pace as exhibited in the reference case.
Thus, the differences between the control and reference cases  are both the existence of fuel
economy targets and the availability of major technological innovations (in addition to
incremental innovations).

   It should be noted that there is neither the requirement nor expectation that manufacturers
allocate major innovations solely to fuel economy improvement and incremental innovations
solely to other vehicle attributes.  The standards give manufacturers the flexibility to choose
what technologies to apply to which vehicle, when to apply them, and the use of each individual
technology.  If major innovations driven by the GHG/fuel economy standards were used to
enhance these other attributes, though, it should be noted that these other attributes would not
have been enhanced in the absence of the standards; those enhancements are ancillary benefits of
the standards.

4.1.3.3 Potential Ancillary Benefits of GHG-Reducing Technologies
                                              4-34

-------
                                                    Baseline and Reference Vehicle Fleets
   Yet another complication associated with assessing an appropriate reference case is the
potential existence of ancillary benefits of GHG-reducing technologies.  These can arise due to
major innovation enabling new features and systems that can provide greater comfort, utility, or
safety.1" The studies discussed above all assume that, other than through innovation, improving
fuel economy reduces power or weight, and thus imposes opportunity costs; and innovation can
be channeled only to fuel economy, weight, or some single-dimensional measure of
performance, such as 0-60 acceleration. When performance is characterized more broadly as a
combination of multiple characteristics, it will often not be possible to strictly maintain
performance along every dimension with the application of technological innovations.  For
example, a new technology may have unequal effects on the various measures of acceleration
performance, so that an attempt to maintain performance along one dimension by resizing the
vehicle powertrain will result in an increase or decrease along other dimensions.  In addition,
some technologies provide ancillary benefits that improve vehicle performance and utility along
dimensions that are unrelated to acceleration and powertrain sizing.  In such cases, the
technologies implemented to reduce GHG emissions enhance other vehicle characteristics,
providing entirely new capabilities and desirable features or  resulting in lower costs for these
features than would be otherwise possible.

   Some examples of the potential ancillary benefits of GHG reducing technologies are listed
here.

       •  Mass reduction can provide benefits of improved braking and handling performance,
          and on towing vehicles can enable additional towing and hauling capability with same
          or similar engine sizing.
       •  Mass reduction achieved through material substitution from non-ferrous metals
          provides greater corrosion resistance.
       •  Accessory Load reductions achieved through the  use of pulse-width modulation
          (PWM) on accessory motors for HVAC  blower fan speeds provide the benefit of
          improved durability.
       •  Air conditioning system improvements achieved through variable displacement
          compressors which adjust automatically  rather than shutting off completely provide
          the benefit of smoother compressor transitions and less noise.
       •  Advanced transmissions with wider overall gear ratios and lower 1 st gear ratios
          provide the benefit of improved launch feel.
       •  Electric power steering (EPS) systems enable automakers to implement customer
          features that utilize automatic steering such as automatic parking features, or trailer
          hitch connection assistance.
       •  EPS systems also provide the capability  for variable ratio steering  systems which
          allow greater steering responsiveness close to center, and reduced effort at large
          steering angles, while also reducing the lock-to-lock turns.
       •  Head-integrated exhaust manifolds and improved thermal management systems
          reduce warm-up time for the cabin and provide greater passenger comfort in cold
          climates.
p It is also possible that these new technologies may have undesirable adverse effects - hidden costs - associated
  with them, such as noise or vibration. EPA's analysis to identify hidden costs through review of professional auto
  reviews, discussed in Chapter 6.4.1.2, did not find evidence of systematic hidden costs of the new technologies.


                                              4-35

-------
                                                    Baseline and Reference Vehicle Fleets
       •  PEVs which can be remotely activated or programmed to precondition the vehicle in
          a garage when plugged in provide greater passenger comfort and convenience.  In
          cold weather, the vehicle can be pre-warmed and defrosted, and in warm weather the
          vehicle can be pre-cooled.
       •  PEV systems with an electric axle on AWD vehicles, or even each individual wheel
          with electric drive motors, can provide torque vectoring for improved driving
          dynamics as the increased torque on the outside wheel is able to steer the car into the
          corner.
       •  LED headlights enable adaptive automotive headlight systems, in which lighting
          intensity and direction can be automatically controlled to road, ambient lighting, and
          weather conditions.
   Additional discussion of the effects of each technology considered in this Draft TAR is
provided in Chapter 5.

4.1.3.4 Estimating Potential Opportunity Costs and Ancillary Benefits

   As this discussion has shown, the standards could potentially lead to opportunity costs in
terms of reduced power or other adversely affected vehicle attributes. At the same time, the
standards could induce major innovations that may be used in part to mitigate those opportunity
costs, and that may in addition lead to ancillary benefits. Because the standards may contribute
both benefits and  costs to other vehicle attributes, measuring the net effect on consumer impacts
requires estimates of the values of these attributes to consumers.

   The most common sources of estimates of willingness to pay for these attributes are models
developed to understand vehicle purchase decisions. These studies quantitatively estimate the
role of various vehicle characteristics, such as size, power, and fuel economy, in those purchase
decisions. The parameters estimated for these characteristics can usually be used to derive
estimates of the value - the willingness to pay (WTP) — of each attribute to consumers. It is
common in this literature, though, for the researchers themselves not to have done the WTP
calculation. In a 1988 study, Greene and Liu27 reviewed the literature to that time; they found,
"The dispersion of estimated attribute values both within and across models is striking," varying
by factors of 5 to  10 or more;  for performance, they considered the variation "wild. .  . from -$8
to $4,081 per 0.01 cubic inches per pound."  To our knowledge, there has not been a study since
that time that has  done a comprehensive review of consumers' willingness to pay for vehicle
attributes.*2 28

   EPA has commissioned a new review of the literature to understand what is known about
consumer valuation of vehicle characteristics.  This review is looking at the metrics various
studies have considered important for consumer vehicle purchase decisions, and is calculating
the WTP values implied by the estimates in those studies. The goal is to determine whether there
are robust WTP values that could be used for monetizing at least some of the opportunity costs
and ancillary benefits. A draft of that report is expected in summer 2016.
Q Greene (2010) conducted a review of consumers' willingness to pay for one attribute, fuel economy, and found
  wide ranges of values.
                                              4-36

-------
                                                   Baseline and Reference Vehicle Fleets
4.1.4   Incorporation of the California Zero Emissions Vehicle (ZEV) Program into the
EPA Reference Fleet

4.1.4.1 The ZEV Regulation in OMEGA

   In its analysis, EPA has considered sales of electrified vehicles as projected to be needed to
meet State Zero Emission Vehicle (ZEV) requirements.  Because these ZEVs are already
required by separate regulations in California and nine_other states, these vehicles are built into
the EPA reference fleet.  This approach reasonably avoids attributing costs to the federal GHG
program which necessarily occur due to another existing requirement and assures that those costs
are not double counted. (Note that this reflects a change from the 2012 FRM where EPA did not
account for compliance with the ZEV regulations in the reference case fleet for the 2017-2025
standards. However, this was because CARB was simultaneously  substantially revising the ZEV
regulation in early 2012 just prior to the release of the 2012 FRM and EPA had not yet acted
upon California's waiver request for the ZEV program).

   This analysis is meant to be one example representation of how the ZEV program
requirements could be fulfilled; it is in no way meant to reflect the exact way in which any given
manufacturer would actually comply with the ZEV program. Rather, it is meant as an
illustration to reflect the potential number and penetration of ZEVs across the national fleet as
part of the reference case. To accomplish this, the baseline fleet with future sales projections had
to be adjusted to account for the projected ZEV sales.  Those sales adjustments are described in
detail below (see 4.1.4.2). The analysis fleets used in OMEGA and in EPA's benefit cost
analysis are shown in Tables 4.24 through Table 4.28.

   Note that, in Tables 4.24 through Table 4.28, EPA shows "Baseline" EV and PHEV sales and
"Additional ZEV Program" EV and PHEV sales. The "baseline" sales are sales projected in
EPA's MY2014-based baseline fleet. In other words, these vehicles are part of the future fleet
described in Section 4.1.2.1. The "additional ZEV program" sales are EV and PHEV sales above
and beyond those projected in Section 4.1.2.1. The "additional ZEV program" sales were taken
from the ICE-only sales that were projected in Section 4.1.2.1. We have not increased the size of
the fleet,  but have  "converted" some ICE-only vehicles to EVs and PHEVs to meet the projected
sales required by the ZEV program in California and nine other states. We describe the process
of doing this in the text following the tables. Importantly, the costs of "converting" the
"additional ZEV program" sales are  attributable to those programs  adopting the ZEV program
and, therefore, those costs are not considered in the EPA analysis. Similarly, any benefits from
those vehicles are not considered explicitly in the EPA analysis. However, there is an implicit
benefit that is considered. Since the ZEV program vehicles are part of the analysis fleet, they
reduce slightly the compliance burden for any manufacturer required to meet the ZEV program
because of their low tailpipe emissions when averaged with other vehicles in that manufacturer's
fleet. We model the fleet in this way because this is how ZEV program vehicles will be treated in
the National Program.
                                             4-37

-------
                                                           Baseline and Reference Vehicle Fleets
          Table 4.25 OMEGA MY2021 Car Fleet using the AEO 2015 Reference Fuel Price Case
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-
Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
ICE-only Car
Sales
282,880
586,667
904,320
1,174,858
768,430
1,090,833
21,101
243,393
214,942
45,378
742,674
131,755

1,093,150
447,866
38,574
7,786,822
Baseline EV
Sales
3,273
6,909
1,355
600
11
0
0
0
3,944
1,344
8,201
0
86,636
1,418
0
0
113,691
Baseline
PHEV Sales
8,770

7,007
26,201
719
0
0
0
0
0
0
0
0
10,630
0
0
53,327
Additional ZEV
Program EV Sales
2,543
2,429
9,952
9,612
10,093
7,396
1,192
2,191
888
0
5,031
1,224
0
13,091
6,599
794
73,035
Additional ZEV
Program PHEV Sales
1,514
11,660
12,378
564
15,312
11,587
1,868
3,433
6,829
374
11,970
1,918
0
13,797
10,339
1,244
104,787
Total Car
Sales
298,980
607,666
935,011
1,211,835
794,566
1,109,815
24,161
249,017
226,604
47,096
767,876
134,897
86,636
1,132,086
464,804
40,612
8,131,662
Note: The analysis fleet differs from the baseline fleet by removing small volume
Ferrari, McLaren, and Lotus) and by adjusting sales to account for projected ZEV
manufacturers (Aston Martin,
sales.
         Table 4.26  OMEGA MY2021 Truck Fleet using the AEO 2015 Reference Fuel Price Case
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-
Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
ICE-only
Truck Sales
110,369
1,438,814
1,358,371
1,323,614
741,722
157,915
103,489
106,222
159,880
29,109
555,586
462,747
0
1,019,912
303,810
46,418
7,917,977
Baseline EV
Sales
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Baseline
PHEV Sales
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Additional ZEV
Program EV Sales

918
585
884
3,992
582

694
0
0
1,215
4,038
0
3,238
0
0
16,147
Additional ZEV
Program PHEV Sales

4,408
727
52
6,057
912

1,087
0
216
2,890
6,327
0
3,413
0
0
26,088
Total
Truck
Sales
110,369
1,444,140
1,359,683
1,324,550
751,770
159,409
103,489
108,003
159,880
29,325
559,691
473,112
0
1,026,564
303,810
46,418
7,960,213
Note: The analysis fleet differs from the baseline fleet by removing small volume manufacturers (Aston Martin,
Ferrari, McLaren, and Lotus) and by adjusting sales to account for projected ZEV sales.
                                                    4-38

-------
                                                           Baseline and Reference Vehicle Fleets
          Table 4.27 OMEGA MY2025 Car Fleet using the AEO 2015 Reference Fuel Price Case
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-
Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
ICE-only Car
Sales
298,264
587,738
881,873
1,231,982
797,320
1,121,220
20,341
249,487
225,277
56,667
786,957
138,497

1,142,185
481,441
39,666
8,058,914
Baseline EV
Sales
3,859
6,678
1,460
768
11
0
0
0
5,065
1,477
8,523
0
103,502
1,616
0
0
132,959
Baseline
PHEV Sales
10,692
0
6,772
27,823
786
0
0
0
0
0
0
0
0
10,384
0
0
56,458
Additional ZEV
Program EV Sales
7,224
10,454
19,758
19,694
21,696
15,384
2,255
4,593
4,488
360
13,423
2,616
0
27,666
14,138
1,645
165,394
Additional ZEV
Program PHEV Sales
4,184
18,041
19,821
7,463
24,901
18,076
2,649
5,397
10,511
823
19,048
3,074
0
25,579
16,612
1,933
178,112
Total Car
Sales
324,223
622,911
929,684
1,287,730
844,715
1,154,680
25,245
259,477
245,341
59,327
827,952
144,187
103,502
1,207,430
512,191
43,244
8,591,837
Note: The analysis fleet differs from the baseline fleet by removing small volume
Ferrari, McLaren, and Lotus) and by adjusting sales to account for projected ZEV
manufacturers (Aston Martin,
sales.
         Table 4.28 OMEGA MY2025 Truck Fleet using the AEO 2015 Reference Fuel Price Case
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-
Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
ICE-only
Truck Sales
101,636
1,459,761
1,286,443
1,277,635
722,752
154,756
95,454
111,360
151,199
32,515
535,267
480,683
0
984,287
311,139
46,908
7,751,796
Baseline EV
Sales
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Baseline
PHEV Sales
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Additional ZEV
Program EV Sales
0
3,793
1,391
1,837
7,149
1,108

1,452
0
186
2,787
8,522
0
6,930
0
0
35,153
Additional ZEV
Program PHEV Sales
0
6,545
1,396
696
8,205
1,302

1,706
0
425
3,954
10,013
0
6,407
0
0
40,649
Total
Truck
Sales
101,636
1,470,099
1,289,230
1,280,168
738,106
157,166
95,454
114,518
151,199
33,126
542,008
499,218
0
997,624
311,139
46,908
7,827,599
Note: The analysis fleet differs from the baseline fleet by removing small volume manufacturers (Aston Martin,
Ferrari, McLaren, and Lotus) and by adjusting sales to account for projected ZEV sales.
                                                    4-39

-------
                                                   Baseline and Reference Vehicle Fleets
   To generate the fleet inclusive of the ZEV program sales, we began with the fleet shown in
the above tables exclusive of the additional ZEV program sales.  That fleet included some EVs
and PHEVs consistent with the sales in the MY2014 baseline fleet as projected forward to MYs
2021 and 2025. Those sales are shown in the tables Table 4.25 through Table 4.28 above. The
additional ZEV program sales shown above, rather than being EVs and PHEVs, were internal
combustion cars and trucks in the original  fleet. For example, Table 4.28 shows additional ZEV
program truck fleet sales of 35,153 EVs and 40,649 PHEVs.  Those combined 75,802 vehicles
were originally ICE vehicles meaning that the baseline ICE sales were the 7,751,796 shown in
column 2 of the table above plus an additional 75,802, or 7,827,598 in total. To "generate" the
projected additional 75,802 ZEV program  vehicles, each model within a manufacturer's fleet was
mapped into a vehicle typeR matching its characteristics and capability. For this analysis, it was
assumed that only vehicle types classified  as non-towing would be considered for conversion
from an ICE to a ZEV to meet the ZEV program requirements. The eight vehicle types
considered for additional ZEV program sales include all of the passenger car vehicle types
(vehicle types 1 through 6) along with the  two small truck and small CUV/SUV vehicle types
(vehicle types 7 and 13). Table 4.29 lists the 19 possible vehicle types including the towing or
non-towing designation and consideration  as a "ZEV-source platform," Rather than selecting
which individual vehicle models or platforms would be the most likely sources, all ICE vehicles
within those eight vehicle types in a manufacturer's fleet were considered as a source for
additional ZEV program sales. Each manufacturer's additional ZEV program sales were then
created by converting, on a platform-level  sales weighted basis across all eligible vehicle types,
the necessary number of ICE vehicles into the respective EV and PHEV sales.  The tables below
are meant to provide clarity with a simple  example of how this was done.8
R We discuss "vehicle types" in Appendix C.
s The Excel spreadsheets used to generate the ZEV program fleet are in the docket and on our website at
             ;j)^^                      the filenames include the keyword "FleetsABC."
                                             4-40

-------
                                                   Baseline and Reference Vehicle Fleets
                   Table 4.29 Vehicle Types Considered for Conversion to ZEVs
Vehicle - Engine - Valve Type
SubCompactAuto - 14 - DOHC
Auto - 14 - DOHC
Auto - V6 - DOHC
Auto - V6 - SOHC
Auto - V8 - DOHC
Auto - V8 - OHV
MPV-I4-DOHC
MPV-V6-DOHC
MPV-V6-SOHC
MPV-V6-OHV
MPV-V8-DOHC
MPV-V8-OHV
Truck -14 -DOHC
Truck -V6- DOHC
Truck - V6 - OHV
Truck -V8- DOHC
Truck -V8- SOHC
MPV-V8-SOHC
Truck -V8- OHV
Vehicle Type
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Towing(T)/ Non-towing(N)
N
N
N
N
N
N
N
T
T
T
T
T
N
T
T
T
T
T
T
ZEV Platform
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
Y
N
N
N
N
N
N
   First, consider a simple manufacturer fleet consisting of seven vehicle models built on five
platforms which we have mapped into three vehicle types with total fleet sales of 600 vehicles,
see Table 4.30.
             Table 4.30 Example Manufacturer Fleet from which ZEVs are to be Created
Platform index
100
100
101
101
102
103
104
Total
Vehicle index
1
2
3
4
5
6
7

Model
A
B
C
D
E
F
G

Fuel
G
G
G
G
G
G
G

VehType
1
1
2
2
1
2
17

Baseline sales
100
100
75
75
100
50
100
600
   For this manufacturer, we will assume that the needed additional ZEV program sales are 50
EVs and, for simplicity, no PHEVs. As noted above, only vehicle types 1-7 and 13 are
considered to be ZEV-source platforms. Thus, the 50 ZEV program vehicles cannot come from
platform 104 since that is vehicle type 17. We determine the number of EVs to create from each
                                             4-41

-------
                                                        Baseline and Reference Vehicle Fleets
platform according to its sales weighting within ZEV-source platforms.1 This is shown in Table
4.31. We also need to know how many vehicles within each vehicle model to convert to a ZEV
program vehicle. This is shown in Table  4.32.

               Table 4.31 Number of Additional ZEV Program Sales from each Platform
Platform index
100
101
102
103
Total
VehType 1
200

100

300
VehType 2

150

50
200
Total
200
150
100
50
500
%in Platform
40%
30%
20%
10%
100%
# of ZEV program sales
20
15
10
5
50
            Table 4.32 Percentage of Additional ZEV Program Sales from Each Vehicle Model
Platform index
100
101
102
103
Model A
50%



Model B
50%



Model C

50%


Model D

50%


Model E


100%

Model F



100%
Total
100%
100%
100%
100%
   With the details shown in Table 4.31 and Table 4.32, we can then convert ICE vehicles into
ZEV program vehicles as shown in Table 4.33.

            Table 4.33 Example Manufacturer's OMEGA Fleet including ZEV Program Sales
Platform
index
100
100
101
101
102
103
104
100
101
102
103
Total sales G
Total sales E
Total sales
Vehicle index
1
2
3
4
5
6
7
8
9
10
11



Model
A
B
C
D
E
F
G
ZEV
ZEV
ZEV
ZEV



Fuel
G
G
G
G
G
G
G
E
E
E
E



VehType
1
1
2
2
1
2
17
1
1
2
2



Baseline Sales
100
100
75
75
100
50
100
0
0
0
0
600
0
600
OMEGA fleet
with ZEV
program sales
90
90
68
68
90
45
100
20
15
10
5
550
50
600
T The ZEV-source platforms are those platforms "mapped" into the 8 "ZEV platform" vehicle types presented in
  Table 4.29. The point of Table 4.29 is to make clear that we are creating ZEV program vehicles in only those
  types of vehicles that we believe to make the most sense. Those types of vehicles being passenger cars and the
  smallest sport and cross-over utility vehicles that have 4-cylinder engines and therefore are not "towing" vehicles.
  The ZEV program vehicles are created only from within those vehicle types and, therefore, the creation of ZEV
  program vehicles is done using sales-weighting within those vehicle types rather than within all vehicles.
                                                  4-42

-------
                                                    Baseline and Reference Vehicle Fleets
   As noted above, we then created each manufacturer's ZEV program fleet by converting, on a
platform-level sales weighted basis, the necessary number of ICE vehicles into the respective EV
and PHEV sales.  Staff considered an alternate approach to look instead at which specific
platforms, or even vehicle models, were the best candidates for conversion to EV/PHEV.
However, that approach was rejected because a problem with that is, by what measure does one
determine the best candidates for conversion?  The smallest cars? The lightest cars?  Those that
already have an EV or PHEV version?  We were concerned that any attempt at determining the
"best" candidates for conversion might be seen as "cherry picking" in order to provide a certain
result.  Some might see us as choosing  all of the smallest vehicles thereby leaving all of the
larger, perhaps dirtier vehicles as ICE vehicles needing costly improvements to comply with the
future standards. Others might see us as choosing all of the largest vehicles thereby leaving all
of the smaller, perhaps cleaner vehicles as ICE vehicles needing less costly improvements to
comply with future standards. Further, there is no clear trend as to which vehicles or platforms
manufacturers are currently using for EV or PHEV platforms. Current and publicly announced
near term models span platforms from subcompact cars to large cars, large SUVs to minivans,
and use of shared or dedicated platforms. Our final decision was to choose equally (by sales
weighting) from each ZEV source platform such that there would be no net impact on the sales
weighted footprint of remaining ICE vehicles needing technology to comply.

4.1.4.2 The ZEV Program Requirements

   The preceding discussion describes how we determined which vehicles would be converted
from ICE technology to EV/PHEV.  Here we discuss how many vehicles to actually convert or,
in other words, what the  additional ZEV program sales are projected to be.

4.1.4.2.1       Overview

   California requires the largest vehicle manufacturers to manufacture ZEV credit producing
vehicles to comply with the increasing number of ZEV credits required through 2025.u The
ZEV credits can be generated by producing battery electric vehicles, fuel cell electric vehicles,
and plug-in hybrid vehicles.  In addition to the requirements applying in California (CA), several
other statesv have used section  177 (SI 77) of the federal Clean Air Act to adopt the California
ZEV requirements (referred to as SI77 ZEV States).  These states, when combined with CA,
account for nearly 30 percent of all new light-duty vehicles sold in the United States.

   Under the ZEV regulation, manufacturers are required  to generate ZEV credits to fulfill an
annual obligation based on their cumulative vehicle sales  as summarized in Table 4.34.
Requirements  are satisfied by producing vehicles that generate credit which,  for MY2018 and
beyond, means a combination of plug-in hybrid electric vehicles (PHEV), battery electric
vehicles (BEV),  and fuel cell electric vehicles (FCEV). Each PHEV, BEV, and FCEV earns
between 0.4 and 4 credits per vehicle depending on its electric range over a test cycle. For
example, a PHEV with a 10 mile electric range earns 0.4 credits and a BEV or FCEV with  a 350
mile test range earns 4.0 credits.
u Title 13, California Code of Regulations, section 1962.2 "Zero-Emission Vehicle Standards for 2018 and
  Subsequent Model Year Passenger Cars, Light-Duty Trucks, and Medium-Duty Vehicles."
v Section 177 ZEV states: Connecticut, Maine, Maryland, Massachusetts, New York, New Jersey, Oregon, Rhode
  Island, and Vermont.
                                              4-43

-------
                                                  Baseline and Reference Vehicle Fleets
   To incorporate the ZEVs into the OMEGA fleet, the ZEV regulation credit requirements were
converted to a vehicle sales requirement as follows:

   1) Determine how many total ZEV credits each manufacturer will need in CA and the SI77
ZEV states for the two years being modeled in OMEGA (MY2021 and MY2025).

   2) Develop a nominal BEV electric range (described in Table 4.33) and a nominal PHEV set
of electric range characteristics (described in Table 4.34) that are projected to be representative
of BEV and PHEV capability in the MY2021-2025 time frame. The range and characteristics
are then used to determine how many ZEV credits each vehicle will generate.

   3) Calculate the incremental ZEV credits needed beyond those generated by any ZEVs
already included in the OMEGA reference fleet projections and expected to be sold in CA and
the SI77 ZEV states.

   4) Determine how many incremental BEVs and PHEVs each manufacturer will need to sell to
satisfy their ZEV credit  obligations for MY2021 and MY2025.
4.1.4.2.2
ZEV Credit Requirement
   Each manufacturer's ZEV credit obligation is calculated by multiplying its projected total
light duty vehicle sales in CA and SI77 ZEV states by the ZEV credit percentage required (see
Table 4.34 below).  The total projected CA and SI77 ZEV states sales volume for each
manufacturer was calculated by multiplying the manufacturer-specific reference fleet national
sales volumes in OMEGA by the current (MY 2014) CA and SI77 ZEV states sales volume
ratio.  For example, if manufacturer "A" is projected to sell 250,000 vehicles nationally in MY
2021, and it's MY 2014 CA and SI77 ZEV state sales are 40 percent of its national sales, its
projected MY 2021 CA and SI77 ZEV state sales would be 100,000 (250,000*40%).  Although
the regulation has flexibilities in the technologies a manufacturer may use to generate credits,
there is a cap on the portion of the credits that can be satisfied with PHEVs as identified in Table
4.34.  For example, if manufacturer "A" sells 100,000 vehicles in CA and the S177 ZEV states in
2021, it is required to generate 12,000 ZEV credits  (100,000*12%) in 2021 and, of those 12,000
ZEV credits, only 4,000 (100,000*4%) can come from PHEVs.  For the purpose of this analysis,
manufacturers are projected to comply with the ZEV requirements by maximizing their ZEV
credits earned using PHEVs and using BEVs to generate the remaining credits.
                        Table 4.34 ZEV Regulation Credit Requirements


Total ZEV
Credit
Required
Max.
Credits
From
PHEVs
ZEV Credit Requirements
2018
4.50%
2.50%
2019
7.00%
3.00%
2020
9.50%
3.50%
2021
12.00%
4.00%
2022
14.50%
4.50%
2023
17.00%
5.00%
2024
19.50%
5.50%
2025
22.00%
6.00%
                                            4-44

-------
                                                    Baseline and Reference Vehicle Fleets
4.1.4.2.3     Projected Representative ofPHEVandBEV Characteristics for MY2021-2025

   The first step to calculate the number of ZEVs needed in the projected fleet to meet the
manufacturer's credit obligation is to determine the type of vehicles that will be used to comply
with the regulation. The primary  characteristic for determining ZEV credits per vehicle is the
urban dynamometer driving schedule (UDDS) test cycle range for BEVs and the UDDS test
cycle "equivalent all electric range" for PHEVs.w  Given that these would be future vehicles for
which actual specifications are not yet known, assumptions were made regarding what future
range(s) might be in the MY 2021 and MY 2025 timeframe. Further simplifications of such
projections were also necessary to fit within the existing model framework of OMEGA including
baseline vehicles and technology  packages. These simplifications include the use of a single
nominal BEV range and a single nominal PHEV range for all manufacturers and all vehicle
classes with characteristics projected to be representative of BEVs and PHEVs in the MY2021 to
2025 timeframe.  Given these  constraints, this projection reflects a scenario for minimum
compliance with the ZEV regulation using a representative nominal BEV and PHEV but not a
'likely' scenario that might reflect a wide variety of different ranges of PHEV and BEV offerings
across manufacturers, vehicle  classes, and model years or the inclusion of FCEVs that have
already begun to enter the market.

   To develop the nominal BEV and PHEV electric range, staff first looked at the relative impact
of battery pack costs for a variety of battery costs (dollars per kilowatt-hour (kWh)). For this
simplified analysis, vehicle energy consumption was assumed to be constant for all vehicle
types; therefore all-electric vehicle range and  battery pack size increase proportionally. The
relative costs to achieve longer range were then compared to the number of ZEV credits earned
for the increased range. The qualitative results are shown in Figure 4.4.  As the figure shows,
building individual BEVs with a longer range directionally results in a lower cost per ZEV credit
earned (i.e., satisfying the ZEV credit obligation with fewer long range BEVs is directionally
more cost-effective than using a larger volume of shorter range BEVs).  And, as Figure 4.4
illustrates, the relative impact  is even larger at lower battery costs. Accordingly, the nominal
BEV and PHEV packages targeted longer range variants of both types of ZEVs rather than
multiple variants of shorter and longer range vehicles.  Note that the range of battery costs used
in the figure (from  $150/kWh  to $300/kWh in the  2021-2025 time frame) is consistent with the
projections of the EPA battery costing analysis for PHEVs and BEVs as shown in Tables 5-84
through 5-88. The reasonableness of EPA's projected costs used in both the 2012 FRM and this
Draft TAR is supported elsewhere, particularly in  Section 5.2.4.4.9 where we evaluate the 2012
FRM battery cost projections,  and in Section 5.3.4.3.7.6 where we discuss Draft TAR battery
cost projections.
w As defined in "California Exhaust Emission Standards and Test Procedures for 2018 and Subsequent Model Zero-
  Emission Vehicles and Hybrid Electric Vehicles, in the Passenger Car, Light-Duty Truck and Medium-Duty
  Vehicle Classes," adopted March 22, 2012, last amended May 30, 2014.
                                             4-45

-------
                                                   Baseline and Reference Vehicle Fleets
         •o
         01
         jj
         M
         "5
         •o

         "6
         ts
         o
         01
         V
         £
         01
                                         Battery Cost:
                                         $300/kWh
                                      A  Batte ry Cost:
                                         $250/kWh
                                     —•— Batte ry Cost:
                                         $200/kWh
                                      >  Batte ry Cost:
                                         $150/kWh
                           100
    200
Range (miles)
300
400
            Figure 4.4 Relative Cost of ZEV Credits for Different Ranges and Battery Costs

   The projected range for the nominal BEV and PHEV in the MY2021 to 2025 timeframe was
developed assuming a constant improvement from the current sales-weighted average range.
The MY2014 BEV sales-weighted label range is -156 miles, as shown in Table 4.33 below;  for
MY2014 PHEVs, the sales-weighted label electric range is -26 miles as shown in Table 4.36.
                                             4-46

-------
                                                   Baseline and Reference Vehicle Fleets
                      Table 4.35 Range Characteristics of BEVs for MY2014
Brand
BMW
Chevrolet
Fiat
Ford
Honda
Mercedes-Benz
Mercedes-Benz
Mitsubishi
Nissan
Tesla
Tesla
Tesla
Toyota
Model
i3
Spark EV
500e
Focus Electric
FitEV
Smart fortwo Convertible
Smart fortwo Coupe
i-MiEV
Leaf
Model S 60
Model S 85
Model SAWD(P85D)
RAV4 EV
Sales-Weighted Average Range (Label Miles)
EPA Label All-electric
Range (miles)
81
82
87
76
70
68
68
50
84
200
270
270
80
155.5
                     Table 4.36 Range Characteristics of PHEVs for MY2014
Brand
Ford
Ford
Cadillac
Chevrolet
Honda
Toyota
Model
C-Max Energi
Fusion Energi
ELR
Volt
Accord Plug-In
Prius Plug-In
Sales-Weighted Average Range (Label Miles)
EPA Label All-electric
Range (miles)
21
21
37
38
13
11
26.2
   For this analysis, the range for future vehicles was estimated to increase at a rate of 5 percent
per year until the sales-weighted range reaches 245 miles which correlates to the maximum
number of ZEV credits earned by any one vehicle. While manufacturers are not expected to
actually redesign vehicles to increase the range every year or to cap the range when they reach
the 245 mile range, this rate of annual improvement is consistent with the improvements
manufacturers have been making over more discrete intervals such as redesigns,  refreshes, or
other updates. For example, new or updated model introductions and announcements for the
Ford Focus EV, VW e-Golf, Nissan Leaf, Tesla Model S, Tesla Model 3, GM Bolt EV, GM
Volt, and BMW i3 have all included increased range compared to their predecessors. The 5
percent rate of growth is an estimated average of both longer and shorter range vehicles. It is not
expected that BEVs with 200+ miles of range, such as some Teslas, will increase their range as
                                             4-47

-------
                                                   Baseline and Reference Vehicle Fleets
quickly as shorter range vehicles such as the BMW i3.  This is supported by the 2.5 percent per
year increase observed in the Model S (85 to 90 kW-h) compared to the 9 percent per year
increase seen by the GM Volt and the BMW i3. Additionally, while some OEMs may continue
offering BEVs with lower ranges, these may be offset by longer range offerings such as
hydrogen fuel cell electric vehicles (FCEV) like those announced by Toyota and Honda with
ranges that well exceed 200 miles.

   Given that the time period of interest for the midterm evaluation is MY2021-2025 and that the
ZEV requirements increase annually, a nominal range for the  single BEV variant to be used for
all model years was determined by calculating the sales-weighted average for the years being
evaluated. Table 4.37 combines the results from Table 4.33 for average electric range with the
projected BEV sales for MY 2021-2025 to calculate a sales-weighted average BEV for MYs
2021-2025.  The sales-weighted average was calculated as 237 miles. Although this projection
results in an estimated 237 mile range, a final range of 200 miles was chosen to account for a
potential slower-than-historical increase in range and to be consistent with an existing
technology package in OMEGA. A 200 mile label range is reasonable given recent
announcements in this magnitude for the Tesla Model 3, GM Bolt EV, and  an announced future
Ford BEV which will all be available prior to MY2021.  ZEV credits are generated based on
UDDS range, not label range, and a review of current certified BEVs indicates a UDDS range to
label range correction factor of between 0.65 and 0.76.  For this analysis, a  value of 0.7 was used
for the nominal BEV. As a result, for the model years being evaluated, all BEV200s are
assumed to have a label range of 200 miles and a UDDS range of 286 miles which generates
3.36 ZEV credits per vehicle.
                 Table 4.37 Projected Sales Weighted BEV Range for MY2021-2025
Model year
2021
2022
2023
2024
2025

EV real-world range
218.1
229.0
240.5
245.0
245.0

BEV sales
(% of whole fleet)
2%
3%
3%
4%
4%
Range Based on Sales
Weighting MY2021-2025
BEV sales
(% of 2021-2025
cumulative sales)
14%
17%
20%
23%
26%
237.5
   The projected ranges for PFLEVs in the MY2021-2025 time frame were calculated in a similar
manner to the BEV ranges with one minor difference. PFLEVs generate credits based not only on
electric range on the UDDS cycle but also on the ability to drive all electrically for at least 10
miles of the US06 supplemental FTP test cycle.  PFLEVs that can meet this US06 criterion earn
an additional 0.2 credits per vehicle. While the reality is that motor, inverter, and battery pack
sizing along with the powertrain architecture all play a role in determining whether a PHEV can
meet this criterion, for this analysis, the ability to meet it was assumed to increase linearly for
vehicles with electric range from 20 to 40 miles (i.e., 0 percent of PFLEVs with 20 mile range, 50
percent of PFLEVs with a 30 mile range, and 100 percent of PFLEVs with 40 mile range can meet
                                             4-48

-------
                                                   Baseline and Reference Vehicle Fleets
the US06 criterion).  The analysis summarized in Table 4.38 shows that, for MY2021-2025, the
sales-weighted average PHEV is projected to have a range of about 41 miles which was rounded
down to a final range of 40 miles to be consistent with an existing PHEV40 technology package
in OMEGA. A PHEV40 is assumed to be 100 percent US06 capable so it generates 1.07 credits
per vehicle after adjusting from a 40 mile label range to an equivalent HDDS range and
including the additional credits for US06 capability. For perspective, the newly revised MY2016
GM Volt already exceeds this capability and other manufacturers are expected to further increase
their range and capability over the next 5-9 years.
        Table 4.38 Projected Sales Weighted PHEV Range and US06 Capability for MY2021-2025
Model year
2021
2022
2023
2024
2025
EV real-world
range
36.8
38.6
40.6
42.6
44.7

PHEV sales
(% of whole fleet)
4%
4%
5%
5%
5%
Range Based on Sales
Weighting MY2021-
2025
PHEV sales
(% of 2021-2025
cumulative sales)
17%
19%
20%
21%
23%
40.9
4.1.4.2.4
Calculation of Incremental ZEVs Needed for ZEV Program Compliance
   Next, the number of ZEV credits that would be generated from vehicles already included in
the projected reference fleet was subtracted from the total credit obligation. Given the projected
reference fleet only included national sales numbers for ZEVs, those numbers were first scaled to
CA and SI77 ZEV state sales using the current (MY2014) manufacturer-specific percentage of
national ZEV sales in CA and the SI77 ZEV states. For this analysis, all manufacturers are
projected to generate ZEV credits using the same nominal sales-weighted BEV and PHEV all
electric ranges and each manufacturer is projected to fulfill their credit requirements without
exercising any of the various additional flexibilities included in the ZEV regulation.  These
earned credits were then subtracted from each manufacturer's  credit obligation to calculate  the
remaining incremental credits needed. For example, if a manufacturer's ZEV credit obligation
for MY2021 is 12,000 credits and the original baseline projected 1000 BEV sales in CA and the
S177 ZEV states, its incremental obligation is 8,640 ZEV credits (12,000 credits -1000
vehicles*3.36 credits/vehicle).

   Finally, the incremental credits needed were translated to the number of additional PHEV and
BEV sales for each manufacturer. For this analysis, it was assumed that each manufacturer
would satisfy the maximum amount of ZEV credits allowed with PHEVs and the remaining
portion with BEVs so both the ZEVs in the original reference  fleet and those incrementally
added take this PHEV limitation into account. No ZEV credit trading and banking was included
in this analysis; each manufacturer was assumed to meet its  ZEV obligation in MY2021 and
MY2025 with vehicles produced for those model years.  An example analysis can be found in
Table 4.39 and Table 4.40.  For the projected sales volumes used in this  draft TAR, the overall
                                             4-49

-------
                                                   Baseline and Reference Vehicle Fleets
effect of the ZEV regulation is an addition of approximately 220,000 and 420,000 ZEVs in the
reference fleet for model years 2021 and 2025, respectively. This increases the percent of ZEVs
in the OMEGA reference fleet from 1.0 percent of national sales to 1.7 percent in MY2021, and
1.2 percent to 3.0 percent in MY2025.
                                             4-50

-------
                                                   Baseline and Reference Vehicle Fleets
Table 4.39 Incremental PHEV40s and BEV200s needed in MY2021

Row Labels
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-
Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Grand Total

Sum of
Annual
Sales -
Cycle 1
409349
2051806
2294695
2536385
1546336
1269224
127650
357020
386483
76422
1327567
608009
86636
2158650
768613
87030
16091875
Reference Fleet Characteristics
National
BEV200
sales
3273
6909
1355
600
11
0
0
0
3944
1344
8201
0
86636
1418
0
0

National
PHEV40
sales
8898
0
7239
26470
744
0
0
0
0
0
0
0
0
10898
0
0

%CA+S177
BEVand
PHEV40
66%
100%
67%
62%
96%
81%
0%
0%
88%
35%
40%
0%
56%
97%
98%
0%

% Total
Sales in
CA+S177
48.0%
21.0%
20.9%
18.0%
38.3%
26.4%
39.2%
33.9%
47.3%
24.1%
30.0%
36.3%
56.8%
34.4%
36.0%
38.3%

CA+S177
BEV200
sales
2144
6909
901
372
10
0
0
0
3471
471
3239
0
48083
1368
0
0

CA+S177
PHEV40
sales
5828
0
4814
16411
714
0
0
0
0
0
0
0
0
10516
0
0

Total ZEVs Needed
TOTAL
BEV200
Sales
Needed
CA+S177
4687
10256
11438
10868
14095
7978
1192
2885
4359
439
9485
5263
1172
17697
6599
794

TOTAL
PHEV sales
CA+S177
7342
16068
17919
17027
22083
12499
1868
4520
6829
688
14860
8245
1836
27726
10339
1244

Incremental ZEVs Needed
Incremental
BEV
2543
3347
10537
10496
14085
7978
1192
2885
888
0
6246
5263
0
16329
6599
794

Incremental
PHEV40s
needed
1514
16068
13105
616
21369
12499
1868
4520
6829
590
14860
8245
0
17210
10339
1244

                            4-51

-------
                                                   Baseline and Reference Vehicle Fleets
Table 4.40 Incremental PHEV40s and BEV200s needed in MY2025

Row Labels
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-
Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Grand Total

Sum of
Annual
Sales -
Cycle 1
425859
2093010
2218913
2567898
1582821
1311846
120699
373995
396540
92453
1369960
643404
103502
2205054
823330
90151
16419435
Reference Fleet Characteristics
National
BEV200
sales
3859
6678
1460
768
11
0
0
0
5065
1477
8523
0
103502
1616
0
0

National
PHEV40
sales
11104
0
7180
28546
834
0
0
0
0
0
0
0
0
10878
0
0

%CA+S177
BEVand
PHEV40
66%
100%
67%
62%
96%
81%
0%
0%
88%
35%
40%
0%
56%
97%
98%
0%

% Total
Sales in
CA+S177
48.0%
21.0%
20.9%
18.0%
38.3%
26.4%
39.2%
33.9%
47.3%
24.1%
30.0%
36.3%
56.8%
34.4%
36.0%
38.3%

CA+S177
BEV200
sales
2527
6678
971
476
11
0
0
0
4457
517
3367
0
57444
1559
0
0

CA+S177
PHEV40
sales
7273
0
4775
17698
800
0
0
0
0
0
0
0
0
10497
0
0

Total ZEVs Needed
TOTAL
BEV200
Sales
Needed
CA+S177
9751
20925
22120
22007
28856
16492
2255
6045
8945
1063
19576
11138
2800
36156
14138
1645

TOTAL
PHEV sales
CA+S177
11458
24586
25991
25858
33906
19378
2649
7103
10511
1249
23002
13087
3291
42483
16612
1933

Incremental ZEVs Needed
Incremental
BEV
7224
14247
21149
21530
28845
16492
2255
6045
4488
546
16210
11138
0
34596
14138
1645

Incremental
PHEV40s
needed
4184
24586
21217
8159
33106
19378
2649
7103
10511
1249
23002
13087
0
31986
16612
1933

                            4-52

-------
                                                   Baseline and Reference Vehicle Fleets
4.2    Development of the CAFE Light Duty Analysis Fleet

4.2.1   Why did NHTSA Develop the Analysis Fleet?

   In considering potential new CAFE standards, NHTSA considers manufacturers' potential
responses to those standards.  To do so, NHTSA uses a modeling system—often referred to as
"the CAFE model" or "the Volpe model"—developed by DOT's Volpe National Transportation
Systems Center (Volpe Center).  NHTSA's CAFE model relies on many inputs, including an
analysis fleet. The analysis fleet is a forecast of the future vehicle market—defined in terms of
specific manufacturers, vehicle models, and vehicle model configurations—during the model
years to be covered in the analysis. As such, the analysis fleet provides a starting point for
NHTSA's analysis.

   The fleet used for today's analysis is the set of vehicles offered for sale in 2015MY, with
individual vehicle models described by attributes like vehicle specifications, technology features,
and sales volumes. The analysis fleet also covers fleet mix and fuel consumption.  Once the
analysis fleet is  defined, NHTSA estimates how each manufacturer could potentially deploy (not
"should," "must," or "will" deploy) additional fuel-saving technology in response to a given
series of attribute-based standards.  With a representative analysis fleet, NHTSA tracks the
application of technology that may benefit fuel economy and CCh emissions in the current fleet.
When NHTSA accounts for how manufacturers may improve fleet fuel economy with additional
technology, a representative analysis fleet prevents the CAFE model from "double counting" the
benefits of a technology. The model does not allow technology to be added to a vehicle already
equipped with that technology.  Beyond the current fleet, the model also uses projections of
future sales from MYs 2016-2030.  Details appear in the input file.  The analysis fleet grounds
assumptions about vehicle sales and technology proliferation and helps NHTSA understand
potential pathways to compliance for attribute-based standards.

   The structure of the analysis fleet file includes vehicle models sold that year, listed by row.
For each vehicle row, the columns list observable and assignable attributes, including technology
used, sales volumes, vehicle platform,  and other inputs for the CAFE model. As discussed
below, the basic data for vehicle configurations are provided by each manufacturer. In many
cases, manufacturers provided details about technologies, platforms, engines, transmissions, and
other vehicle information. In some cases, the model required information that was not
volunteered by manufacturers. In these instances, NHTSA/Volpe supplemented the analysis
fleet file with information available from commercial and public sources.

4.2.2   How the MY2015 Analysis Fleet Was Developed

4.2.2.1 Background

   In CAFE rulemakings since 2001, NHTSA has used either confidential, forward-estimating
product plans from manufacturers or publicly available data on vehicles already sold. These two
sources present  a tradeoff: confidential product plans provide a comprehensive representation of
what vehicles a  manufacturer expects to produce in coming years, accounting for plans to
introduce new vehicles and fuel-saving technologies and, for example, plans to discontinue other
vehicles and even brands. However, for competitive reasons, most of this information is
provided on a confidential basis and must be redacted prior to publication with rulemaking
documentation.  Since 2010, NHTSA has based its analysis fleets almost exclusively on
                                             4-53

-------
                                                   Baseline and Reference Vehicle Fleets
information from commercial and public sources.  Therefore, unlike an analysis fleet based
primarily on confidential business information (CBI), an analysis fleet based primarily on public
sources can be released to the public, allowing any interested parties to reproduce NHTSA's
analysis.  However, being "anchored" in an earlier model year, such an analysis fleet holds
vehicle characteristics unchanged over time and may not reflect manufacturers' actual plans to
apply fuel-saving technologies (e.g., a manufacturer may apply turbocharging to improve not just
fuel economy, but also to improve vehicle performance), or manufacturers' plans to change
product offerings by introducing some vehicles and brands and discontinuing other vehicles and
brands. For example, in the 2012-2016 Final Rule the 2008 Model Year fleet was used, while
for the 2017-2025 Final Rule both the 2008 and 2010 Model Year fleets were used.  In addition
to reflecting the near dissolution of Chrysler due to market turmoil in that year, the 2008-based
fleet included a significant proportion of models and brands discontinued between 2008 and
2010.

4.2.3  NHTSA Decision to use 2015 Foundation for Analysis Fleet

   NHTSA chose to use the 2015 model year as the foundation for today's analysis fleet because
the data include the most recent possible mix of commercially available technologies and vehicle
configurations, and the data may be made available to the public.  If NHTSA began with
information from an earlier model year, the information could be disclosed, but the analysis fleet
would neither include new vehicles recently introduced (e.g., the Ford F-150 that was redesigned
for 2015), nor would the data include the most recent estimated sales mix. If NHTSA used 2016
model year data, the agency would have needed to use product planning information that could
not be made available to the public.

   Although model year 2015 vehicles were still in production when DOT staff compiled
available information regarding the 2015 fleet, such that final production and fuel economy
values may be slightly different for specific model year 2015 vehicle models and configurations
than are indicated in today's analysis, other vehicle characteristics (e.g., footprint, curb weight,
technology content) important to DOT's analysis should ultimately be the same or virtually the
same as indicated here.  Although final CAFE compliance data is available for earlier model
years, even that data can be  subject to later revision (e.g., if errors in fuel economy tests are
discovered). In any event, considering also the range of important changes in model year 2015
(discussed below) to product offerings, DOT's judgment is that using available data regarding the
2015 model year provides the most realistic characterization of the 2015 market.  Insofar as
future product offerings are  likely to be more similar to vehicles produced in 2015 than to
vehicles produced in earlier model years, DOT's judgment is further that using available data
regarding the 2015 model year provides the most realistic publicly releasable foundation for
constructing a forecast of the future vehicle market.

   NHTSA will consider options regarding the set of vehicles upon which to base development
of the analysis fleet to be used for subsequent modeling to evaluate potential new CAFE
standards. For example, one option will be to rely primarily on  model year 2015 data, making
updates to reflect final production volumes giving  the actual sales of each model and any other
new information about characteristics of specific vehicles. Another option will be to develop an
updated analysis fleet based on any information that can be obtained regarding, for example,
vehicles produced in the 2016 model year. NHTSA seeks comment on these and any other
                                             4-54

-------
                                                    Baseline and Reference Vehicle Fleets
options, and on the tradeoffs between, on one hand, fidelity with manufacturers' actual plans
and, on the other, the ability to make detailed analysis inputs and outputs publicly available.

4.2.4   Developments in 2015

   Many new, technologically advanced models were introduced in 2015 Model Year. For
instance, Ford released an aluminum-bodied F150. Acura, BMW, Hyundai, Kia, Lexus, Porsche,
and Volkswagen released new hybrid, plug-in hybrid, and alternative fuel vehicles.
Additionally, manufacturers redesigned many high-volume vehicles for the 2015 model year.

   The following list includes new vehicles, significantly refreshed vehicles, and discontinued
vehicles for 2015:
                   Table 4.41 Summary of Portfolio Revisions by Manufacturer.
Manufacturer


BMW


Daimler
FCA

Ford
General Motors




Honda


New Model Entrants (2015)
2-Series
235i
4-Series, M4
i3
i8
X4
Mercedes GLA
Alfa Romeo 4C
Ram Promaster
Jeep Renegade
Lincoln MKC
Ford Transit Wagon
Cadillac ATS, coupe
Chevrolet City Express
Chevrolet Colorado
Chevrolet Impala, CNG
Chevrolet Trax
GMC Canyon
Acura RLX, hybrid
Acura TLX
Honda Fit



Significant Redesigns (2015)
3-Series
X3
X6


Mercedes C-Class
Dodge Charger
Dodge Challenger
Ford F-150
Ford Expedition
Ford Mustang, 2.3L
Lincoln Navigator
Cadillac Escalade
Chevrolet Tahoe
Chevrolet Suburban
GMC Yukon



Honda CRV


Retired Models






Chrysler 200
Dodge Avenger
Ford E-150
Ford E-250
Ford E-350



Acura ILX, hybrid
Acura TL
Acura TSX
Honda Insight
Honda Accord, PHEV
Honda Fit, EV
Honda FCX Clarity
                                              4-55

-------
                                                   Baseline and Reference Vehicle Fleets

Hyundai Kia
JLR
Mazda
Mitsubishi
Nissan
SUBARU
Tesla
TOYOTA
Volvo
VWA

Hyundai Tucson, Fuel Cell
Kia K900
Kia Soul, EV



Infiniti Q40
Subaru WRX

Lexus NX
Lexus RC

Audi A3, Diesel
Volkswagen e-Golf
Porsche 918 Spyder
Porsche Cayenne, HEV


Land Rover LR2


Nissan Murano
Infiniti QX70
Subaru Legacy
Subaru Outback
Tesla Model S, AWD


Volkswagen Golf
Honda Ridgeline



Mitsubishi i-MiEV
Nissan Cube
Nissan Maxima
Subaru Tribeca

Scion xD
Toyota FJ
Toyota Rav4, EV

Volkswagen Routan
4.2.5   Manufacturer-Provided Information for 2015

   In 2015, NHTSA/Volpe Center staff worked with the Alliance of Automobile Manufacturers
and the Association of Global Automakers to invite individual manufacturers to provide
information on the 2015 model year fleet, including a range of vehicle characteristics, as well as
mid-model year estimates of 2015 production volumes.  In April 2015, NHTSA/Volpe Center
staff provided a template of the input file for the CAFE model, indicating relevant characteristics
of vehicles, engines and transmissions. By fall 2015, virtually all manufacturers provided
extensive included fuel type, combined fuel economy, regulatory class, body style, footprint,
curb weight, powertrain specifications and features, and sales volumes. Many manufacturers
provided substantially more information about their vehicles, including drag coefficient, peak
power and torque, and other specific technologies applied.  NHTSA/Volpe Center staff contacted
manufacturers to clarify and correct some information, and integrated the information into a
single input file for use in the CAFE model.

   NHTSA seeks information that could be used to refine its representation of the 2015 fleet, or
to develop a similarly-detailed representation of a more recent fleet.
                                             4-56

-------
                                                   Baseline and Reference Vehicle Fleets
4.2.6   Other Data

4.2.6.1 Redesign/Refresh Schedules

   Redesign schedules play an important role in the application of new technologies. Many
technologies that may improve fuel economy or reduce CCh emissions may be difficult to
include without a major product redesign.  Therefore, the CAFE model includes redesign
schedules as an input, and the model limits the introduction of most technologies on a vehicle to
major redesign years or refresh years.  In addition to nameplate refresh and redesign schedules,
the CAFE model also accounts for platform refresh and redesign schedules.

   NHTSA did not request future product plans from manufacturers.  NHTSA used information
from Ward's Automotive and other sources to project redesign cycles through 2022.  For years
2023-2030, NHTSA extended redesign schedules based on Ward's projections, segment, and
platform history, and anticipated competitive pressures. For some products with a history of
extended production runs, NHTSA/Volpe Center staff estimated that the duration between future
major redesigns could be shortened by a year or two.

   In some  cases, NHTSA judged the  Ward's data to be incomplete, or misleading. For instance,
Ward's identified some newly imported vehicles as new platforms, but the international platform
was midway through the product lifecycle.  While new to the U.S. market, treating these vehicles
as new entrants would have resulted in artificially short redesign cycles if carried forward, in
some cases. Similarly, Ward's labeled some product refreshes as redesigns, and vice versa.  In
these limited cases, NHTSA revised the Ward's forecast to reflect more realistic redesign and
refresh schedules, for the purpose of the CAFE model.
   Table 4.42 Estimated Average Production Life For Freshly Redesigned Vehicle, By Manufacturer, By
                                        Segment.

BMW
Daimler
FCA
Ford
General Motors
Honda
Hyundai Kia
JLR
Mazda
Mitsubishi
Nissan
SUBARU
Tesla
TOYOTA
Volvo
VWA
Small Car
5.8
7.1
5.5
7.9
5.5
4.9
5.0
7.3
6.5
5.7
6.0
5.0

5.6

7.8
Medium
Car
6.5
6.2
6.7
6.5
6.1
4.7
4.9
7.6
4.2

7.1
5.3

6.4
8.3
7.0
Small
SUV
6.0
5.6
7.0
8.6
5.1
4.5
5.3
6.6
5.0
9.6
7.7
5.1

5.8
8.3
6.7
Medium
SUV
5.6
5.4
6.8
7.5
7.2
5.9
6.3
6.3
6.3

6.1


6.3
8.3
6.9
Pickup


8.1
5.9
4.4


6.3


9.7


9.5


                                             4-57

-------
                                                   Baseline and Reference Vehicle Fleets
   NHTSA Seeks Information that could be used to refine its Representation of the Future
Schedules for Freshening and Redesigning Specific Vehicles.

4.2.6.2 Technologies

   Manufacturers can add technology to a vehicle to improve fuel economy. Each technology
may be more or less effective in reducing fuel consumption, depending on complementary
equipment and vehicle attributes.  As discussed below, Argonne National Laboratory supported
NHTSA's analysis by using Autonomie—Argonne's full vehicle simulation tool—to estimate
the impact of a wide range of potential combinations of different technology, producing a
database of results informing inputs to the CAFE model.  The CAFE model uses these inputs to
estimate the potential benefits  of applying specific combinations  of technologies to specific
vehicles in the analysis fleet.

   The analysis fleet includes many technologies, including vehicle technologies, engine
technologies, and transmission types. For instance, vehicle technologies include mass reduction,
aerodynamic drag reduction, low rolling resistance tires, and others. Engine technologies cover
core powertrain technologies.  Internal combustion engines have  attributes for fuel type, engine
aspiration, valvetrain configuration, compression ratio, number of cylinders, size of
displacement, and others. Hybrid and electric powertrains are also described in tiers.
Transmission technologies include arrangements like manual, 6-speed automatic, 8-speed
automatic, continuously variable transmission, and dual-clutch transmissions. With a portfolio
of descriptive technologies, NHTSA can summarize the analysis  fleet, and project how vehicles
in that fleet may improve over time via the application of advanced technology.

   In many cases, technology is clearly observable, but in some cases technology levels less
discrete in nature. For the latter, like tiers of mass reduction, NHTSA conducted careful analysis
to describe the level of technology already used  in a given vehicle. Similarly, NHTSA uses
engineering judgement to determine if higher mass reduction tiers may be used practicably and
safely in a given vehicle.

   Most manufacturers provided a summary of observable technology used in each of their
vehicles. In some cases, NHTSA/Volpe supplemented supplied information with data available
to the public, typically from manufacturer media sites.  In limited cases, manufacturers did not
supply adequate information, and NHTSA/Volpe Center  staff used information from commercial
and publicly available information.

4.2.6.3 Engine Utilization

   Manufacturers submitted many details about  engines and transmissions to NHTSA. NHTSA
used submissions to understand the current level of technology in the fleet and to estimate
powertrain families.

   NHTSA catalogued engine and transmission  specifications as  part of the CAFE model input.
For engines, NHTSA recorded number of cylinders, displacement, valvetrain configuration,
aspiration, fuel type, compression ratio, power output, and others. For transmissions, NHTSA
recorded number of forward gears, automatic or manual,  driveline configuration  (front-wheel
drive, rear-wheel drive, all-wheel drive), and others. With an index of current equipment in the
fleet, the CAFE model can project pathways for manufacturers to adapt and to adopt
technologies and comply with regulations.
                                             4-58

-------
                                                    Baseline and Reference Vehicle Fleets
   Similar to vehicle platforms, the CAFE model considers engine platforms.  Manufacturer
submissions varied widely in the degree to which engines were identified as unique, shared, or
sharing common components. In some cases, manufacturers designated each engine in each
application as a unique powertrain. For instance, a manufacturer may have listed two engines for
a pair that share designs for the engine block, the crank shaft, and the head because the accessory
drive components, oil pans, and engine calibrations differ between the two. In practice, many
engines share parts, tooling, and assembly resources, and manufacturers often coordinate design
updates between two similar engines.  For the all engine portfolios, NHTSA/Volpe Center staff
tabulated engine families. By grouping  engines together, the CAFE model explores future
product portfolios with reasonable powertrain complexity.

   NHTSA assigned engines to families based on data driven criteria. If engines share a
common cylinder count and configuration, valvetrain, and fuel type NHTSA considered
grouping engines together. Additionally, if the compression ratio, horsepower, and displacement
differed by no more than 15 percent, the engines were considered to be  the same for the purposes
of redesign and sharing.  Similarly, in some cases NHTSA consolidated the number of
transmission designs for a manufacturer. As a result, for manufacturers that submitted highly
atomized engine and transmission portfolios, there is a practical cap on  powertrain complexity
and the ability of the manufacturer to optimize (a.k.a. "right size") engines perfectly for each
vehicle configuration.

4.2.7  Estimated Technology Prevalence in the MY2015 Fleet

   The following tables show the estimated prevalence of major technologies, by sales volume
weighting, in the MY2015 Light Duty analysis fleet. Numbers provided may differ from actual
penetration rates based on projected sales and technology take rates.  Separate tables cover
conventional engine technologies, electrification technologies, and transmission technologies.
                                             4-59

-------
                                                    Baseline and Reference Vehicle Fleets
                        Table 4.43  Engine Technologies by Manufacturer.
Manufacturer
BMW
Daimler
FCA
Ford
General Motors
Honda
Hyundai Kia
Jaguar/ Land Rover
Mazda
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volvo
VWA
Light Duty Fleet
Diesel
4
5
3
0
0
0
0
0
0
0
0
0
-
0
0
14
1
DOHC
100
99
68
100
64
51
100
100
100
89
100
100
-
100
100
100
85
VVT
96
79
96
100
91
51
100
100
100
98
100
100
-
99
100
81
92
VVL
95
0
18
0
7
100
0
0
92
11
6
0
-
1
0
30
19
SGDI
95
93
0
61
87
48
85
100
86
0
3
4
-
5
37
81
45
Cylinder
Deactivation
0
0
14
0
38
32
1
0
0
0
0
0
-
0
0
2
12
Turbo- or Super-
Charging
100
69
6
43
11
0
1
100
0
3
2
4
-
1
92
87
17
   Few manufacturers rely on diesel engines for a large portion of sales. All manufacturers have
deployed DOHC and VVT across the majority of the light duty fleet. Adoption of VVL, SGDI,
cylinder deactivation, and air intake charging vary widely across the fleet and across
manufacturers.
                                              4-60

-------
                                                    Baseline and Reference Vehicle Fleets
                     Table 4.44  Electrification Technologies by Manufacturer.
Manufacturer
BMW
Daimler
FCA
Ford
General Motors
Honda
Hyundai Kia
Jaguar/ Land Rover
Mazda
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volvo
VWA
Light Duty Fleet
SS12V
93
85
0
0
7
0
0
92
0
0
0
0
-
0
0
0
6
BISG/CISG
0
0
0
0
0.1
0
0
0
0.5
0
0
0
-
0
0
0
0
SHEV
0
0
0
2
0
1
2
0
0
0
0
2
-
9
0
0
2
PHEV
0.1
0
0
0.7
0.5
0
0
0
0
0
0
0
-
0.2
0
0.3
0.2
EV
0.1
0.8
0.5
0.1
0.1
0
0.1
0
0
0
1.2
0
100
0
0
0.5
0.4
   Many manufacturers have offered some type of alternative, electric powertrain to the market;
however, electrification technologies currently have very modest market share.  A few
manufacturers have reported use of 12V start-stop systems, but very few report use of BISG or
CISG systems. Many manufacturers offer some combination of strong hybrids and plug-in
hybrids, but only Toyota has sales in these categories approaching 10 percent of total sales
volume.  Most manufacturers have dabbled with commercializing electric vehicles, but only
Tesla remains fully committed to pure battery electric vehicle technology.  Vehicles with
electrification technologies continue to form a small fraction of the total light duty fleet.
                                              4-61

-------
                                                    Baseline and Reference Vehicle Fleets
                      Table 4.45 Transmission Technology by Manufacturer.
Manufacturer
BMW
Daimler
FCA
Ford
General Motors
Honda
Hyundai Kia
Jaguar/ Land Rover
Mazda
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volvo
VWA
Light Duty Fleet
Manual
4
0
3
6
1
3
2
0
9
8
2
7

1
0
7
3
CVT
0
0
1
2
1
63
0
0
0
90
83
93

16
0
2
20
AMI or DCT
3
0
1
6
0
1
2
0
0
0
0
0

0
0
91
4
Auto, 6+ speeds
93
100
94
86
98
33
96
100
91
3
15
0

83
100
0
73
   The biggest trend for transmissions is that manufacturers are offering more speeds in
automatics.  Many six, seven, eight, and nine-speed automatic transmissions have entered the
fleet, and manufacturers have announced publicly that ten-speed automatics will be widely
available soon.  Manufacturers who have limited deployment of six speed or higher automatic
transmissions have committed to continuously variable transmissions.  Despite the promise of
high efficiency, early launches of dual-clutch transmissions have been plagued with drivability
complaints, and the technology has seen limited application.  Manual transmissions remain a
niche technology for specialty performance vehicles and entry level vehicle packages.
Conventional transmissions with six or more speeds makeup approximately 73 percent of the
2015 analysis fleet.

4.2.8   Engine and Platform Sharing

   Over the past several decades, manufacturers have expanded product offerings to consumers
at a rapid rate. Manufacturers share and standardize components, systems, tooling, and assembly
processes within their products (and occasionally with the products of another manufacturer) to
cost effectively maintain vibrant portfolios.  A "platform" refers to engineered underpinnings
shared on several differentiated products.

4.2.8.1 Platform Sharing

   The concept of platform sharing has evolved with time. Years ago, manufacturers rebadged
vehicles and offered more exotic options on premium nameplates. Today, manufacturers share
parts across highly differentiated vehicles. Engineers design chassis platforms with the ability to
vary wheelbase, ride height, and even driveline configuration.  Assembly lines can produce
                                              4-62

-------
                                                   Baseline and Reference Vehicle Fleets
hatchbacks and sedans with large overlaps in manufacturing capacity.  Engines made on the
same line may power small cars or mid-size sport utility vehicles. Many manufacturers,
including Ford, General Motors and Toyota have publicized strategies to reduce complexity with
expanded use of common platforms.  Now, vehicles with different looks and different
capabilities may share the same platform.

   Although NHTSA's analysis, like past CAFE analyses, considers vehicles produced for sale
in the U.S., the agency notes that these platforms are not constrained to vehicle models built for
sale in the United States; many manufacturers have developed, and use, global platforms. And
the number of global platforms is shrinking across the industry. Several automakers (for
example, General Motors and Ford) either plan to, or already have, reduced their number of
platforms to fewer than ten and account for the overwhelming majority of their production
volumes on that small number of platforms.

   The CAFE model accounts for platform sharing and complexity management within the
context of production for sale in the U.S.  The model restricts significant advances in some
technologies, like major mass reduction, to major redesign years. If one vehicle on the platform
receives a  treatment of technology, other vehicles on the platform also receive the technology as
part of their next major redesign or refresh.

4.2.8.2 Engine Sharing & Inheritance

   Similar to vehicle platforms, manufacturers create engines that share parts.  For instance,
common engine block castings may be bored out with marginally different diameters to create
engines with an array of displacements. Head assemblies for different displacement engines may
share many components across the engine family.  Crankshafts may be finished with the same
tools, to similar tolerances. One engine family may appear on many vehicles on a platform, and
changes to that engine may or may not carry through to all the vehicles. Some engines are
applied across a range of vehicle platforms.

   The CAFE model currently accounts for sharing of engines by "truing up" technology among
vehicles that share the same engine. If such vehicles have different design schedules, and a
subset of vehicles using a given engine add engine technologies in the course of a redesign or
freshening that occurs in an early model year (e.g., 2018), other vehicles using the same engine
"inherit" these technologies at the soonest ensuing freshening or redesign.  This is consistent
with a view that, over time, most manufacturers are likely to find it more practicable to shift
production to a new version of an engine than to indefinitely continue production of a "legacy"
engine.

   The CAFE model does not currently attempt to simulate the potential that, having no further
regulatory need to improve fuel economy, a manufacturer might shift the application of
technologies that improve technical efficiency to favor performance rather than fuel economy.
Therefore, the model's representation of the "inheritance" of technology can lead to estimates
that a manufacturer might eventually exceed fuel economy standards as technology continues to
propagate  across shared platforms and engines. Historical CAFE compliance data shows
examples of extended periods  during which some manufacturers exceeded one or both  standards.
On the other hand, notwithstanding the potential that doing so would reintroduce complexity that
would come at some cost (e.g., to replace a naturally aspirated engine with a smaller
turbocharged engine, and subsequently split the newer engine into versions with multiple
                                             4-63

-------
                                                   Baseline and Reference Vehicle Fleets
displacements), NHTSA recognizes that buyers could continue to place enough value on vehicle
performance and utility that a manufacturer would, having achieved compliance, take advantage
of opportunities to cost-effectively shift technical capability in those directions. Still, the
prospect of "splintering" engines and platforms may limit the extent to which manufacturers
attempt to finely balancing fuel economy and performance for each vehicle configuration.

  NHTSA will consider options to further refine its representation of sharing and inheritance of
technology, possibly including model revisions to account for tradeoffs between fuel economy
and performance when applying technology. The agency seeks comments on the sharing- and
inheritance-related aspects of its analysis fleet and the CAFE model, and information that would
support refinement of the current approach or development and implementation of alternative
approaches.

4.2.9   Class Types and Assignment

  The CAFE model makes use of four distinct class assignments: Regulatory Class, Safety
Class, Technology Class, and Technology Cost Class.

4.2.9.1 Regulatory Class

  Regulatory Class is a straightforward classification by Passenger Car or Light Truck (PC or
LT). Assignment to PC  or LT is defined by the criteria set forth in the corporate average fuel
economy rules.

4.2.9.2 Safety Class

  Each vehicle in the input fleet receives a Safety Class designation based on vehicle body style
and vehicle weight. NHTSA uses safety class to conduct safety analysis, discussed  separately.

4.2.9.3 Technology Class

  Technology Class maps vehicle models in the analysis fleet to a set of Argonne simulation
results that provide effectiveness values for each technology.  Argonne currently supports five
Technology Classes: (1) small car, (2) small SUV, (3) medium car, (4) medium SUV and (5)
pickup.  NHTSA assigns technology classes in the following way:

   •   All vehicles with Body Style = Pickup are  classified as a Pickup. All body-on-frame
       vehicles are classified as Pickups, so some Vans and SUVs appear in the Pickup
       technology class.
   •   Big SUVs with unibody construction are medium SUVs. Medium SUVs also include
       vehicles with van body styles and vehicles with minivan body styles. Generally, SUVs
       with a larger than average footprint are  designated medium SUVs.
   •   The small  SUV technology class includes all vehicles with a wagon body style. In
       addition, SUVs that have a smaller than average footprint also earn a small SUV
       technology class assignment.
   •   Passenger cars with a greater than mean footprint are medium cars. The medium car
       technology class includes convertibles,  coupes, hatchbacks, and sedans.
   •   Passenger cars with a less than mean footprint are small cars.  The small car technology
       class includes convertibles, coupes, hatchbacks, and sedans.
                                             4-64

-------
                                                   Baseline and Reference Vehicle Fleets
4.2.9.4 Technology Cost Class

   Technology Cost Class accounts for costs that vary by engine configuration (e.g. SGDI,
VVT), and therefore provides a code for the number of cylinders, banks, and whether or not a
vehicle uses an OHV valve train configuration. For example, 4C1B indicates an inline 4-
cylinder engine with a conventional valvetrain, while 8C2B_ohv indicates a V8 engine with an
OHV valvetrain configuration.

   NHTSA seeks comment on this approach to grouping specific vehicles for these different
analytical requirements, recommendations regarding any alternative approaches, and information
that could be used to refine the assignment of specific vehicles to specific categories.

4.2.10 Mass Reduction and Aero Application

   Unlike other technologies like valvetrain configurations or transmission arrangements, the
degree of mass reduction already applied to a vehicle is not always straightforward to assign as a
generic "level," Vehicles with lower mass and less aerodynamic drag often have higher
performance. More so than other technologies, vehicle mass and aerodynamics are the product
of hundreds of engineering decisions, material choices, design strategies and manufacturing
approaches that together makeup a vehicle.  The utility a vehicle provides a customer affects a
vehicle's mass and aerodynamic characteristics: the general shape, number of openings, surface
features of the car, and optional equipment factor into mass and aerodynamic performance.

   NHTSA recognizes that in many cases manufacturers have already implemented mass savings
technologies and drag reductions on many of their 2015MY products.  As a result, not all
vehicles in the analysis fleet have the same opportunities to further reduce mass and improve
aerodynamic drag in future years.  To account for the diverse progress on mass reduction and
aerodynamics among the analysis fleet, NHTSA assigned each vehicle a level of mass reduction
and aerodynamic treatment relative to a baseline case. NHTSA has adopted a relative
performance approach to assess the application of mass reduction and aerodynamic technologies.

4.2.10.1      Mass Reduction

   NHTSA developed cost curves for glider weight savings on baseline sedans and pick-ups.  In
order for NHTSA's cost curves to be used effectively in the NHTSA/Volpe model, vehicles in
the analysis fleet must start at a position on the estimated cost curve that reflects the level of
mass reduction technology currently used on the platform.  This section describes the assignment
process and summarizes the mass reduction assignment results.

   NHTSA/Volpe Center staff developed regression models to estimate curb weights based on
other observable  attributes.  With regression outputs in hand, Volpe evaluated the distribution of
vehicles in the analysis fleet. Additionally, NHTSA/Volpe evaluated vehicle platforms based on
the sales-weighted residual of actual vehicle  curb weights vs. predicted vehicle curb weights.
Based on the actual curb weights relative to predicted curb weights, NHTSA/Volpe assigned
platforms (and the  subsequent vehicles) a 2015MY mass reduction level.

   For the curb weight regressions, NHTSA/Volpe  Center staff grouped vehicles in the analysis
fleet into three separate body design categories for analysis: 3-Box, 2-Box, and Pick-up.
                                             4-65

-------
                                                    Baseline and Reference Vehicle Fleets
                           Table 4.46 Mass Reduction Body Style Sets
3-Box
Coupe
Sedan
Convertible
2-Box
Hatchback
Wagon
Sport Utility
Minivan
Van (LT)
Pick-up
Pick-up (LT)
   NHTSA/Volpe Center staff leveraged many documented variables in the analysis fleet as
independent variables in the regressions.  Continuous independent variables included footprint
(wheelbase x track width), and powertrain peak power.  Binary independent variables included
strong HEV (yes or no), PHEV (yes or no), BEV or FCV (yes or no), all-wheel drive (yes or no),
rear-wheel drive (yes or no), and convertible (yes or no). Additionally, for PHEV and BEV /
FCV vehicles the capacity of the battery pack was included in the regression as a continuous
independent variable. In some of the body design categories, the analysis fleet did not cover the
full spectrum of independent variables. For instance, in the pickup body style regression, there
were no front-wheel drive vehicles in the analysis fleet, so the regression defaulted to all-wheel
drive and left an independent variable for rear-wheel drive.
                                             4-66

-------
                                                  Baseline and Reference Vehicle Fleets
Table 4.47 Regression Statistics for Curb Weight (Ibs.)

Observations
Adjusted R Square
Standard Error
REGRESSION STATISTICS
Intercept
Footprint (sqft)
Power (hp)
Strong HEV (1,0)
PHEV(1,0)
BEVorFCV(l,0)
Battery pack size ( KwH)
AWD(1,0)
RWD(1,0)
Convertible (1,0)
3-Box
822
0.865
228.7
Coefficients
-1581.63
100.50
1.22
200.36
259.28
602.33
-2.48
294.51
117.20
273.65
Standard Error
98.5
2.2
0.1
46.3
96.8
215.0
4.1
24.5
23.7
25.3
4^
O
5!
-16.06
44.79
14.85
4.33
2.68
2.80
-0.60
12.03
4.94
10.84
3
o
i>
Q_
0.0000
0.0000
0.0000
0.0000
0.0075
0.0052
0.5461
0.0000
0.0000
0.0000
Lower 95%
-1775.0
96.1
1.1
109.5
69.3
180.3
-10.6
246.4
70.6
224.1
Upper 95%
-1388.3
104.9
1.4
291.2
449.2
1024.3
5.6
342.6
163.8
323.2
2-Box
584
0.883
332.8
Coefficients
-1930.09
104.72
3.09
358.97
462.90
374.24
-1.32
353.91
208.02
-
Standard Error
142.5
3.6
0.2
80.3
169.7
152.1
3.7
33.4
54.1
-
4^
O
5!
-13.54
28.69
13.42
4.47
2.73
2.46
-0.36
10.59
3.84
-
|
§
Q_
0.0000
0.0000
0.0000
0.0000
0.0066
0.0142
0.7187
0.0000
0.0001
-
Lower 95%
-2210.0
97.5
2.6
201.3
129.5
75.5
-8.5
288.3
101.7
-
Upper 95%
-1650.2
111.9
3.5
516.6
796.3
673.0
5.9
419.5
314.3
-
Pick-up
453
0.461
318.1
Coefficients
1857.77
41.67
1.57

-

-

-240.32
-
Standard Error
194.3
3.2
0.3

-

-

30.2
-
o
&>
9.56
12.92
5.11
-
-
-
-
-
-7.96
-
3
o
i>
Q_
0.0000
0.0000
0.0000

-

-

-
-
Lower 95%
1475.9
35.3
1.0

-

-

-299.7
-
Upper 95%
2239.7
48.0
2.2

-

-

-181.0
-
                         4-67

-------
                                                     Baseline and Reference Vehicle Fleets
   The regression for pickup body style did not include independent variables for strong HEV,
PHEV, BEV or FCV, battery pack size, or convertible.  No vehicles in the analysis fleet matched
these criteria for the pick-up body style. Additionally, with the inclusion of the 2015MY Ford F-
150, a large portion of the pickup sample set is known to have adopted a significant amount of
weight savings technology.

   Each of the three regressions produced outputs that were effective for identifying vehicles
with significant amount of mass reduction technology in the 2015MY analysis fleet. Many of
the coefficients for independent variables provided clear insight into the average weight penalty
for the utility feature. In some cases, like battery size, the relatively small sub-sample size and
high collinearity with other variables confounded the coefficients.  This was especially true for
advanced PFLEV's and BEV's, which are often vehicles that include high levels of weight saving
technology on the vehicle glider.  By design, no independent variable directly accounted for the
degree of weight savings technology applied to the vehicle.  The residuals of the regression
captured weight reduction efforts and noise from other sources.
                    Predicted Curb vs. Actual Curb Weight by Body Style
        8,000
                                                                             3-Box

                                                                             2-Box

                                                                             Pick-up

                                                                             Regression

                                                                             MR3

                                                                             MRS

                                                                             Linear (Regression)

                                                                             Linear (MR3)

                                                                             Linear (MRS)
             0      1,000    2,000    3,000    4,000    5,000   6,000    7,000
                              Predicted Curb Weight{lbs.)
   Figure 4.5 shows a plot of results from each of the three regressions on a predicted curb
weight vs. actual curb weight.  Points above the thick dashed "regression" line represent vehicles
heavier than predicted; points below the thick dashed "regression" line represent vehicles lighter
than predicted. For points with actual curb weight below the predicted curb weight,
NHTSA/Volpe Center staff used the residual as a percent of predicted weight to get a sense for
the level current mass reduction technology used in the vehicle, as described in inputs to the
CAFE model (MRO, MR1, MR2, MRS, MR4, and MRS).

   Generally, the residuals of the regressions as a percent of predicted weight appropriately
stratified vehicles by mass reduction level. Most vehicles showed positive residuals or had
                                              4-68

-------
                                                    Baseline and Reference Vehicle Fleets
actual curb weights very close to the predicted curb weight.  Very few vehicles in the analysis
fleet were identified with the highest levels of mass reduction. Most vehicles with the largest
negative residuals have adopted advanced weight savings technologies at the most expensive end
of the cost curve.
8,000
7,000
_ 6,000
i/i
]T 5,000
op
5 4,000
D
^ 3,000
^
** 2,000
1,000
0
(
Predicted Curb vs. Actual Curb Weight by Body Style






!





1
0 ° X.






£j!P





A A at
pr




$
?"%> 0




p1
W^'
jTfin





x *





3-Box
A 2-Box
n Pick-up
Regression
MR3
MRS
	 Linear (Regression)
Linear (MRS)
) 1,000 2,000 3,000 4,000 5,000 6,000 7,000
Predicted Curb Weight (Ibs.)
                 Figure 4.5  Mass Reduction Regression Residual Plot by Body Style
   The CAFE model trues up levels of applied mass reduction within a platform, so vehicles that
share the same platform receive a common starting point for mass reduction. This approach for
assigning platforms levels of mass reduction reflects the observation that many weight savings
opportunities, for instance in body and chassis structure, are shared across the platform. The
platform approach also dampens the impact of potential weight variation by trim level on the
analysis. To determine the starting level of mass reduction for each platform NHTSA/Volpe
staff computed a sales-weighted average residual of all the vehicle variants for each platform.
Based on the MY2015 platform average residual, NHTSA/Volpe staff assigned  an initial level of
mass reduction to the platform and corresponding vehicles.
                       Table 4.48  Mass Reduction Levels by Residual Error
Mass Reduction
Technology
Assignment
MRO
MR1
Residual as a Percent
of Predicted Curb
Weight
Predicted
-3.75%
                                              4-69

-------
                                                    Baseline and Reference Vehicle Fleets
MR2
MRS
MR4
MRS
-5.625%
-7.5%
-11.25%
-15.0%
   With an 'MR' assignment, the CAFE model factors in that vehicles approach additional
weight savings opportunities from different starting points, and vehicles may face incrementally
higher or lower costs to shed additional weight.
\
8,000
7,000
_^ 6,000
t/i
.a
+T 5,000
ap
01
5 4,000
_a
3
^ 3,000
^
•H*
u
<
2,000
1,000
0
(
Vehicle Mass Reduction (MR) Assignments by Common Platform













A
(2
X





^
'-''•''




-*Z&
fctH
?\





A;Sj
-------
                                                   Baseline and Reference Vehicle Fleets
       Table 4.49 Vehicle Platforms with Highest Estimated Levels of Mass Reduction Technology
CAFE MR Group
MRS
MR4
MRS
MR2
NHTSA/VoIpe Platform Code
VWA Veneno
VWAPorshe 918
GM Sigma
BMW 13
FCA4
BMW 18
VWA Aventador
GMY
Toyota_B
Nissan FF-1
Daimler Daimler_R197
Hyundai Kia HKJ5
General Motors MST
Hyundai Kia HK_UB
Mazda SkyActive_BM
Mazda NC
Ford Ford_F
Toyota FR_S
VWAVW MSS
Hyundai Kia HK_Sedona
Hyundai Kia HK_PS
Mazda SkyActive_GJ
Daimler Daimler MRA
Honda HONDA_PILOT
VWA Veyron
JLRXJ
Daimler Daimler_W246
Nissan FF-3
Example Nameplate
Lamborghini Veneno Roadster
Porsche 918 Spyder PHEV
Cadillac CTS-V Wagon
BMW i3 PHEV
Alfa Romeo 4C
BMW i8 PHEV
Lamborghini Aventador
Chevrolet Corvette
Toyota Prius C
Nissan Versa
Mercedes SLS AMG GT Roadster
Hyundai Elantra
GMC Canyon
Kia Rio
Mazda 3
Mazda MX-5
Ford F-150
Toyota FR-S
Audi R8
Kia Sedona
Kia Soul
Mazda 6
Mercedes C 300
Honda Odyssey
Bugatti Veyron
Jaguar XJ
Mercedes CLA 250
Nissan Altima
MR Residua 1%
-27.6%
-26.5%
-25.7%
-16.9%
-15.9%
-15.3%
-15.0%
-11.4%
-11.2%
-10.8%
-10.5%
-9.6%
-9.3%
-9.1%
-9.1%
-8.8%
-8.2%
-8.1%
-8.1%
-7.8%
-7.5%
-7.2%
-6.8%
-6.8%
-6.6%
-6.1%
-6.0%
-5.9%
   MRS vehicles included the BMW i3, BMW i8, and some exotics. The Chevrolet Corvette
received an MR4. The newly redesigned Ford F-150 and the recently redesigned GMC Canyon
received MRS. The Mazda6 was binned as MR2. The Honda Civic was assigned MR1, with a
platform residual very near the boundary for MR2.  The 2011MY Honda Accord and the
2014MY Chevy Silverado served as benchmark vehicles as NHTSA developed cost curves for
weight savings. The actual vs. predicted weight for each benchmark vehicle falls very near the
predicted curb weight based on their independent variable vehicle attributes, and each vehicle
would be assigned MRO. Both the MY2015 Honda Accord and MY2015 Chevy Silverado are
MRO vehicles. The table below summarizes the initial levels of mass reduction assigned for each
manufacturer's MY2015 light-duty fleet.
                                             4-71

-------
                                                    Baseline and Reference Vehicle Fleets
        Table 4.50 2015MY Mass Reduction Level by Manufacturer as a Percent of Vehicle Sales
Manufacturer
VWA
General Motors
BMW
FCA
TOYOTA
Nissan
Daimler
Hyundai Kia
Mazda
Ford
Honda
JLR
Tesla
Mitsubishi
Volvo
SUBARU
MRO
99.68%
95.71%
99.69%
91.33%
97.58%
17.33%
59.35%
32.13%
9.76%
76.44%
52.84%
93.95%
0.00%
100.00%
100.00%
100.00%
MR1
0.00%
0.00%
0.00%
8.64%
0.00%
32.64%
0.00%
26.47%
32.77%
0.00%
29.86%
0.00%
100.00%
0.00%
0.00%
0.00%
MR2
0.01%
0.00%
0.00%
0.00%
0.00%
40.61%
40.61%
0.00%
19.33%
0.00%
17.30%
6.05%
0.00%
0.00%
0.00%
0.00%
MRS
0.23%
3.21%
0.00%
0.00%
2.42%
9.42%
0.03%
41.40%
38.15%
23.56%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
MR4
0.05%
1.06%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
MRS
0.04%
0.02%
0.31%
0.02%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
0.00%
   NHTSA seeks comment on this approach to assigning initial levels of mass reduction, and
recommendations regarding any alternative approaches, taking into account the agency's
representation of costs and fuel consumption impacts of additional mass reduction. The agency
seeks any additional information that could be used to refine the agency's approach or develop
and implement alternative approaches.

   As part of the mass reduction regression analysis, NHTSA/Volpe staff evaluated trends in
residuals. Based on prior work in the industry and observations from this analysis, a more
detailed summary of residuals with respect to vehicle footprint, luxury content, and company
heritage is included below.
                                              4-72

-------
                                                    Baseline and Reference Vehicle Fleets
All Vehicle Platforms Residual Histogram
T C
t/1
£ 20
£ 2°
*j
to
°- ir
0) ±;>
u
1 m
4-
3 i 1 1 I I
0 1 . .11111
tfv tfv tfv tfv tfv tfv tfv ^v ^v
ooooooooo
;
c

'
;
;
c


1
1 1 . 1 . 1 . 1
V tfv ^v. ^v. ^v. S^. S^. S"v. S^.
ooooooooo
lDrxJ!X3'*OlDrxjOd'd:O'*Od(NtDO'^:Od(NtD
rnmrvirvirvirHrH iHtHrvirvirxjmm
Regression Residual: Lightweight (left), Overweight (right)
• All Vehicle Platforms
              Figure 4.7 Mass Reduction Residual Histogram for All MY2015 Platforms
4.2.10.1.1    Mass Reduction Residual Analysis for Footprint

   NHTSA/Volpe staff identified a meaningful trend in the regression residuals for vehicle
footprint: vehicles under 41 square foot footprint tended to have large residuals as a percentage
of predicted weight.  The two smallest vehicles were estimated to be the most overweight based
on content modeled in the regression.
                                              4-73

-------
                                                     Baseline and Reference Vehicle Fleets
        Table 4.51 Mass Reduction Platform Residuals for Platforms with the Smallest Footprint
Sales
Weighted
Platform
Average
Footprint
(sq.ft.)
26.8
34.8
36.1
37.4
38.7
39.9
40.0
40.1
40.2
40.5
40.8
40.9
41.1
41.2
41.5
42.1
Platform%
Residual
33.1%
32.3%
11.2%
-8.8%
1.1%
6.2%
-3.8%
-11.2%
-0.6%
-1.1%
-1.3%
0.0%
-15.9%
3.9%
-10.8%
-3.6%
Rank of
Smallnessof
Platform
Footprint
(out of 128)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Rank of
Heaviness on
a Residual%
Basis
(out of 128)
1
2
7
112
48
15
91
119
69
72
76
60
123
24
118
90
Example Vehicle from Platform
Smart ForTwo
Fiat 500
Chevrolet Spark
Mazda MX-5
Mini Cooper Coupe
Porsche 911 Carrera
Honda CR-Z
Toyota Prius C
Ford Fiesta
Mini Cooper Hardtop, 4-door
Porsche Boxster
Chevrolet Sonic
Alfa Romeo 4C
BMWZ4
Nissan Versa
Mitsubishi Lancer
Assigned
MR Value
MRO
MRO
MRO
MRS
MRO
MRO
MR1
MRS
MRO
MRO
MRO
MRO
MRS
MRO
MRS
MRO
   The NHTSA/Volpe staff proposes that this trend is a result of limited crush space in the
smallest vehicles, so on a relative basis the smallest vehicles may include more mass in structure
for a given set of content than their larger counterparts.  As shown in the table above, and the
figure below, this trend subsides after the platform exceeds a sales weighted average footprint of
about 41 square feet.
                                               4-74

-------
                                                      Baseline and Reference Vehicle Fleets
                         Mass Reduction Residual Trends in Footprint
ion Residua
Reg
41
U)
KJ
H»
r
X



X

X
>

x
f
X
tf i^yX^TSvM.

< X
xv x
A



x
x




   _^
   CL.
   •~
   O
   i_i
   3
   £
-10%




X



^ »
< X "x X
X x XX
x$

X



X







   •=  -40%
          20           30           40           50           60           70           80
                              Sales-Weighted Average Platform Footprint (sqft)

                                       X Platform Residual

                     Figure 4.8 Mass Reduction Platform Residuals vs. Footprint

   Chapter 8 discusses the agencies updated assessment of the effects of vehicle mass and size
on overall societal safety.  The complex relationship between a vehicle's mass, size, and fatality
risk varies in different types of crashes, and NHTSA and others have been examining this
relationship for over a decade.  The principal findings and conclusions of NHTSA's updated
mass-size safety analysis are that mass reduction in heavier light-duty trucks, while holding
footprint constant, reduces societal fatality risk, whereas mass reduction in lighter passenger cars
increases overall societal fatality risk.  The agencies investigated the amount of mass reduction
that is projected to maintain overall fleet safety. For the Draft TAR analyses, the agencies have
limited the amount of mass reduction applied to passenger cars to achieve  a safety neutral
outcome.  Therefore technology pathways shown by the agencies' analyses have a neutral effect
on overall fleet safety. Based on such results, additional application of mass reduction
technology is restricted,  according to three criteria shown in Table 4.52.
    Table 4.52 Criteria for Limiting Additional Application of Mass Reduction Technology in the CAFE
                                           Analysis
  Platforms with a sales weighted average of less than 2800 Ibs. may not apply more mass reduction technology.
  SmallCar vehicles may not add new MR technology to proceed past MR2.
  MediumCar vehicles may not add new MR technology to proceed past MR2.
   As a result of these criteria, the model will not apply excessive mass reduction technologies to
small vehicles.
                                                4-75

-------
                                                      Baseline and Reference Vehicle Fleets
4.2.10.1.2     Mass Reduction Residual Analysis for Low and High Price Platforms

   In 2015, the California Air Resources Board published a study, "Technical Analysis of
Vehicle Load Reduction Potential for Advanced Clean Cars"29 that evaluated the distribution of
applied mass reduction technology in the fleet. The study used similar modeling techniques as
used for today's CAFE analysis.  As part of that study, skewed residuals of 1.6 percent were
observed for luxury sedans, and this was reasonably explained optional luxury content. With the
result of those findings in mind, the NHTSA/Volpe evaluated the residuals for platforms with
low base prices and with high base prices to investigate if some form of additional content
should be accounted for in the regression.
             Table 4.53 Mass Reduction Average Residual by Average Platform Base Price

All Vehicle Platforms
Platform MSRP Average
Base Price
$30k or Less
$30k - $50k
$50k or Greater
Average Residual
-0.6%
-0.5%
-0.5%
-0.8%
Platform Count
128
52
37
39
   While option content may add weight on a vehicle basis, the CAFE analysis assigns levels of
mass reduction at a platform level.  Trends in the residuals do not provide strong evidence that
some variable for premium content is needed to correct for a predicted weight bias among high
priced vehicles.
      25
   in

   !»
   |
   Q.
   _OJ  lj
   O
   15
      10
               All Vehicle Platforms vs. Platforms with Base Price of $30k or Less
                    I  ,
.1    I
                                            q    q   q    q
                                            •f    O       O   ^   00   (N   '-JT1
mrnrvirvirvirHrH                          tHtHrvirvirxjmm
         Regression Residual: Lighter than projected (left), Heavier than projected (right)

                    • All Vehicle Platforms    $30kor Less Platforms
      Figure 4.9 Mass Reduction Residual Distribution of Platforms with Base Price of $30k or Less
                                               4-76

-------
                                                        Baseline and Reference Vehicle Fleets
      25
    5 20
      15
    QJ
    > 10
    c
    3
    o
    u
                 All Vehicle Platforms vs. Platforms with Base Price of $30k-$50k
                    I  ,
                          .1    I
         ooooooooooooooooooo
         ^DrMCo'^o^DrMCO^o^cdrM^DO^cdrM^D
         m   m   rM    rM    rM   t-l   *H                           t-H    t-H    rM   rM   rM    nn   nn


                  Regression Residual: Lighter than projected (left), Heavier than projected (right)




                              • All Vehicle Platforms     $30k-$50k Platforms
    Figure 4.10 Mass Reduction Residual Distribution of Platforms with Base Price between $30k-$50k
   Many of the largest residuals represent high priced platforms, and many of the smallest

residuals also represent high priced platforms.  Lower priced platforms tended to have actual

weights clustered closer to the predicted weight and hence residuals with lower variance.
              All Vehicle Platforms vs. Platforms with Base Price of $50k and Above
      25

    •s,


    I 20
   Q-

   _0)

   Ij
   •s

   I
   o
15




10




 5




 0
                       II
I.I,
^O    ^O   vO

pop


(T>    (V>   rM
                                          CO
                                                   q
                                                   d
                                                  q
                                                  •-f
q

CO
                                                                 pj
                                                                                    CO
                                                                                    "l
                                                                                         g   g
                                                                                   H
                                                                                   m
                   Regression Residual: Lighter than projected (left), Heavier than projected (right)



                            • All Vehicle Platforms     $50kor Greater Platforms




     Figure 4.11  Mass Reduction Residual Distribution of Platforms with Base Price of $50k and Above
                                                  4-77

-------
                                                    Baseline and Reference Vehicle Fleets
4.2.10.1.3    Mass Reduction Residual Trends for Company Heritage

   The NHTSA/Volpe did observe a notable skew based on company heritage. Many vehicle
platforms with Asian parent companies demonstrate a residual skew towards lightweight designs,
or negative residuals when compared with vehicles of other heritage. For the purposes of this
analysis, FCA platforms were binned as "North American" heritage.
              Table 4.54 Mass Reduction Average Residual by Parent Company Heritage

All Vehicle Platforms
North America
Platform Parent
Corrmanv Heritace Europe
Asia
Average Residual Platform Count
-0.6% 128
0.1% 42
0.7% 47
-2.9% 39
Residuals for Platforms with North American Heritage
T C
1/1
E 20
I
°- ^^
_OJ lhl
y
!c
01
>1 n
•s
E
3 c
S 1 I
1 1 ill




I.I. 1 . 1
iDiNOJ'tfoiDrxJai'^O'tfairxJiriO'^odrxJiD
rnmrvirvirvirHrH tHiHrvirvirvimm
Regression Residual: Lighter than projected (left), Heavier than projected (right)
• All Vehicle Platforms North American Parent
          Figure 4.12 Mass Reduction Residuals for Platforms with North American Heritage
                                             4-78

-------
                                                         Baseline and Reference Vehicle Fleets
    E 20
    o
   > 10
   4-
   o
   —•
   c
   3  5
   u
                         Residuals for Platforms with European Heritage
                                 .1
                               .1,
         ooooooooooooooooooo
         ^D    (N   00   ^    O    ^D    (N   00    ^    O    ^   00   (N    {D    O    ^   00    (N    'X1
         rDrDrvJrvJrNjTHTH                           THtHrNlrNlrNjmm

                   Regression Residual: Lighter than projected (left), Heavier than projected (right)



                                • All Vehicle Platforms   • Europe an Parent



              Figure 4.13 Mass Reduction Residuals for Platforms with European Heritage
   Platforms with European heritage exhibit large variance and a modest skew towards positive
residuals.
      25
    E" 20
   Q- , q


   13
   .^
   OJ
   > 10
    0
    U
         lD
         fO
                           Residuals for Platforms with Asian Heritage
                     I  -
       I
- 1    I
         ooooooooooooooooooo
                                                                                     r\i    m    m
                   Regression Residual: Lighter than projected (left), Heavier than projected (right)
                                   All Vehicle Platforms    Asian Parent
                Figure 4.14 Mass Reduction Residuals for Platforms with Asian Heritage
                                                  4-79

-------
                                                   Baseline and Reference Vehicle Fleets
4.2.10.2
Aerodynamic Application
   Similar to mass reduction, NHTSA/Volpe Center staff used a relative performance approach
to assign the current aerodynamic technology level to a vehicle. Different body styles offer
different utility and have varying levels of baseline form drag. Additionally, frontal area is a
major factor in aerodynamic forces, and the frontal area varies by vehicle. NHTSA/Volpe
considered both frontal area and body style as utility factors that affect aerodynamic forces.
NHTSA/Volpe computed an average coefficient of drag (Cd) for each body style segment in the
2015MY analysis fleet from drag coefficients published by manufacturers.  By comparing
coefficients of drag among vehicles that share body styles, the NHTSA/Volpe was able to
estimate the level of aerodynamic improvement already present on specific vehicles.

   NHTSA/Volpe Center staff assigned levels of aerodynamic technology to the 2015 fleet on a
relative basis, based on the average aerodynamic drag coefficient (Cd) by body style and
manufacturer reported drag coefficients. NHTSA  calculated the average Cd for each body style
by grouping vehicles by body  style and then averaging the manufacturer reported or publicly
available drag coefficients for each group.

   In order for a vehicle to achieve AERO 10, the aerodynamic drag coefficient needs to be at
least 10 percent below the calculated average drag coefficient for the body style. In order to
achieve AERO20, the Cd needs to be at least 20 percent better than the body style average.  No
aerodynamic application was assumed for vehicles with no manufacturer reported Cd.

   The table below summarizes the best, worst, and average recorded Cd for each body style.
The table also lists the thresholds for AERO 10 and AERO20 that were used to assign an
aerodynamic tech level for each vehicle.
                     Table 4.55 Aerodynamic Drag Coefficients by Body Style
Body style
Sedan
Coupe
Minivan
Hatchback
Convertible
Wagon
Sport Utility
Van
Pickup
Sample Size
437
175
23
88
92
32
346
21
361
Body style
Average Cd
0.302
0.319
0.326
0.333
0.334
0.342
0.363
0.389
0.395
Body style
Lowest Cd
0.240
0.240
0.290
0.250
0.290
0.290
0.300
0.337
0.360
Body style
Highest Cd
0.370
0.440
0.360
0.370
0.410
0.380
0.540
0.415
0.420
AERO10
0.271
0.287
0.293
0.300
0.301
0.308
0.327
0.350
0.355
AERO20
0.241
0.255
0.261
0.266
0.267
0.274
0.290
0.311
0.316
                                             4-80

-------
                                                     Baseline and Reference Vehicle Fleets
0.600
0.550
TD <-^
t ii 0.500
O ftQ
S" 2 0.450
DC Q
5 "5 0.400
13 c
u .oj 0.350
1 ^ 0.300
ro 0
S u 0.250
0.200
Aerodynamic Coefficient of Drag (Cd) by Vehicle Body Style




e



j
9 r



3
1 i
1 1
• 8 c
f T




s E

) |




c
{





)
| [
1 *





(
(
(








i ||
) I
' 1




c ai c -g ii c-^Qjc Q.
^ Q. SJ ra ^ O .-= TJ ™ 3
i o ^
Vehicle Bodystyle

           Figure 4.15 Distribution of Aerodynamic Drag Coefficients by Vehicle Body Style
   Based on the results of the CAFE input assignment process, most manufacturers have the
opportunity to further improve aerodynamic performance for a large portion of the fleet.
          Table 4.56 Aerodynamic Application by Manufacturer as a Percent of MY2015 Sales
Manufacturer
BMW
Daimler
FCA
Ford
General Motors
Honda
Hyundai Kia
JLR
Mazda
Mitsubishi
Nissan
SUBARU
Tesla
TOYOTA
Volvo
VWA
AEROO
86.7%
41.7%
99.4%
100.0%
99.8%
90.5%
97.9%
100.0%
100.0%
72.9%
93.4%
100.0%
0.0%
74.4%
88.8%
99.2%
AERO 10
13.3%
23.0%
0.6%
0.0%
0.2%
9.5%
2.1%
0.0%
0.0%
27.1%
6.6%
0.0%
0.0%
19.5%
11.2%
0.8%
AERO 20
0.0%
35.3%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
100.0%
6.2%
0.0%
0.0%
                                               4-81

-------
                                                   Baseline and Reference Vehicle Fleets
   NHTSA seeks comment on this approach to assigning initial levels of aerodynamic
performance, and recommendations regarding any alternative approaches, taking into account the
agency's representation of costs and fuel consumption impacts of additional aerodynamic
improvements.  The agency seeks any additional information that could be used to refine the
agency's approach or develop and implement alternative approaches.

4.2.11  Projecting Future Volumes for the Analysis Fleet

   In order to analyze the impact of alternative fuel economy standards in future model years, it
was necessary to estimate vehicle production volumes for each manufacturer (and the models
they offer for sale) in those years. Because the standards are based on the harmonic average of a
manufacturer's fuel economy targets, which are themselves a function of vehicle footprint, the
specific mix of vehicle footprints and regulatory classes that a manufacturer produces in each
model year determines the standard for each manufacturer in that year.

   The CAFE model operates at the level of specific model variants offered by each
manufacturer (insofar as they vary by either footprint, fuel economy, or both), so any projection
of future vehicle volumes must have a comparable resolution. For example, the MY2015
analysis fleet contains several variants of the Ford Fusion, where model variants are
distinguished by drive type (FWD or AWD), engine type (cylinders, displacement), and degree
of hybridization. So it was critical that our projection of future volumes produced estimated
volumes for each variant of the Ford Fusion, rather than simply "the Fusion" or, even more
coarsely, Ford's total volume within a market segment (of which "the Fusion" is a part).

   To generate sales volumes for future model years, we combined three distinct sources of
information about volumes.  The first, and most fundamental, of these is the Mid-Model Year
reports and attribute data that manufacturers supplied to NHTSA. These data informed decisions
about the granularity of the model variants (how many different types of the Ford Fusion, for
example, need to appear in the analysis fleet for modeling) and the relative sales of variants
within a model and market segment for each manufacturer.

   The second source of information used to project volumes is a proprietary production volume
forecast that NHTSA purchased from IHS/Polk that covers the years from 2013 to 2032. This
forecast contains volume projections for each vehicle model that is currently offered for sale in
the United States (below 14,000 Ibs GVW), as well as some legacy models that were phased out
over the last two model years, and future models that have not yet been introduced in the U.S.
market. Despite the high degree of resolution in the Polk forecast, modifications were required
in order to match the level of resolution in the MY2015  analysis fleet.  In particular, the model-
level volume projections in the IHS/Polk forecast were insufficient to account for instances
where one variant of a single model is regulated as a passenger car and another (typically a 4WD
version) as a light truck. In those cases, we manually split the volume forecasts into a passenger
car and a light truck variant based on the shares present in the Mid-Model Year submissions
from manufacturers. We also treated the latest years of the forecast (2029 - 2032) as being
static. While the Polk forecast shows changes in manufacturer market shares in those years,
some of them abrupt, the discontinuities created by those changes are undesirable for a sequence
                                             4-82

-------
                                                     Baseline and Reference Vehicle Fleets
of years that should primarily be driven by trend at that pointx. However, the majority of the
information in the Polk forecast was used, unaltered, to inform the volume projections for the
analysis fleet.

   The third source of volumes comes from a special set of runs of the National Energy
Modeling System, NEMS, which forms the basis of the Energy Information Administration's
Annual Energy Outlook 2015 (AEO 2015). These runs, rather than simulating fuel economy
responses to the augural standards for 2022 - 2025 that NHTSA proposed in 2012, freeze the
fuel economy standards at their 2021 level for the remainder of the model run, which continues
to 2040.  From these runs, we used the total volumes of passenger cars and light trucks
(separately), synthesizing the three sources to approximately preserve these volumes for all
future model years.

   The three data sources were combined sequentially, and the process is depicted graphically in
Figure 4.16, which shows the three data sources in blue and constructed elements in green.
                                        Manufacturer/Model
                                        Level Martcet Shares
                                                                 Future Vehicle
                                                                Volumes 2015-
                                                                     2032
                 Special AEO
               2015 NEMS Run
                                           Total Annual
                                           Vehicle Sates
                Figure 4.16 Data Sources and Construction of the Production Forecast
   We constructed the manufacturer shares in each market segment by combining the AEO total
volumes of passenger cars and light trucks for each (calendar) year with the IHS/Polk volumes
for each manufacturer and body style within each of the passenger car/light truck categories. We
distributed those volumes to each manufacturer's collection of unique model variants in each
body style category based on each model's relative share in the data submitted by the
manufacturers. It was necessary to ensure that the multiple categorizations of vehicle models
x In order to provide a forecast that covers all the years of concern for the Draft TAR, Polk combined information
  from a short-term forecasting model, and a long-term forecasting model. The years that would logically be driven
  by results from the long-term forecasting model were deemed insufficiently volatile for use in the primary
  forecast.
                                              4-83

-------
                                                    Baseline and Reference Vehicle Fleets
across the three sources were synchronized - so regulatory class, body style, and luxury brand
(explained below) were added manually to either the IHS/Polk forecast, the manufacturer
submissions, or both. We attempted to preserve the inherent market preferences represented by
the relative market shares of different vehicle segments in the IHS/Polk forecast. However, there
are many possible characterizations of these  segments, most of them essentially arbitrary. With
that in mind, we chose to use vehicle body style as a proxy for market segment in both the
IHS/Polk forecast and the manufacturer data, ensuring that vehicle models were consistently
categorized across the two sources. Since vehicle body style is a strong indicator of buyer usage
and needs, it seemed a reasonable proxy for the market segments in which these vehicles exist.

   In addition to offering a variety of body styles, many manufacturers have developed luxury
brands that produce higher-end versions of models available in their other brands.  Ford and
Lincoln, for example, produce the Expedition and the Navigator, respectively, which share
engines, transmissions and a common platform, but differ in styling and price.  To the extent that
the IHS/Polk forecast shows migration either to, or away from, luxury versions of comparable
models between 2015 and 2032, we felt that  distinction worth capturing in the synthesized
forecast.  It is less detailed than accounting for volumes within all of a manufacturer's brands
(General Motors produces Buick, Chevrolet, GMC and Cadillac, for example), but superior to
allocating luxury-brand volumes to non-luxury models (or vice versa).

   We calculated the percentage of passenger car and light truck volumes, respectively, in the
IHS/Polk forecast at the level of manufacturer, body style, and luxury brand (or not). Then we
used the total number of passenger cars and light trucks from the AEO runs to calculate the total
sales of each manufacturer's body  style offerings, stratified by luxury (or not) and regulatory
class. Those volumes were then allocated to the model variants in the market data file, based on
the share of volumes for each model variant in the manufacturer, body style, luxury (or not)
stratum. This process was applied such that the total volumes of passenger cars and light trucks
estimated to be produced for the U.S. market aligns with corresponding volumes from AEO2015.

   This process resulted in a market forecast that is broadly consistent with all three sources,
without identically preserving the volumes, or shares, of any one.  A consequence of the
remixing described above is that, in some instances, we show manufacturers exiting the market
(completely) for some body styles in future model years.  The IHS/Polk forecast shows models
entering and leaving the fleet, but we do not  explicitly account for either in the synthesized
forecast. In the case of new model  entrants, the volumes associated with those were allocated to
the remaining models in the manufacturer submissions that already exist within that body style,
luxury, and regulatory class group based on their relative shares. In the case of models exiting
the market segment, those volumes were also re-allocated to the models in that segment as of
model year 2015.  This implies that a manufacturer will always offer all of the current model
variants in a given segment (as defined above)  in future years,  as long as the forecast shows them
offering at least one model in that segment. If the Polk forecast shows a manufacturer  exiting a
market segment (as we've defined them) completely in some future year, then those volumes  are
not re-allocated to any models and are essentially lost to the manufacturer. While this was a rare
occurrence, there are a few instances where this occurs in the synthesized forecast - particularly
for later years.
                                              4-84

-------
                                                           Baseline and Reference Vehicle Fleets
   The forecast used in NHTSA's Draft TAR analysis can be seen in full detail by downloading
the CAFE Model's market data file. However, high level summaries of market shares by
manufacturer appear in Figure 4.17 and Figure 4.18 for model years 2015 and 2025, respectively.
   ReguL Manufacturer
   PC  BMW
       Daimler
       FCA
       Ford
       General Motors
       Honda
       Hyundai Kia
       JLR
       Mazda
       Mitsubishi
       Nissan
       SUBARU
       TOYOTA
       Volvo
       VWA
   LT  BMW
       Daimler
       FCA     |
       Ford     |
       General Motors)
       Honda
       Hyundai Kia
       JLR
       Mazda
       Mitsubishi
       Nissan
       SUBARU
       TOYOTA
       Volvo
       VWA
   Regulatory.Class
                                                      9%   10%   11%
                                                      MY 2015 Market Share
                         Figure 4.17 MY2015 Market Shares by Manufacturer
   While some manufacturers are forecast to gain (or lose) market share between MY2015 and
MY2025 (VW, for example, is forecast to gain small shares in both passenger car and light truck
markets over the next decade), the changes are not dramatically different for any manufacturer
relative to their current market shares.
      . Manufacturer
       BMW
       Daimler
       FCA
       Fnrd
       General Mot<
       Honda
       Hyundai Kia
       JLR
       Mazda
       Mitsubishi
       Nissan
       SUBARU
       TOYOTA
       Volvo
       VWA
       BMW
       Daimler
       FCA     |
       Ford
       General Motors)
       Honda
       Hyundai Kia |
       JLR
       Mazda
       Mitsubishi
       Nissan    |
       SUBARU   |
       TOYOTA   |
       Volvo     ||
       VWA
   Regulatory.Class
                                                     9%   10%   11%
                                                      MY 2025 Market Share
                         Figure 4.18 MY2025 Market Shares by Manufacturer
                                                    4-85

-------
                                                  Baseline and Reference Vehicle Fleets
   NHTSA seeks comment on the information and methods used to develop these estimates of
future production volumes for specific vehicles, and recommendations and additional
information that could be used to refine this approach or develop and implement alternative
approaches.
                                            4-86

-------
                                                           Baseline and Reference Vehicle Fleets
References
1 EPA's Omega Model and input sheets are available at http://www.epa.gov/oms/climate/models.htm; Available in
the docket (Docket EPA-HQ-OAR-2015-0827).
2 "Technical Analysis of Vehicle Load Reduction Potential for Advanced Clean Cars," Controltec, LLC, for CARB,
April 29, 2015.
3 U.S. Environmental Protection Agency (2012). "Regulatory Impact Analysis: Final Rulemaking for 2017-2025
Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards." EPA-
420-R-12-016, Chapter 10.
4 The baseline Excel file ("2014-2025 Production Summary and Data with Definitions") is available in the docket
(Docket EPA-HQ-OAR-2015-0827).
5 Department of Energy, Energy Information Administration, Annual Energy Outlook (AEO) Available at
http://www.eia.gov/forecasts/aeo/tables_ref.cfm (last accessed June. 15, 2016).
6 The baseline Excel file ("2014-2025 Production Summary and Data with Definitions") is available in the docket
(Docket EPA-HQ-OAR-2015-0827).
7 77 Federal Register 62843-62844 (October 15, 2012) and Regulatory Impact Analysis, pages 3-18 to 3-23.
8 National Research Council. (2015) Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles. The National Academies Press, pp. S-9, 9-2, 10-29, 10-38.
9 National Research Council. (2015) Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles. The National Academies Press, pp. S-9, 9-2, 10-29, 10-38.
10 Knittel, C. R. (2011). "Automobiles on Steroids: Product Attribute Trade-Offs and Technological Progress in the
Automobile Sector." American Economic Review 101(7): pp. 3368-3399, Docket EPA-HQ-OAR-2010-0799-11946;
Klier, T. and Linn, J. (2016). "The Effect of Vehicle Fuel Economy Standards on Technology Adoption." Journal of
Public Economics 133: 41-63; McKenzie, D. and Heywood, J. B. (2015). "Quantifying efficiency technology
improvements in U.S. cars from 1975-2009." Applied Energy 157: 918-928; Wang, Y. (2016). "The Impact of
CAFE Standards on Automobile Innovation in the US." Working paper.
11 Cooke, Dave (2016). "The Trade-off between Fuel Economy and Performance: Implications for the Mid-term
Evaluation of the National Program." Union of Concerned Scientists working paper.
12 Knittel, C. R. (2011). "Automobiles on Steroids: Product Attribute Trade-Offs and Technological Progress in the
Automobile Sector " American Economic Review 101(7): pp. 3368-3399, Docket EPA-HQ-OAR-2010-0799-11946;
Klier, T. and Linn, J. (2016). "The Effect of Vehicle Fuel Economy Standards on Technology Adoption." Journal of
Public Economics 133: 41-63; McKenzie, D. and Heywood, J. B. (2015). "Quantifying efficiency technology
improvements in U.S. cars from 1975-2009." Applied Energy 157: 918-928; Wang, Y. (2016). "The Impact of
CAFE Standards on Automobile Innovation in the US." Working paper.
13 MacKenzie, Don, and John Heywood (2012). "Acceleration Performance Trends and Evolving Relationship
Between Power, Weight, and Acceleration in U.S. Light-Duty Vehicles." Transportation Research Record 2287:
122-131.
14 Jaffe, A. B., Newell, R. G., & Stavins, R. N. (2003). Technological change and the environment. In Handbook of
environmental economics 1, ed. K.-G. Maler and J. R. Vincent. Elsevier.
15 Jaffe, A. B., Newell, R. G., & Stavins, R. N. (2003). Technological change and the environment. In Handbook of
environmental economics 1, ed. K.-G. Maler and J. R. Vincent. Elsevier.
16 Dosi, G. and Nelson, R. R. (2010). "Technical Change and Industrial Dynamics as Evolutionary Processes."
Handbook of Econometrics 1: pp. 52-114.
17 Abernathy, W. J., and Utterback, J. M. (1978). Patterns of industrial  innovation. Technology review, 80, pp. 254-
228; Henderson, R. M., and Clark, K. B. (1990). "Architectural innovation: The reconfiguration of existing product
technologies and the failure of established firms." Administrative science quarterly, 9-30.
18 Zhao, H. (Ed.).  (2009). Advanced Direct Injection Combustion Engine Technologies and Development: Gasoline
and Gas Engines (Vol. 1). ISBN 9781845697327. Elsevier. pp. 1-3.
19 U.S. EPA. (2014). Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends:
1975 through 2014. p. 70.
20 Blumstein, Carl and Margaret Taylor (2013). "Rethinking the Energy-Efficiency Gap: Producers, Intermediaries,
and Innovation," Energy Institute at Haas Working Paper 243, University of California at Berkeley, Docket EPA-
HQ-OAR-2014-0827-0075; Tirole, Jean (1998). The Theory of Industrial Organization  Cambridge, MA: MIT
Press, pp. 400, 402, Docket EPA-HQ-OAR-2014-0827-0089.
                                                    4-87

-------
                                                            Baseline and Reference Vehicle Fleets
21 Gal-Or, E. (1987). "First Mover Disadvantages with Private Information." The Review of Economic Studies 54(2):
pp. 279-292; Liu, Z. (2005). "Stackelberg Leadership with Demand Uncertainty." Managerial and Decision
Economics 26: pp. 345-350.
22 Popp, D., Newell, R. G., and Jaffe, A. B. (2010). "Energy, the environment and technological change." In
Handbook of the economics of innovation 2, ed. B. H. Hall, and N. Rosenberg. Elsevier.
23 Holmes, T. I, Levine, D. K., and Schmitz, J. A., Jr. (2012). "Monopoly and the Incentive to Innovate When
Adoption Involves Switchover Disruptions." American Economic Journal: Microeconomics 4(3): pp. 1-33.
24 National Research Council. (2015) Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles. The National Academies Press, pp.  S-9, 9-2, 10-29, 10-38; Klier, T. and Linn, J. (2016). "The Effect
of Vehicle Fuel Economy Standards on Technology Adoption." Journal of Public Economics 133: 41-63; Wang, Y.
(2016). "The Impact of CAFE Standards on Automobile Innovation in the US." Working paper.
25 Boncimino, A. (April 10, 2015). "Auto industry rushes to meet 2025 fuel efficiency guidelines." Upstate Business
Journal, http://upstatebusinessjournal.com/news/auto-industry-rushes-meet-2025-fuel-efficiency-guidelines/;
Buchholz, K. (March 13, 2015). "New technology for exhausting jobs." Automotive Engineering Magazine.
http://articles.sae.org/13974/.
26 Powell, W. W., and Giannella, E. (2010). "Collective Invention and Inventor Networks," In Handbook of the
Economics of Innovation, Volume 1.  Ed. B. Hall and N. Rosenberg Elsevier.
27 Greene, David, and Jin-Tan Liu (1988). "Automotive Fuel Economy Improvements and Consumers' Surplus."
Transportation Research A 22A(3): 203-218. Docket EPA-HQ-OAR-2010-0799-0703.
28 Greene, David (2010). "How Consumers Value Fuel Economy: A Literature Review." EPA-420-R-10-008.
Docket EPA-HQ-OAR-2010-0799-0711.
29 http://www.arb.ca.gov/researcli/apr/pasl/13-313.pdf.
                                                     4-8

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Table of Contents

Chapter 5:  Technology Costs, Effectiveness, and Lead-Time Assessment	5-1
  5.1    Overview	5-1
  5.2    State of Technology and Advancements Since the 2012 Final Rule	5-7
     5.2.1   Individual Technologies and Key Developments	5-7
     5.2.2   Engines: State of Technology	5-12
       5.2.2.1   Overview of Engine Technologies	5-13
       5.2.2.2   Sources of Engine Effectiveness Data	5-15
       5.2.2.3   Low Friction Lubricants (LUB)	5-16
       5.2.2.4   Engine Friction Reduction (EFR1, EFR2)	5-17
       5.2.2.5   Cylinder Deactivation (DEAC)	5-17
       5.2.2.6   Variable Valve Timing (VVT) Systems	5-17
         5.2.2.6.1   Intake Cam Phasing (ICP)	5-18
         5.2.2.6.2   Coupled Cam Phasing (CCP)	5-18
         5.2.2.6.3   Dual Cam Phasing (DCP)	5-18
         5.2.2.6.4   Variable Valve Lift (VVL)	5-18
       5.2.2.7   GDI, Turbocharging, Downsizing and Cylinder Deactivation	5-19
       5.2.2.8   EGR	5-28
       5.2.2.9   Atkinson Cycle	5-29
       5.2.2.10  Miller Cycle	5-33
       5.2.2.11  Light-duty Diesel Engines	5-36
       5.2.2.12  Thermal Management	5-39
       5.2.2.13  Reduction of Friction and Other Mechanical Losses	5-40
       5.2.2.14  Potential Longer-Term Engine Technologies	5-41
     5.2.3   Transmissions: State of Technology	5-42
       5.2.3.1   Background	5-42
       5.2.3.2   Transmissions: Summary of State of Technology and Changes since the FRM 5-
       43
       5.2.3.3   Sources of Transmission Effectiveness Data	5-44
       5.2.3.4   Sources of GHG Emission Improvements: Reduction in Parasitic Losses, Engine
       Operation, and Powertrain System Design	5-46
       5.2.3.5   Automatic Transmissions (ATs)	5-48
       5.2.3.6   Manual Transmissions (MTs)	5-51
       5.2.3.7   Dual Clutch Transmissions (DCTs)	5-52
       5.2.3.8   Continuously Variable Transmissions (CVTs)	5-53
       5.2.3.9   Transmission Parasitic Losses	5-56
         5.2.3.9.1   Losses in ATs	5-56
         5.2.3.9.2   Losses in DCTs	5-56
         5.2.3.9.3   Losses in CVTs	5-57
         5.2.3.9.4   Neutral Idle Decoupling	5-57
       5.2.3.10  Transmission Shift Strategies	5-58
       5.2.3.11  Torque Converter Losses and Lockup Strategy	5-58
     5.2.4   Electrification: State of Technology	5-59
       5.2.4.1   Overview of Electrification Technologies	5-62
       5.2.4.2   Non-Battery Components of Electrified Vehicles	5-64
         5.2.4.2.1   Propulsion Components	5-65

-------
                          Technology Cost, Effectiveness, and Lead-Time Assessment
     5.2.4.2.2  Power Electronics	5-66
     5.2.4.2.3  Industry Targets for Non-Battery Components	5-69
  5.2.4.3  Developments in Electrified Vehicles	5-71
     5.2.4.3.1  Non-hybrid Stop-Start	5-71
     5.2.4.3.2  Mild Hybrids	5-74
     5.2.4.3.3  Strong Hybrids	5-79
     5.2.4.3.4  Plug-in Hybrids	5-82
     5.2.4.3.5  Battery Electric Vehicles	5-92
  5.2.4.4  Developments in Electrified Vehicle Battery Technology	5-103
     5.2.4.4.1  Battery Chemistry	5-104
     5.2.4.4.2  Pack Topology, Cell Capacity and Cells per Module	5-106
     5.2.4.4.3  Usable Energy Capacity	5-110
     5.2.4.4.4  Thermal Management	5-115
     5.2.4.4.5  Pack Voltage	5-116
     5.2.4.4.6  Electrode Dimensions	5-117
     5.2.4.4.7  Pack Manufacturing Volumes	5-118
     5.2.4.4.8  Potential Impact of Lithium Demand on Battery Cost	5-121
     5.2.4.4.9  Evaluation of 2012 FRM Battery Cost Projections	5-122
  5.2.4.5  Fuel Cell Electric Vehicles	5-128
     5.2.4.5.1  Introduction to FCEVs	5-128
     5.2.4.5.2  FCEV Cost Estimation	5-130
       5.2.4.5.2.1  Fuel Cell System Cost	5-131
       5.2.4.5.2.2  Hydrogen Storage Cost	5-135
       5.2.4.5.2.3  Combined Fuel Cell and Hydrogen Storage Systems Cost	5-135
       5.2.4.5.2.4  Market Projections	5-137
     5.2.4.5.3  FCEV Performance Status and Targets	5-139
     5.2.4.5.4  Onboard Hydrogen  Storage Technology	5-141
     5.2.4.5.5  FCEV Commercialization Status	5-141
     5.2.4.5.6  Outlook for National FCEV Launch	5-142
5.2.5   Aerodynamics: State of Technology	5-143
  5.2.5.1  Background	5-143
  5.2.5.2  Aerodynamic Technologies in the FRM	5-143
  5.2.5.3  Developments since the FRM	5-144
     5.2.5.3.1  Industry Developments	5-145
     5.2.5.3.2  Joint Test Program with Transport Canada	5-147
     5.2.5.3.3  CARS Control-Tec Study	5-150
     5.2.5.3.4  EPA Study of Certification Data	5-150
     5.2.5.3.5  Conclusions	5-152
5.2.6   Tires: State of Technology	5-152
  5.2.6.1  Background	5-152
  5.2.6.2  Tire Technologies  in the FRM	5-153
  5.2.6.3  Developments since the FRM	5-154
     5.2.6.3.1  Industry Developments	5-155
     5.2.6.3.2  Control-Tec Analysis of Trends in Tire Technologies	5-157
     5.2.6.3.3  Canada Tire Testing Program	5-157
  5.2.6.4  Conclusions	5-158

-------
                          Technology Cost, Effectiveness, and Lead-Time Assessment
5.2.7  Mass Reduction: State of Technology	5-158
  5.2.7.1  Overview of Mass Reduction Technologies	5-158
  5.2.7.2  Developments since the 2012 FRM	5-162
  5.2.7.3  Market Vehicle Implementation of Mass Reduction	5-163
  5.2.7.4  Holistic Vehicle Mass Reduction and Cost Studies	5-166
    5.2.7.4.1  EPA Holistic Vehicle Mass Reduction/Cost Studies	5-169
      5.2.7.4.1.1  Phase 2 Low Development Midsize CUV Updated Study and
      Supplement 5-170
      5.2.7.4.1.2  Light Duty Pickup Truck Light-Weighting  Study	5-173
    5.2.7.4.2  NHTSA Holistic Vehicle Mass Reduction/Cost Studies	5-176
      5.2.7.4.2.1  Updated Midsize Car Lightweight Vehicle  Study	5-176
      5.2.7.4.2.2  Light Duty Pickup Truck Light-Weighting  Study	5-179
    5.2.7.4.3  ARE Holistic Vehicle Mass Reduction/Cost Study	5-184
    5.2.7.4.4  Aluminum Association Midsize CUV Aluminum BIW Study	5-185
    5.2.7.4.5  DOE/Ford/Magna MMLV Mach 1 and Mach 2 Lightweighting Research
    Projects  5-187
      5.2.7.4.5.1  Mach I	5-189
      5.2.7.4.5.2  Mach 2	5-192
    5.2.7.4.6  Technical Cost Modeling Report by DOE/INL/IBIS on 40 Percent-45
    Percent Mass Reduced Vehicle	5-194
    5.2.7.4.7  Studies to Determine Mass Add for IfflS Small Overlap	5-195
      5.2.7.4.7.1  NHTS A Mass Add Study for a Passenger Car to Achieve a "Good"
      Rating on the IfflS Small Overlap	5-196
      5.2.7'.4.7'.2  Transport Canada Mass Add Study for a Light Duty Truck to Achieve a
      "Good" Rating on the IIHS Small Overlap	5-197
5.2.8  State of Other Vehicle Technologies	5-200
  5.2.8.1  Electrified Power Steering: State of Technology	5-200
    5.2.8.1.1  Electrified Power Steering in the 2012 FRM	5-200
    5.2.8.1.2  Developments since the FRM	5-200
  5.2.8.2  Improved Accessories: State of Technology	5-200
  5.2.8.3  Secondary Axle Disconnect: State of Technology	5-201
    5.2.8.3.1  Background	5-201
    5.2.8.3.2  Secondary Axle Disconnect in the FRM	5-202
    5.2.8.3.3  Developments since the FRM	5-203
  5.2.8.4  Low-Drag Brakes:  State of Technology	5-206
    5.2.8.4.1  Background	5-206
    5.2.8.4.2  Low Drag Brakes in the FRM	5-206
    5.2.8.4.3  Developments since the FRM	5-206
5.2.9  Air Conditioning Efficiency and Leakage Credits	5-207
  5.2.9.1  A/C Efficiency Credits	5-208
    5.2.9.1.1  Background on the A/C Efficiency Credit Program	5-208
    5.2.9.1.2  Idle Test Procedure	5-208
    5.2.9.1.3  AC 17 Test Procedure	5-209
    5.2.9.1.4  Manufacturer Uptake of A/C  Efficiency Credits since the 2012 FRM.. 5-210
    5.2.9.1.5  Evaluation of the AC 17 Test Procedure	5-211
    5.2.9.1.6  Conclusions and Future Work	5-215

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
    5.2.9.2   A/C Leakage Reduction and Alternative Refrigerant Substitution	5-216
       5.2.9.2.1   Leakage	5-216
       5.2.9.2.2   Low-GWP Refrigerants	5-216
       5.2.9.2.3   Conclusions	5-218
  5.2.10  Off-cycle Technology Credits	5-218
    5.2.10.1   Off-cycle Credits Program	5-218
       5.2.10.1.1   Off-cycle Credits Program Overview	5-218
    5.2.10.2   Use of Off-cycle Technologies to Date	5-220
5.3     GHG Technology Assessment	5-223
  5.3.1   Fundamental Assumptions	5-223
    5.3.1.1   Technology Time Frame and Measurement Scale for Effectiveness and Cost... 5-
    223
    5.3.1.2   Performance Assumptions	5-224
    5.3.1.3   Fuels	5-227
    5.3.1.4   Vehicle Classification	5-228
  5.3.2   Approach for Determining Technology Costs	5-229
    5.3.2.1   Direct Manufacturing Costs	5-229
       5.3.2.1.1   Costs from Tear-down Studies	5-229
       5.3.2.1.2   Electrified Vehicle Battery Costs	5-231
       5.3.2.1.3   Specific DMC Changes since the 2012 FRM	5-232
       5.3.2.1.4   Approach to Cost Reduction through Manufacturer Learning	5-232
    5.3.2.2   Indirect Costs	5-237
       5.3.2.2.1   Methodologies for Determining Indirect Costs	5-237
       5.3.2.2.2   Indirect Cost Estimates Used in this Analysis	5-239
    5.3.2.3   Maintenance and Repair Costs	5-243
       5.3.2.3.1   Maintenance Costs	5-243
       5.3.2.3.2   Repair Costs	5-244
    5.3.2.4   Costs Updated to 2013 Dollars	5-245
  5.3.3   Approach for Determining Technology Effectiveness	5-245
    5.3.3.1   Vehicle Benchmarking	5-246
       5.3.3.1.1   Detailed Vehicle Benchmarking Process	5-246
         5.3.3.1.1.1  Engine  Testing	5-247
         5.3.3.1.1.2  Transmission Testing	5-248
       5.3.3.1.2   Development of Model Inputs from Benchmarking Data	5-251
         5.3.3.1.2.1  Engine  Data	5-251
         5.3.3.1.2.2  Engine  Map	5-251
         5.3.3.1.2.3  Inertia	5-252
         5.3.3.1.2.4  Transmission Data	5-253
         5.3.3.1.2.5  Gear Efficiency and Spin Losses	5-253
         5.3.3.1.2.6  Torque Converter	5-254
       5.3.3.1.3   Vehicle Benchmarking Summary	5-255
    5.3.3.2   ALPHA Vehicle Simulation Model	5-256
       5.3.3.2.1   General ALPHA Description	5-256
       5.3.3.2.2   Detailed ALPHA Model Description	5-257
         5.3.3.2.2.1  Ambient System	5-258
         5.3.3.2.2.2  Driver System	5-258

-------
                          Technology Cost, Effectiveness, and Lead-Time Assessment
       5.3.3.2.2.3  Powertrain System	5-259
       5.3.3.2.2.3.1  Engine Subsystem	5-259
       5.3.3.2.2.3.2  Electric Subsystem	5-260
       5.3.3.2.2.3.3  Accessories Subsystem	5-261
       5.3.3.2.2.3.4  Transmission Subsystem	5-261
       5.3.3.2.2.3.4.1   Transmission Gear Selection	5-261
       5.3.3.2.2.3.4.2   Clutch Model	5-262
       5.3.3.2.2.3.4.3   Gearbox Model	5-262
       5.3.3.2.2.3.4.4   Torque Converter Model	5-262
       5.3.3.2.2.3.4.5   Automatic Transmission & Controls	5-262
       5.3.3.2.2.3.4.6   DCT Transmission & Control	5-263
       5.3.3.2.2.3.4.7   CVT Transmission & Control	5-263
       5.3.3.2.2.3.4.8   Driveline	5-263
       5.3.3.2.2.3.5  Vehicle System	5-263
     5.3.3.2.3  Energy Auditing	5-264
     5.3.3.2.4  ALPHA Simulation Runs	5-265
     5.3.3.2.5  Post-processing	5-265
     5.3.3.2.6  Vehicle Component Vintage	5-266
     5.3.3.2.7  Additional Verification	5-267
  5.3.3.3  Determining Technology Effectiveness for MY2022-2025	5-268
  5.3.3.4  Lumped Parameter Model	5-271
     5.3.3.4.1  Lumped Parameter Model Usage in OMEGA	5-272
5.3.4   Data and Assumptions Used in GHG Assessment	5-275
  5.3.4.1  Engines: Data and Assumptions for this Assessment	5-275
     5.3.4.1.1  Low Friction Lubricants (LUB)	5-276
     5.3.4.1.2  Engine Friction Reduction (EFR1,EFR2)	5-276
     5.3.4.1.3  Cylinder Deactivation (DEAC)	5-277
     5.3.4.1.4  Intake Cam Phasing (ICP)	5-278
     5.3.4.1.5  Dual Cam Phasing (DCP)	5-279
     5.3.4.1.6  Discrete Variable Valve Lift (DVVL)	5-279
     5.3.4.1.7  Continuously Variable Valve Lift (CVVL)	5-279
     5.3.4.1.8  Investigation of Potential Future Non-HEV Atkinson Cycle Engine
     Applications	5-280
     5.3.4.1.9  GDI, Turbocharging, Downsizing	5-283
  5.3.4.2  Transmissions: Data and Assumptions  for this Assessment	5-294
     5.3.4.2.1  Assessment of Automated Transmissions (AT, AMT, DCT, CVT)	5-295
     5.3.4.2.2  Technology Applicability and Costs	5-299
  5.3.4.3  Electrification: Data and Assumptions for this Assessment	5-300
     5.3.4.3.1  Cost and Effectiveness for Non-hybrid Stop-Start	5-300
     5.3.4.3.2  Cost and Effectiveness for Mild Hybrids	5-301
     5.3.4.3.3  Cost and Effectiveness for Strong Hybrids	5-302
     5.3.4.3.4  Cost and Effectiveness for Plug-in Hybrids	5-304
     5.3.4.3.5  Cost and Effectiveness for Electric Vehicles	5-304
     5.3.4.3.6  Cost of Non-Battery Components for xEVs	5-305
     5.3.4.3.7  Cost of Batteries for xEVs	5-313
       5.3.4.3.7.1  Battery Sizing Methodology for BEVs and PHEVs	5-313

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
         5.3.4.3.7.2  Battery Sizing Methodology for HEVs	5-341
         5.3.4.3.7.3  ANL BatPaC Battery Design and Cost Model	5-341
         5.3.4.3.7.4  Assumptions and Inputs to BatPaC	5-343
         5.3.4.3.7.5  Battery Cost Projections for xEVs	5-346
         5.3.4.3.7.6  Discussion of Battery Cost Projections	5-354
         5.3.4.3.7.7  Battery Pack Costs Used in OMEGA	5-355
         5.3.4.3.7.8  Electrified Vehicle Costs Used In OMEGA (Battery + Non-battery
         Items)      5-361
    5.3.4.4  Aerodynamics: Data and Assumptions for this Assessment	5-363
    5.3.4.5  Tires: Data and Assumptions for this Assessment	5-364
    5.3.4.6  Mass Reduction: Data and Assumptions for this Assessment	5-365
       5.3.4.6.1  Cost Curves	5-365
         5.3.4.6.1.1  Cost Curve for Cars and CUVs	5-368
         5.3.4.6.1.2  Cost Curve for Light Duty Trucks	5-383
       5.3.4.6.2  Mass Reduction in the Baseline MY2014 Fleet	5-394
         5.3.4.6.2.1  Vehicles with MY2008 and MY2014 Production	5-395
         5.3.4.6.2.2  MY2014 Vehicles without MY2008 Counterparts	5-399
         5.3.4.6.2.3  MY2014 Cost Curve Adjustments Due to Vehicle Baseline MY2014-
         MY2008 Curb Weight  Differences	5-399
         5.3.4.6.2.4  Safety Regulation Mass Increase Estimate Post MY2014	5-402
       5.3.4.6.3  Effectiveness of Mass Reduction	5-403
       5.3.4.6.4  Mass Reduction Costs used in OMEGA	5-403
    5.3.4.7  Other Vehicle Technologies	5-411
       5.3.4.7.1  Electrified Power Steering: Data and Assumptions for this Assessment	5-
       411
       5.3.4.7.2  Improved Accessories: Data and Assumptions for this Assessment	5-412
       5.3.4.7.3  Secondary Axle Disconnect: Data and Assumptions for this Assessment.. 5-
       412
       5.3.4.7.4  Low Drag Brakes: Data and Assumptions for this Assessment	5-413
    5.3.4.8  Air Conditioning: Data and Assumptions for this Assessment	5-413
    5.3.4.9  Cost Tables for Individual Technologies Not Presented Above	5-413
5.4    CAFE Technology Assessment	5-415
  5.4.1   Technology Costs Used in CAFE Assessment	5-415
    5.4.1.1  Direct Costs	5-415
       5.4.1.1.1  Improved Low Friction Lubricants and Engine Friction Reduction Levels 2
       &3 (LUBEFR2 & LUBFFR3)	5-415
       5.4.1.1.2  Automatic Transmission Improvements Levels 1 & 2 (ATI1 & ATI2) 5-416
       5.4.1.1.3  High Compression Ratio Engine	5-416
       5.4.1.1.4  Advanced Diesel Engine (ADSL) Engine	5-416
       5.4.1.1.5  7-speed Manual Transmission	5-416
       5.4.1.1.6  6-speed Automatic Transmission	5-416
       5.4.1.1.7  8-speed Automatic Transmission	5-417
       5.4.1.1.8  6-speed Dual Clutch Transmission	5-417
       5.4.1.1.9  8-speed Dual Clutch Transmission	5-417
       5.4.1.1.10  Continuously Variable Transmission	5-418
       5.4.1.1.11  Belt Integrated Starter Generator	5-418

-------
                          Technology Cost, Effectiveness, and Lead-Time Assessment
     5.4.1.1.12  Crank Integrated Starter Generator	5-418
     5.4.1.1.13  Electric Power Steering	5-418
     5.4.1.1.14  Improved Accessories (IACC1 & IACC2)	5-418
     5.4.1.1.15  Low Drag Brakes	5-419
     5.4.1.1.16  Secondary Axle Disconnect	5-419
     5.4.1.1.17  Low Rolling Resistance Tires	5-419
     5.4.1.1.18  Aerodynamic Drag Reduction	5-419
     5.4.1.1.19  Mass Reduction	5-419
       5.4.1.1.19.1   Light Duty Pickup Truck Light-Weighting Study	5-424
  5.4.1.2   Indirect Costs	5-428
     5.4.1.2.1   Methodologies for Determining Indirect Costs	5-428
     5.4.1.2.2   Indirect Cost Multipliers Used in this Analysis	5-430
     5.4.1.2.3   NHTSA's Application of Learning Curves	5-434
  5.4.1.3   Technology Cost Summary Tables	5-438
     5.4.1.3.1   Basic Gasoline Engine Costs	5-439
     5.4.1.3.2   Gasoline Turbo Engine Costs	5-443
     5.4.1.3.3   Other Advanced Gasoline Engine Technologies	5-446
     5.4.1.3.4   Diesel Engine Costs	5-448
     5.4.1.3.5   Transmission Costs	5-449
     5.4.1.3.6   Electric Vehicle and Accessory Costs	5-453
     5.4.1.3.7   Vehicle Technology Costs	5-456
5.4.2  Technology Effectiveness Modeling Method and Data Used in CAFE Assessment 5-
457
  5.4.2.1   Volpe Model Background	5-458
  5.4.2.2   Autonomie Vehicle Simulation Tool	5-460
     5.4.2.2.1   Overview	5-460
     5.4.2.2.2   Plant Model Overview	5-462
       5.4.2.2.2.1  Internal Combustion Engine Model	5-462
       5.4.2.2.2.2  Transmission Models	5-464
       5.4.2.2.2.3  Electric Machine Models	5-467
       5.4.2.2.2.4  Energy Storage Models	5-467
       5.4.2.2.2.5  Chassis Models	5-468
       5.4.2.2.2.6  Tire Models	5-469
       5.4.2.2.2.7  Auxiliaries Model	5-469
       5.4.2.2.2.8  Driver Models	5-469
       5.4.2.2.2.9  Environment Models	5-469
     5.4.2.2.3   Control Overview	5-470
       5.4.2.2.3.1  Transmission Shifting Algorithm	5-470
       5.4.2.2.3.2  Torque Converter Lock-up Assumptions	5-478
       5.4.2.2.3.3  Fuel Cut-off Algorithm	5-480
       5.4.2.2.3.4  Vehicle Level Control for Electrified Powertrains	5-480
  5.4.2.3   Vehicle Model Validation	5-491
     5.4.2.3.1   Vehicle Benchmarking	5-491
     5.4.2.3.2   Vehicle Validation Examples	5-495
       5.4.2.3.2.1  Transmission Shifting Algorithm	5-495
       5.4.2.3.2.2  Powersplit HEV	5-498

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
          5.4.2.3.3  Pre-transmission HEV	5-499
            5.4.2.3.3.1  Range Extender PHEV	5-500
       5.4.2 A  Simulation Modeling Study Overview	5-501
       5.4.2.5  Selection of Technologies for Modeling	5-502
       5.4.2.6  Modeling Assumptions	5-503
          5.4.2.6.1  Vehicle Level	5-503
          5.4.2.6.2  Gasoline and Diesel Engines	5-504
       5.4.2.7  Description of Engine Technologies Evaluated	5-512
          5.4.2.7.1  Friction reduction	5-512
          5.4.2.7.2  Cylinder Deactivation	5-513
          5.4.2.7.3  Turbocharged Engines	5-513
       5.4.2.8  Transmissions	5-514
       5.4.2.9  Torque Converter	5-520
       5.4.2.10  Electric Machines	5-521
          5.4.2.10.1  Energy Storage Systems	5-523
          5.4.2.10.2  Fuel Cell Systems	5-524
       5.4.2.11  Light-weighting	5-525
       5.4.2.12  Rolling Resistance	5-525
       5.4.2.13  Aerodynamic	5-525
       5.4.2.14  Accessory Loads	5-526
       5.4.2.15  Driver	5-526
       5.4.2.16  Electrified Powertrains	5-526
          5.4.2.16.1  Electrified Powertrain Configurations	5-527
          5.4.2.16.2  Parallel Hybrid Vehicle	5-528
          5.4.2.16.3  Power Split Hybrid Vehicle	5-529
            5.4.2.16.3.1  Voltec Gen 1 Plug-in Hybrid Vehicle	5-530
          5.4.2.16.4  Series Fuel Cell HEV	5-532
          5.4.2.16.5  Powertrain Electrification Selection	5-533
       5.4.2.17  Drive Cycles and Vehicle Simulation Conditions	5-534
       5.4.2.18  Vehicle Sizing Process	5-534
       5.4.2.19  Autonomie Outputs	5-541
       5.4.2.20  Individual Vehicle  Simulation Quality Check	5-543

Table of Figures
Figure 5.1 Light-duty Vehicle Engine Technology Penetration since the 2012 Final Rule	5-13
Figure 5.2 Comparison of BTE for A Representative MY2008 2.4L 14 NA DOHC PFI 4-valve/cyl. Engine with
            Intake Cam Phasing (Left) and a GM Ecotec 2.5L NA GDI Engine with Dual Camshaft Phasing
            (Right)	5-20
Figure 5.3 Comparison of BTE for A Representative MY2010 3.5L V6 NA PFI 4-valve/cyl. Engine (Left) and a
            Toyota 2GR-FSE GDI/PFI Engine with Dual Camshaft Phasing (Right)	5-21
Figure 5.4 Graphical Representation Showing How Cylinder Deactivation Moves Engine Operation to Regions of
            Operation with Improved Fuel Consumption over the UDDS Regulatory Drive Cycle (shaded area). 5-
            22
Figure 5.5 Comparison of BTE for A Representative MY2010 5.4L V8 NA PFI 3-valve/cyl. Engine (Left) and a
            Ford 2.7L V6 Ecoboost Turbocharged, GDI Engine With Dual Camshaft Phasing (Right)	5-24
Figure 5.6 Comparison of BTE for A Representative MY2010 2.4L NA PFI Engine (Left) and A Modern, l.OL
            Turbocharged, Downsized GDI Engine (Right)	5-25
Figure 5.7 Comparison of BTE for A Representative MY2010 2.4L NA PFI Engine (Left) and A Modern, 1.5L
            Turbocharged, Downsized GDI Engine (Right)	5-26

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
Figure 5.8 Typical Turbocharger Compressor Map Showing How Pressure And Flow Characteristics Can Be
            Matched Over a Broader Range of Engine Operation Via Surge Improvement and Higher Operational
            Speed	5-27
Figure 5.9 Cross Sectional View of a Honeywell VNT Turbocharger.  The Moveable Turbine Vanes And Servo
            Linkage Are Highlighted In Light Red	5-27
Figure 5.10 A Functional Schematic Example of a Turbocharged Engine Using Two Variants of External EGR.5-29
Figure 5.11 Comparison of the Timing of Valve Events for Otto-Cycle (black and orange lines) and LIVC
            Implementations of Atkinson- Or Miller-Cycle (black and green lines)	5-30
Figure 5.12 Diagrams Of Cylinder Pressure Vs. Cylinder Volume For A Conventional Otto-Cycle SI Engine
            (orange line) Compared To A LIVC Implementation of Atkinson Cycle (green line) Highlighting the
            Reduction in Pumping Losses	5-31
Figure 5.13 Comparison of BTE for a Representative MY2010 2.4L NA PFI Engine (left) and a 2.5L NA GDI
            LIVC Atkinson Cycle Engine (right) tested by EPA	5-32
Figure 5.14 A Comparison of BSFC Maps Measured For The 2.0L 13: 1CR SKYACTIV-G Engine (left) and
            Modeled For A l.OLRicardo "EGRB Configuration" (right)	5-33
Figure 5.15 Comparison of BTE for Downsized, Turbocharged GDI Engines	5-34
Figure 5.16 Comparison of BTE for A Representative MY2010 3.5L NA PFI V6 Engine (Left) And A Downsized
            2.0L 14 Miller Cycle Engine (Right)	5-35
Figure 5.17 Comparison of BTE for 2015 Turbocharged, Downsized GDI (left) and 2017 Miller Cycle (right)
            variants of the same engine family, the 2.0L VWEA888	5-36
Figure 5.18 Comparison Of BTE For A Downsized SI 2.0L 14 Miller Cycle Engine (Left)  And A 1.7L 14
            Turbocharged Diesel Engine With HPCR, Low  And High Pressure Loop Cegr, And VNT
            Turbocharger (Right)	5-38
Figure 5.19 Exhaust Manifold Integrated Into a Single Casting with the Cylinder Head	5-40
Figure 5.20 Transmission Technology Production Share, 1980 - 2014	5-43
Figure 5.21 Average Torque Losses (Left) And Efficiency (Right) In Each Gear For An Eight-Speed 845RE
            Transmission From A Ram, Tested At 100 °C And With Line Pressures Matching Those Measured
            In-Use In The Vehicle. Torque Losses Were Averaged Over 1000 Rpm - 2500 Rpm. This
            Transmission Is A Clone of the ZF8HP45	5-45
Figure 5.22 Engine Operating Conditions for Six-Speed (Left) and Eight-Speed (Right)  Automatic Transmissions
            on the FTP-75 Drive Cycle	5-47
Figure 5.23 ZF 8HP70 Automatic Transmission	5-49
Figure 5.24 Average Number of Transmission Gears for New Vehicles (excluding CVTs)	5-50
Figure 5.25 Generic Dual Clutch Transmission	5-52
Figure 5.26 (a) Toyota CVT (b) Generic CVT sketch	5-54
Figure 5.27 ZF Torque Converter Cutaway	5-59
Figure 5.28 Hybrid System Direct Manufacturing Cost Projection (ICCT, 2015)	5-82
Figure 5.29 Battery Gross Capacity and Estimated AER or Equivalent for PHEVs	5-86
Figure 5.30 Comparison of Motor Power of 2012-2016MY Production PHEVs and FRM Estimates	5-90
Figure 5.31 Comparison of 2012-2016MYPHEV Battery Capacities to 2012 FRM Estimates	5-91
Figure 5.32 Battery Gross Capacity and EPA Estimated Range for BEVs	5-95
Figure 5.33 Comparison of Motor Power of 2012-2016MY Production BEVs and FRM  Estimates	5-99
Figure 5.34 Comparison of 2012-2016MYBEV Battery Gross Capacities to FRM Estimates	5-100
Figure 5.35 Comparison of 2012 FRM-Projected Battery Capacity to MY2012-2016 BEVs (Smaller Vehicles)5-101
Figure 5.36 Comparison of 2012 FRM-Projected Battery Capacity to MY2012-2016 BEVs (Larger Vehicles) 5-102
Figure 5.37 Comparison of 2012 FRM Projected Battery Cost Per kWh to Estimates Reviewed by Nykvist &
            Nilsson	5-123
Figure 5.38 Comparison of Estimated GM/LG Pack-Level Costs to 2012 FRM Estimates forEV150	5-125
Figure 5.39 Projection of Potential Cost Reductions for Fuel Cell System	5-131
Figure 5.40 Cost Break-Down for Catalyst in an 80kw Fuel Cell System at 1,000 And 500,000  System Annual
            Production Rates	5-132
Figure 5.41 Parameterization of SA Fuel Cell System Cost Analysis (Not Including Storage Tanks) According To
            Production Volume and System Net Power	5-134
Figure 5.42 Projection of Potential Cost Reductions for 700 Bar Compressed Hydrogen  Storage Tank System 5-135
Figure 5.43 Combined Fuel Cell and Tank System Cost Estimates across Design Space of All Possible Systems
            within Domain of SA Simplified Cost Models	5-137

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
Figure 5.44 Mean Costs for All Possible Delineated Systems With Up To Two Tanks, Between 4 and 5 kg Onboard
            Storage, lOOkW Net Power, And At Least 3,000 Units per Year	5-138
Figure 5.45 CARB Estimates Of California and Global FCEV New Vehicle Sales Estimates and Share of Total
            New Vehicle Sales	5-139
Figure 5.46 Distribution of Estimated CdA forMYs 2008 and 2014 Derived from Certification Data	5-151
Figure 5.47 Relationship between Wet Grip Index and Rolling Resistance for Winter Tires from Transport
            Canada/NRCan Study	5-156
Figure 5.48 Change in Adjusted Fuel Economy, Weight and Horsepower for MY1975-2015	5-159
Figure 5.49 Estimated Vehicle Material Change over Time 2012-2025 -Ducker Worldwide422	5-161
Figure 5.50 Forecast of Automotive Market Consumption of Composites	5-161
Figure 5.51 Magnesium Growth Expectations through 2025 (Ducker Worldwide)	5-162
Figure 5.52 Footprint (square feet) Change and Weight 2007-2014	5-164
Figure 5.53 Mass Reduction Cost Curve ($/lb.) for 2017-2025 LD GHG Joint Technical Support Document... 5-166
Figure 5.54 Original Phase 2 Low Development Midsize CUV Lightweighting Cost Curve	5-171
Figure 5.55 Revised Cost Curve for the Midsize CUV Light Weighted Vehicle	5-172
Figure 5.56 Cost Curve Figure from CAR: "A Cost Curve for Lightweighting That Is Broadly Supported"	5-173
Figure 5.57 Light Duty Pickup Truck Lightweighting Study Results	5-174
Figure 5.58 Light Duty Pickup Truck Lightweighting Cost Curve	5-175
Figure 5.59 Light Duty Pickup Truck Lightweighting Study Secondary Mass	5-176
Figure 5.60 NHTSA Passenger Car Updated Cost Curve (DMC($/kg) v %MR)442	5-177
Figure 5.61 NHTSA Revised Passenger Car DMC Curve (($/kgv %MR) and ($/vehicle v %MR))	5-178
Figure 5.62 NHTSA Draft Light Duty Pickup Truck Lightweighting (AHSS Frame with Aluminum Intensive) Cost
            Curve (DMC $/kg v %MR)	5-182
Figure 5.63 NHTSA Light Truck Cost Curve (Direct Manufacturing Costs) $/vehicle vs %MR	5-184
Figure 5.64 Phase 2 High Development BIW - Lotus Engineering	5-185
Figure 5.65 Midsize CUV Baseline vs Midsize CUV Aluminum Intensive Vehicle	5-186
Figure 5.66 Summary Table of Mass Reduction and Cost for Aluminum BIW and Closure Components	5-187
Figure 5.67 MMLV Structures Weight Comparison BIW, Closure, Chassis, Bumper	5-188
Figure 5.68 Machll Mixed Material BIW and Closure Design (brown is carbon fiber)459	5-193
Figure 5.69 Technical Cost Modeling Results for 40 Percent to 45 Percent Lightweighting Scenario (Based on
            Mach l/Mach2 Project Technologies)	5-195
Figure 5.70 Post-test Laboratory Vehicle of IfflS Small Overlap Test	5-195
Figure 5.71 MY2013 Silverado 1500 IfflS Small Overlap Test Crash Before and During	5-198
Figure 5.72 Converting the Actual CrashEventto aModel	5-198
Figure 5.73 Light Weighted Model in the IfflS Small Overlap Crash Test	5-199
Figure 5.74 Results of the Project Models from Baseline to Light Weighted on the IIHS Small Overlap462	5-199
Figure 5.75 Summary of AWD Efficiency Improvement Potentials	5-204
Figure 5.76 Contribution of Individual AWD Driveline Components to Total Additional Vehicle Mass	5-205
Figure 5.77 Variability of AC17 Round Robin Testing on 2011 Ford Explorer, A/C On	5-212
Figure 5.78 Variability of AC17 Round Robin Testing on 2011 Ford Explorer, A/C Off	5-212
Figure 5.79 Variability of AC 17 Round Robin Testing on 2011 Ford Explorer, Delta between A/C on and Off 5-213
Figure 5.80 The "Null Technology Package" and Measurement Scale for Cost and Effectiveness	5-224
Figure 5.81 Chevy Malibu Undergoing Dynamometer Testing	5-247
Figure 5.82 Engine  Test Cell  Setup	5-248
Figure 5.83 Engine Map Points	5-248
Figure 5.84 GM6T40 Transmission during Testing	5-249
Figure 5.85 Transmission Efficiency Data at 93 C and 10 Bar Line Pressure	5-249
Figure 5.86 Torque  Converter Torque Ratio and Normalized K Factor Versus Speed Ratio	5-250
Figure 5.87 Transmission Spin Losses at 93C	5-251
Figure 5.88 Chevy Malibu 2.5LBSFC Map	5-252
Figure 5.89 Engine  Spin down Inertia Test	5-253
Figure 5.90 Gear Efficiency Data At 93 C and 10 Bar Line Pressure	5-254
Figure 5.91 Torque  Converter Drive and Back-Drive Torque Ratio and Normalized K Factor versus Speed Ratio . 5-
            255
Figure 5.92 ALPHA Model Top Level View	5-258
Figure 5.93 ALPHA Conventional Vehicle Powertrain Components	5-259

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
Figure 5.94  Sample ALPHA Energy Audit Report	5-264
Figure 5.95  Example: Difference in 2016, Between Bags 1 and 3 of the FTP, from the Test Car List	5-266
Figure 5.96  Example ALPHA Model UDDS Simulation Observation Display	5-268
Figure 5.97  2.0L 14 Mazda SKYACTIV-G engine Undergoing Engine Dynamometer Testing at the EPA-NVFEL
            Facility	5-275
Figure 5.98  Comparison of a 2.0L Mazda SKYACTIV-G engine with a 13:1 geometric compression ratio to engine
            simulation results of a comparable engine with a 1-point increase in geometric compression ratio
            (14:1) and cooled, low-pressure EGR	5-281
Figure 5.99  Comparison of a 2.0L Mazda SKYACTIV-G engine with a 13:1 geometric compression ratio to engine
            simulation results of a comparable engine with a 1-point increase in geometric compression ratio
            (14:1), cooled, low-pressure EGR and cylinder deactivation with operation on 2 cylinders at below 5-
            barBMEP and 1000-3000 rpm	5-281
Figure 5.100 Mazda 2.0L SKYACTIV-G engine with 14:1 geometric compression ratio, cooled low-pressure
            external EGR system, DCO ignition system, and developmental engine management system
            undergoing engine dynamometer testing at the U.S. EPA-NVFEL facility in Ann Arbor, MI	5-282
Figure 5.101 Contour plot of B SFC in g/kW-hr versus engine speed and BMEP for the Ricardo "EBDI" engine
            equipped with sequential turbocharging, DCP, DWL, cEGR, IBM, and with a 10:1 compression ratio
            using 98 RON Indolene	5-285
Figure 5.102 BSFC Multiplier Used For Scaling Engine Maps In The Ricardo Study Based On The Ratio:
            Displacement[New]Displacement[Baseline]	5-286
Figure 5.103 Schematic Representation of the Development of BSFC Mapping for TDS24	5-287
Figure 5.104 Comparison between a 1.15L 13 version of TD S24  (left) and the 1. 5L turbocharged, GDI engine used
            in the 2017 Civic (right)	5-288
Figure 5.105 Comparison between a 1.15L 13 version of TDS24  (left) and the 2017 Golf 1.5L EA211  TSIEVO
            Engine	5-289
Figure 5.106 Comparison between a 1.51L 13 version of TDS24  (left)  and the 2017 Audi A3 2.0L 888-3B Engine
            (right)	5-289
Figure 5.107 Comparison of the Different Transmission Types	5-296
Figure 5.108 EPA PEV Battery and Motor Sizing Method	5-316
Figure 5.109 Average LDV Fuel Economy Based On Inertia Weight from MY2008 FE Trends Data	5-319
Figure 5.110 Acceleration Performance of MY2012-2016 PEVs  Compared To Targets Generated By Malliaris
            Equation	5-328
Figure 5.111 Comparison of Draft TAR Projected BEV Battery Capacities to MY2012-2016 BEVs	5-337
Figure 5.112 Comparison of Draft TAR Projected PHEV Battery Capacities to MY2012-2016 PHEVs	5-337
Figure 5.113 Projected BEV Battery Capacity per Unit Curb Weight Compared To Comparable BEVs	5-340
Figure 5.114 Comparison of Estimated Pack-Converted  GM/LG Costs to 2012 FRM EV150 and Draft TAR EV200
            Projections	5-354
Figure 5.115 2012  FRM Mass Reduction Direct Manufacturing Cost Curve ($/lb)	5-366
Figure 5.116 CAR Figure for "General Auto Manufacturer Cost Curve to Lightweight Vehicles"	5-368
Figure 5.117 2012  NHTSA Passenger Car and EPA Midsize CUV Lightweighting Study Cost Curves	5-369
Figure 5.118 EPA Updated Midsize CUV Direct Manufacturing Cost Curve from Midsize CUV Study	5-370
Figure 5.119 NHTSA Updated Passenger Car Direct Manufacturing Cost Curve from Passenger Car Study574 5-370
Figure 5.120 Car DMC  Curve from Car Data shown in CUV Methodology ($/kg vs %MR), Engineered Solution
            AHSS BIW & Aluminum Closures and Chassis Frames	5-375
Figure 5.121 DMC Curve for 2008 Era Car/CUV (2013$/kg v %MR) - HSS BIW, Al intensive	5-376
Figure 5.122 DMC Curve for 2008 Era Car/CUV (2013$/vehicle for a 3000 pound vehicle) -AHSS BIW, Al
            Intensive	5-376
Figure 5.123 GM Investment Conference Call "Vehicles with More Efficiency at Better Margins"	5-377
Figure 5.124 DMC Curve Adjusted for Car/CUV with 5 Percent Baseline Mass Reduction for MY2014 ($/kg)5-378
Figure 5.125 DMC Curve Adjusted for Car/CUV with 5 percent Baseline Mass Reduction for MY2014 ($/veh)... 5-
            379
Figure 5.126 Resultant Passenger Car Cost Curve (2013$/lb, 3000 pound vehicle shown)	5-380
Figure 5.127 Resultant Passenger Car Cost Curve (2013$/vehicle, 3000 pound vehicle shown)	5-380
Figure 5.128 Car/CUV DMC Curve Extended to Points with Aluminum BIW	5-381
Figure 5.129 U.S.  EPA Light Duty Pickup Truck Direct Manufacturing Cost Curve, MY2008 Design	5-384
Figure 5.130 NHTSA Light Duty Truck Direct Manufacturing Cost Curves, MY2014 Design	5-385

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
Figure 5.131 NHTSA Light Duty Truck (MY2014) Data Points in EPA Cost Curve Methodology ($/kg v %MR) for
            Aluminum Intensive with AHSS Frame	5-389
Figure 5.132 NHTSA Light Duty Truck (MY2014) Data Points in EPA Cost Curve Methodology ($/vehicle v
            %MR)	5-389
Figure 5.133 EPA Adjusted MY2011 LOT Cost Curve for 2014 LOT Design (-2.6%)	5-390
Figure 5.134 Combined Direct Manufacturing Cost Curve using EPA LOT and NHTSA LOT	5-391
Figure 5.135 Direct Manufacturing Cost Curve for 2008 Era Light Duty Trucks (2013$/kg vs %MR)	5-392
Figure 5.136 MY2008 Light Duty Truck DMC (2013$/Vehicle for a 6000 pound truck) vs Mass Reduction	5-392
Figure 5.137 Resultant Light duty Truck Cost Curve (2013$/lb, 6000 pound vehicle shown)	5-393
Figure 5.138 Resultant Light Duty Truck Cost Curve (2013$/vehicle, 6000 pound vehicle shown)	5-393
Figure 5.139 Car/CUV DMC ($/kg) Curve for MY2014 Vehicle with 5 Percent Lower Curb Weight Than MY2008
            (Vehicle Type 5)	5-400
Figure 5.140 Total Car/CUV DMC ($/vehicle) Curve for MY2014 Vehicle with 5 Percent Lower Curb Weight
            ThanMY2008 (Vehicle Type 5 of 1916kg)	5-401
Figure 5.141 NHTSA Passenger Car Cost Curve	5-423
Figure 5.142 Direct Manufacturing Costs for Light-Weighting Approaches Analyzed	5-424
Figure 5.143 NHTSA Draft Light Duty Pickup Truck Lightweighting (AHSS Frame with Aluminum Intensive)
            Cost Curve (DMC $/kgv%MR)	5-426
Figure 5.144 NHTSA Light Truck Cost Curve ($/Vehicle vs. % Mas Reduction)	5-428
Figure 5.145 RPE History 1972-1997 and 2007	5-429
Figure 5.146 Hypothetical Illustration of Cumulative Production Based Learning	5-435
Figure 5.147 Volpe Model Engine and Transmission Decision Trees	5-458
Figure 5.148 Model Input - Replacing Decision Trees  and Synergies with Individual Simulations	5-459
Figure 5.149 Autonomie Directly Feeds the Volpe Model	5-459
Figure 5.150 Autonomie Vehicle Model Organization	5-462
Figure 5.151 Turbo-charged Engine Response for a 1L Engine	5-463
Figure 5.152 Engine Operating Regions for Throttled Engines	5-464
Figure 5.153 Engine Operating Regions for Un-throttled Engines	5-464
Figure 5.154 Automatic Gearbox Model Input / Output	5-464
Figure 5.155 Dual Clutch Gearbox Model Input / Output	5-465
Figure 5.156 CVT Model Block Diagram	5-466
Figure 5.157 High Energy Battery Model Schematic	5-468
Figure 5.158 Shifting Controller Schematic in Autonomie	5-470
Figure 5.159 Shifting Calculations in Autonomie	5-472
Figure 5.160 Upshifting Gear Map (left), Upshifting Vehicle Speeds (right)	5-472
Figure 5.161 Example Engine Speed Range in Economical Driving, and Economical Shift	5-473
Figure 5.162 Maximum Engine Torque at Wheels and Performance Upshift Speeds	5-474
Figure 5.163 Design of Upshifting and Downshifting Speed Curves for Two  Adjacent Gears	5-475
Figure 5.164 Generic Shift Process for Automatic Transmission	5-475
Figure 5.165 Torque Hole in Autonomie during Shifting Event	5-476
Figure 5.166 Example of Engine Operating Conditions to Prevent Lugging	5-477
Figure 5.167 5-Speed Automatic Up (plain lines) and Down (dotted lines) Shifting Map	5-477
Figure 5.168 6-Speed Automatic Up (plain lines) and Down (dotted lines) Shifting Map	5-478
Figure 5.169 8-Speed Automatic Up (plain lines) and Down (dotted lines) Shifting Map	5-478
Figure 5.170 Torque Converter Efficiency Example	5-479
Figure 5.171 Torque Converter Lockup Control Algorithm	5-480
Figure 5.172 Engine Fuel Cut-off Analysis Based on Test Data (data source APRF)	5-480
Figure 5.173 Hybrid Electric Vehicle Principles [source: www.gm.com]	5-481
Figure 5.174 Engine-On Condition - 2010 Prius Example Based on 25 Test Cycles (data source APRF)	5-482
Figure 5.175 SOC Regulation Algorithm - 2010 Prius Example Based  on 25 Test Cycles (data source APRF) 5-483
Figure 5.176 Example of Engine Operating Target-2010 Prius Example Based on 25 Test Cycles (data source
            APRF)	5-483
Figure 5.177 Cycles Wheel Torque vs. Vehicle Speed, 2014 Jetta HEV Based on Test Cycles (data source APRF) 5-
            484
Figure 5.178 SOC vs. Time (left) Engine Power vs. Wheel Power (right) (data source APRF)	5-485
Figure 5.179 2013 Prius PHEV Wheel Speed and Demand Torque, Based on Test Cycles (data source APRF) 5-485

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
Figure 5.180 2013 Prius PHEV Output Power of the Battery for SOC Balancing Based on Test Cycles (data source
            APRF)	5-486
Figure 5.181 2013 Prius PHEV Battery Output Power According to SOC based on Test Cycles (data source APRF)
            	5-486
Figure 5.182 2013 Toyota Prius PHEV Engine Operating Target Based on Test Cycles (data source APRF).... 5-487
Figure 5.183 Engine On Points - 2011 GM Volt PHEV Example Based on Test Cycles (data source APRF)... 5-488
Figure 5.184 EV Operating Mode - 2011 GM Volt PHEV Based on Test Cycles (data source APRF)	5-488
Figure 5.185 HEV Operating Mode - 2011 GM Volt PHEV Based on Test Cycles (data source APRF)	5-489
Figure 5.186 Battery Output Power - 2011 GM Volt PHEV Based on Test Cycles (data source APRF)	5-489
Figure 5.187 Engine Operating Targets - 2011 GM Volt PHEV Based on Test Cycles (data source APRF)	5-490
Figure 5.188 Component Operating Conditions of a FC Von the Urban ED C using Dynamic Programming	5-491
Figure 5.189 Illustration of testing at 95°F with sun emulation (left) and at 20°F cold ambient temperature (right). 5-
            493
Figure 5.190 Data Dissemination and Project Partners	5-494
Figure 5.191 Map of Downloadable Dynamometer Database	5-494
Figure 5.192 2013 Sonata 6ATX Simulation and Testing Results on UDDS (0-505 s) (data source APRF)	5-496
Figure 5.193 Simulation and Testing Results for 6ATX (left) and 8ATX (right) (test data source APRF)	5-496
Figure 5.194 Comparison of Simulation and Test Results over the NEDC (test data source APRF)	5-497
Figure 5.195 Simulation and Test Results Compared for a Honda Civic HEV (test data source APRF)	5-498
Figure 5.196 Simulation and Testing Results over the UDDS for 2010 Toyota Prius HEV (test data source APRF) 5-
            498
Figure 5.197 Simulation and Testing Results over the UDDS for 2013 Jetta DCT HEV (test data source APRF)... 5-
            499
Figure 5.198 Simulation and Testing Results over the UDDS for 2011 GM Volt PHEV (test data source APRF)... 5-
            500
Figure 5.199 Simulation and Testing Results for a 2011 GM Volt PHEV (test data source APRF)	5-501
Figure 5.200 IAV Gasoline Engine 1  Map	5-505
Figure 5.201 IAV Gasoline Engine2 Map (left), Incremental Improvement vs Engl (right)	5-506
Figure 5.202 IAV Gasoline Engines  Map (left), Incremental Improvement vs Eng2 (right)	5-506
Figure 5.203 IAV Gasoline Engine4 Map (left), Incremental Improvement vs Eng3 (right)	5-507
Figure 5.204 IAV Gasoline EngineSb Map (left), Incremental Improvement vs Engl (right)	5-507
Figure 5.205 IAV Gasoline Engine6aMap (left), Incremental Improvement vs Eng5b (right)	5-508
Figure 5.206 IAV Gasoline Engine7aMap (left), Incremental Improvement vs Eng6a (right)	5-508
Figure 5.207 IAV Gasoline EngineSaMap (left), Incremental Improvement vs Eng7a (right)	5-508
Figure 5.208 IAV Gasoline Enginel2 Map (left), Incremental Improvement vs Engl (right)	5-509
Figure 5.209 IAV Gasoline EnginelS Map (left), Incremental Improvement vs Engl2 (right)	5-509
Figure 5.210 IAV Gasoline Enginel4 Map (left), Incremental Improvement vs Engl3 (right)	5-510
Figure 5.211 IAV Gasoline EnginelS Map (left), Incremental Improvement vs Engl4 (right)	5-510
Figure 5.212 IAV Gasoline Enginel6 Map (left), Incremental Improvement vs Engl5 (right)	5-511
Figure 5.213 Diesel IAVEnginel7 Map (left), Incremental Improvement vs Engl6 (right)	5-511
Figure 5.214. High Compression Ratio Engine Map Developed From Dynamometer Test  Data	5-512
Figure 5.215 Turbo  Charged Engine Response for a One Liter Engine	5-513
Figure 5.216 Fuel Economy and Performance  Variations with Choice  of Progression Factor for a 6-Speed
            Transmission	5-517
Figure 5.217 Gear Ratios Obtained with Three Values of Progression Factor for a 6-Speed Transmission	5-517
Figure 5.218 Comparison of Actual Gear Ratios and Gear Ratios Calculated	5-518
Figure 5.219 Comparison of Actual Gear Ratios and Gear Ratios Calculated	5-518
Figure 5.220 Torque Converter Specification Example	5-520
Figure 5.221 Electric Machine Map for Micro- and Mild HEV (data source ORNL)	5-522
Figure 5.222 Electric Machine Maps for Full HEV and split PHEVs (data source ORNL)	5-522
Figure 5.223 Electric Machine Maps for EREV PHEVs (data source ORNL)	5-522
Figure 5.224 Electric Machine Map for BEV and FCHEV (data source ORNL)	5-523
Figure 5.225 Fuel Cell System Efficiency	5-525
Figure 5.226 Electric Drive Configuration Capabilities	5-528
Figure 5.227 Power Split Hybrid Electric Vehicle	5-530
Figure 5.228 Genl Voltec Operating Modes [www.gm.com]	5-532

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
Figure 5.229 Series Fuel Cell Hybrid Electric Vehicle	5-533
Figure 5.230 Cold Start Penalty between Bag 1 and 3 on the FTP Cycle Based on 2016 EPA Certification Data.... 5-
            534
Figure 5.231 Conventional Powertrain Sizing Algorithm	5-536
Figure 5.232 Split Hybrid Electric Powertrain Sizing Algorithm	5-537
Figure 5.233 Split Plug-in Hybrid Electric Powertrain Sizing Algorithm	5-538
Figure 5.234 Series-Split Hybrid Electric Powertrain Sizing Algorithm	5-539
Figure 5.235 Battery Electric Powertrain Sizing Algorithm	5-540
Figure 5.236 Fuel Cell Series Hybrid Electric Powertrain Sizing Algorithm	5-540
Figure 5.237 Organization of Simulation Results	5-542
Figure 5.238 Example of QA/QC Distribution Plot	5-544


Table of Tables

Table 5.1  U.S. DRIVE Targets for Electric Content Cost and Specific Power	5-70
Table 5.2 Trends in EPA-Estimated Range of PHEVs	5-85
Table 5.3  Driving Range of 2012-2016MYBEVs	5-93
Table 5.4 Lithium-ion Battery Chemistries Available in ANL BatPaC	5-104
Table 5.5  Estimated SOC swings for selected MY2012-2016 BEVs	5-114
Table 5.6 Examples of Conversion Factors for Cell Costs to Pack Costs	5-124
Table 5.7 Comparison of GM/LGChem Pack-Converted Cell Costs to FRM EV150 Pack Cost	5-124
Table 5.8 Summary of Published Evidence of Battery Pack Cost and Pricing	5-127
Table 5.9 FCEV System and Production Rate Input Parameters for Assessment of Potential Costs For CARB-
            Modified SA Simplified Cost Models	5-136
Table 5.10 Updated DOE Status and Targets for Automotive Fuel Cell and Onboard Hydrogen Storage Systems.. 5-
            140
Table 5.11 Hydrogen Storage Performance and Cost Targets and Status for Various Technologies	5-141
Table 5.12 Aerodynamic Technologies Observed in Vehicles Investigated at the 2015 NAIAS	5-146
Table 5.13 Aerodynamic Technology Effectiveness from Phase 1 of Joint Aerodynamics Program	5-148
Table 5.14 Examples of Mass Reduction in Selected Recent Redesigns (Compared to MY2008 Design)	5-165
Table 5.15 Agencies Sponsored Mass Reduction Project List since FRM	5-168
Table 5.16 Components for LWV Solution	5-179
Table 5.17 Components for LWV Solution	5-183
Table 5.18 Costs Per Kilogram at Various %MR Points	5-183
Table 5.19 Direct Manufacturing Costs at MRO-MR5	5-184
Table 5.20 Summary of the Automotive Aluminum 2025	5-187
Table 5.21 Gaps Identified by MMLV Project	5-189
Table 5.22 Safety Tests Performed on the Mach-1	5-191
Table 5.23 Mach-I Components to Maintain Frontal Crash Performance	5-192
Table 5.24 Mach II Design Vehicle Summary459	5-193
Table 5.25 Estimated Mass Increase to Meet IIHS SOL for 2010 Vehicle Classes	5-196
Table 5.26 Estimated Mass Increase to Meet IIHS SOL for 2020 Vehicle Classes	5-197
Table 5.27 Hardware Bench Testing Standards under Development by SAE  Cooperative Research Program... 5-215
Table 5.28 Trends in Fleetwide Mobile Air Conditioner Leakage Credits and Average Leakage Rates	5-216
Table 5.29 Off-cycle Technologies for Cars and Light Trucks	5-220
Table 5.30 Off-cycle Technologies and Credits for Solar/Thermal Control Technologies for Cars and Light Trucks
             	5-220
Table 5.31 Percent of 2014 Model Year Vehicle Production Volume with Credits from the Menu, by Manufacturer
            & Technology (%)	5-221
Table 5.32 Off-Cycle Technology Credits from the Menu, by Manufacturer  and Technology (g/mi)	5-222
Table 5.33 Test Fuel Specifications for Gasoline without Ethanol (from 40 CFR §86.113-04)	5-227
Table 5.34 Petroleum Diesel Test Fuel (from 40 CFR §86.113-94)	5-227
Table 5.35 EPA Vehicle Classes	5-228
Table 5.36 Learning Effect Algorithms Applied to Technologies Used in this Analysis	5-235
Table 5.37 Year-by-year Learning Curve Factors for the Learning Curves Used in this Analysis	5-236

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.38 Indirect Cost Multipliers Used in this Analysis	5-239
Table 5.39 Warranty and Non-Warranty Portions of ICMs	5-240
Table 5.40 Indirect Cost Markups (ICMs) and Near Term/Long Term Cutoffs Used in EPA's Analysis	5-240
Table 5.41 Mass Reduction Markup Factors used by EPA in this Draft TAR	5-242
Table 5.42 Mass Reduction Indirect Cost Curves used by EPA for Cars Using ICMs	5-242
Table 5.43 Mass Reduction Indirect Cost Curves used by EPA for Trucks Using ICMs	5-242
Table 5.44 Mass Reduction Indirect Cost Curves used by EPA for Cars Using RPEs	5-243
Table 5.45 Mass Reduction Indirect Cost Curves used by EPA for Trucks Using RPEs	5-243
Table 5.46 Maintenance Event Costs & Intervals (2013$)	5-244
Table 5.47 Implicit Price Deflators and Conversion Factors for Conversion to 2013$	5-245
Table 5.48 Benchmark Vehicle Description	5-247
Table 5.49 Example OMEGA Vehicle Technology Packages (values are for example only)	5-272
Table 5.50 Example Baseline Vehicle (values are for example only)	5-273
Table 5.51 Example Package Application Process (values are for example only)	5-273
Table 5.52 Example LPM Calibration Check	5-274
Table 5.53 Costs for Engine Changes to Accommodate Low Friction Lubes (dollar values in 2013$)	5-276
Table 5.54 Costs for Engine Friction Reduction Level 1 (dollar values in 2013$)	5-277
Table 5.55 Costs for Engine Friction Reduction Level 2 (dollar values in 2013$)	5-277
Table 5.56 Costs for Cylinder Deactivation (dollar values in 2013$)	5-277
Table 5.57 Costs for Intake Cam Phasing (dollar values in 2013$)	5-278
Table 5.58 Costs for Dual CamPhasing (dollar values in 2013$)	5-279
Table 5.59 Costs for Discrete Variable Valve Lift  (dollar values in 2013$)	5-279
Table 5.60 Costs for Continuously Variable Valve Lift (dollar values in 2013$)	5-280
Table 5.61 Direct Manufacturing Costs (DMC) for Atkinson-2 Technology (2010$)	5-283
Table 5.62 Costs for Atkinson-2  Technology, Exclusive of Enablers such as Direct Inject and Valve Timing
             Technologies (dollar values in 2013$)	5-283
Table 5.63 Specification of Ricardo 3.2L V6 Turbocharged, GDI "EBDI" Proof-of-concept Engine	5-284
Table 5.64 FRM to Draft TAR Engine Technology Package Effectiveness Comparison	5-290
Table 5.65 Costs for Gasoline Direct Injection on an 13 &I4 Engine (dollar values in 2013$)	5-290
Table 5.66 Costs for Gasoline Direct Injection on a V6 Engine (dollar values in 2013$)	5-290
Table 5.67 Costs for Gasoline Direct Injection on a V8 Engine (dollar values in 2013$)	5-290
Table 5.68 Costs for Turbocharging, 18/21 bar, I-Configuration Engine (dollar values in 2013$)	5-291
Table 5.69 Costs for Turbocharging, 18/21 bar, V-Configuration Engine (dollar values in 2013$)	5-291
Table 5.70 Costs for Turbocharging, 24 bar, I-Configuration Engine & for Miller-cycle I-Configuration Engine
             (dollar values in 2013$)	5-291
Table 5.71 Costs for Turbocharging, 24 bar, V-Configuration Engine & for Miller-cycle V-Configuration Engine
             (dollar values in 2013$)	5-291
Table 5.72 Costs for Downsizing as part of Turbocharging & Downsizing (dollar values in 2013$)	5-292
Table 5.73 Costs for Turbocharging & Downsizing (2013$)	5-293
Table 5.74 Costs for Miller Cycle (2013$)	5-293
Table 5.75 Costs for Cooled EGR (dollar values in 2013$)	5-293
Table 5.76 Costs for Valvetrain Conversions from non-DOHC to DOHC (dollar values in 2013$)	5-294
Table 5.77 Standard Car Effectiveness	5-295
Table 5.78 Transmission Level Map	5-297
Table 5.79 TRX11 and TRX12 Null Engine Effectiveness	5-298
Table 5.80 TRX21 and TRX22 Null Engine Effectiveness	5-298
Table 5.81 Costs for Transmission Improvements for  all Vehicles (dollar values in 2013$)	5-299
Table 5.82 Comparison of Transmission Costs Using  the 2012 FRM Methodology to Draft TAR Costs for
             Transmissions (2013$)	5-299
Table 5.83 GHG Technology Effectiveness of Stop-Start	5-301
Table 5.84 Costs for Stop-Start for Different Vehicle Classes (dollar values in 2013$)	5-301
Table 5.85 GHG Technology Effectiveness of Mild Hybrids	5-302
Table 5.86 GHG Technology Effectiveness of Strong Hybrids	5-304
Table 5.87 Linear Regressions of Strong & Plug-in Hybrid Non-Battery System Direct Manufacturing Costs vs Net
             Mass Reduction Applicable in MY2012 (2013$)	5-306

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.88 Linear Regressions of Battery Electric Non-Battery System Direct Manufacturing Costs vs Net Mass
            Reduction Applicable in MY2016 (2013$)	5-306
Table 5.89 Costs for MHEV48V Non-Battery Items (dollar values in 2013$)	5-306
Table 5.90 Costs for Strong Hybrid Non-Battery Items (dollar values in 2013$)	5-307
Table 5.91 Costs for 20 Mile Plug-in Hybrid Non-Battery Items (dollar values in 2013$)	5-308
Table 5.92 Costs for 40 Mile Plug-in Hybrid Non-Battery Items (dollar values in 2013$)	5-309
Table 5.93 Costs for 75 Mile BEV Non-Battery Items (dollar values in 2013$)	5-310
Table 5.94 Costs for 100 Mile BEV Non-Battery Items (dollar values in 2013$)	5-311
Table 5.95 Costs for 200 Mile BEV Non-Battery Items (dollar values in 2013$)	5-312
Table 5.96 Costs for In-Home Charger Associated with 20 Mile Plug-in Hybrid (dollar values in 2013$)	5-312
Table 5.97 Costs for In-Home Charger Associated with 40 Mile Plug-in Hybrid (dollar values in 2013$)	5-312
Table 5.98 Costs for In-Home Charger Associated with All BEVs (dollar values in 2013$)	5-313
Table 5.99 Costs for Labor Associated with All In-Home Chargers for Plug-in & BEV (dollar values in 2013$).... 5-
            313
Table 5.100 Baseline ICE-Powertrain Weight Assumptions (Pounds), By Vehicle Class	5-317
Table 5.101 Example Net Curb Weight Reduction for BEVs and PHEVs With 20% Applied Mass Reduction
            Technology	5-318
Table 5.102 U.S. Drive Targets for Non-Battery Specific Power for 2015 and 2020	5-324
Table 5.103 Examples of Pack-Level Specific Energy Calculated By BatPac for Selected PEV Configurations (0%
            WR)	5-325
Table 5.104 Examples of Pack-Level Specific Energy Calculated By BatPac for Selected PEV Configurations (20%
            WR)	5-326
Table 5.105 Power-to-ETW Ratios Assigned to xEVs in the FRM	5-326
Table 5.106 Estimated 0-60 mph Target Acceleration Times  Corresponding to FRM Assumptions for xEV hp/lb
            ETW	5-327
Table 5.107 PEV Acceleration Performance Intended in the FRM and Projected Probable  Performance	5-329
Table 5.108 Changes in PEV Power-To-Weight Ratios and 0-60 Targets for Draft TAR	5-330
Table 5.109 Changes to Baseline Curb Weights from FRM MY2008 to Draft TARMY2014	5-331
Table5.110 PEV Battery Sizing Assumptions and Changes from FRM to Draft TAR	5-334
Table 5.111 Example Changes in Projected PEV Battery Capacity and Motor Power, FRM to Draft TAR (20%
            weight reduction case)	5-336
Table 5.112 Draft TAR Projected Battery Capacities and Assumed Curb Weights, 0% Nominal Weight ReductionS-
            338
Table 5.113 Draft TAR Projected Battery Capacities and Assumed Curb Weights, 20% Nominal Weight Reduction
             	5-338
Table 5.114 Battery Design Assumptions Input to BatPaC and Changes from 2012 FRM to 2016 Draft TAR.. 5-346
Table 5.115 Average Change in Projected Battery Pack DMC from 2012 FRM to 2016 Draft TAR	5-347
Table 5.116 Estimated Direct Manufacturing Costs inMY2025 forEV75 Battery Packs	5-348
Table 5.117 Estimated Direct Manufacturing Costs inMY2025 forEVlOO Battery Packs	5-349
Table 5.118 Estimated Direct Manufacturing Costs inMY2025 forEV200 Battery Packs	5-350
Table 5.119 Estimated Direct Manufacturing Costs in MY2025 for PHEV20 Battery Packs	5-351
Table 5.120 Estimated Direct Manufacturing Costs inMY2025 forPHEV40 Battery Packs	5-352
Table 5.121 Estimated Direct Manufacturing Costs in MY2017 for strong HEV Battery Packs	5-353
Table 5.122 Linear Regressions of Strong Hybrid Battery System Direct Manufacturing Costs vs Net Mass
            Reduction Applicable in MY2017 (2013$)	5-355
Table 5.123 Linear Regressions of Battery  Electric Battery System Direct Manufacturing  Costs vs Net Mass
            Reduction Applicable in MY2025 (2013$)	5-355
Table 5.124 Costs for MHEV48V Battery (dollar values in 2013$)	5-355
Table 5.125 Costs for Strong Hybrid Batteries (dollar values in 2013$)	5-356
Table 5.126 Costs for 20 Mile Plug-in Hybrid Batteries (dollar values in 2013$)	5-357
Table 5.127 Costs for 40 Mile Plug-in Hybrid Batteries (dollar values in 2013$)	5-358
Table 5.128 Costs for 75 Mile BEV Batteries (dollar values in 2013$)	5-358
Table 5.129 Costs for 100 Mile BEV Batteries (dollar values in 2013$)	5-359
Table 5.130 Costs for 200 Mile BEV Batteries (dollar values in 2013$)	5-360
Table 5.131 Full System Costs for  48V Mild Hybrids (2013$)	5-361
Table 5.132 Full System Costs for  Strong Hybrids (2013$)	5-361

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.133 Full System Costs for 20 Mile Plug-in Hybrids, Including Charger & Charger Labor (2013$)	5-361
Table 5.134 Full System Costs for 40 Mile Plug-in Hybrids, Including Charger & Charger Labor (2013$)	5-362
Table 5.135 Full System Costs for 75 Mile BEVs, Including Charger & Charger Labor (2013$)	5-362
Table 5.136 Full System Costs for 100 Mile BEVs, Including Charger & Charger Labor (2013$)	5-362
Table 5.137 Full System Costs for 200 Mile BEVs, Including Charger & Charger Labor (2013$)	5-363
Table 5.138 Costs for Aero Technologies (dollar values in 2013$)	5-364
Table 5.139 Costs for Lower Rolling Resistance Tires (dollar values in 2013$)	5-365
Table 5.140 Car/CUV DMC Development	5-369
Table 5.141 Methodology Differences Between Original 2012 Car and CUV Mass Reduction Studies	5-372
Table 5.142 Designation of Primary and Secondary Mass reduction for 2012 NHTSA Accord-based Passenger Car
            Study	5-373
Table 5.143 Technologies/Approaches that Result in a Mass Reduction and a DMC Savings	5-377
Table 5.144 Estimate of Percent Change in Mass Reduction Compared to 2008 Estimates	5-378
Table 5.145 Three Aluminum Intensive Vehicle Design Summary - DMC ($), %MR and $/kg	5-382
Table 5.146 LOT DMC Curve Development	5-383
Table 5.147 Comparison of MY2011 and MY2014 Crew Cab Silverado 1500582	5-383
Table 5.148 Light Duty Truck Study Cost Curve Methodology Comparison	5-386
Table 5.149 Re-Designation of Secondary Technologies Listed in NHTSA Light Duty Truck Lightweighting Report
             	5-388
Table 5.150 Draft TAR Mass Reduction Baseline Revisions	5-394
Table 5.151 Footprint Density per Vehicle Class (Ib/sqft and kg/sqft)	5-396
Table 5.152 Examples of Mass Footprint Adjustment (single vehicle)	5-397
Table 5.153 Additional Safety Mass  Added for 2014 Vehicles	5-398
Table 5.154 Examples of Safety Mass Reduction Allotted and Weight Reduction Change (single vehicle)	5-399
Table 5.155 Example of Calculations for Adjusting Car/CUV DMC Curve for 5 Percent Baseline Mass ReductionS-
            401
Table 5.156 Future Safety Regulation Reference.  Mass Increase Expectations	5-402
Table 5.157 Costs for 5 Percent Mass Reduction for Non-towing (Car curve) Vehicle Types (2013$)	5-404
Table 5.158 Costs for 10 Percent Mass Reduction for Non-towing (Car curve) Vehicle Types  (2013$)	5-405
Table 5.159 Costs for 15 Percent Mass Reduction for Non-towing (Car curve) Vehicle Types  (2013$)	5-406
Table 5.160 Costs for 20 Percent Mass Reduction for Non-towing (Car curve) Vehicle Types  (2013$)	5-407
Table 5.161 Costs for 5 Percent Mass Reduction for Towing (Truck curve) Vehicle Types (2013$)	5-408
Table 5.162 Costs for 10 Percent Mass Reduction for Towing (Truck curve) Vehicle Types (2013$)	5-409
Table 5.163 Costs for 15 Percent Mass Reduction for Towing (Truck curve) Vehicle Types (2013$)	5-410
Table 5.164 Costs for 20 Percent Mass Reduction for Towing (Truck curve) Vehicle Types (2013$)	5-411
Table 5.165 Costs for Electric Power Steering (dollar values in 2013$)	5-412
Table 5.166 Costs for Improved Accessories Level 1 (dollar values in 2013$)	5-412
Table 5.167 Costs for Improved Accessories Level 2 (dollar values in 2013$)	5-412
Table 5.168 Costs for Secondary Axle Disconnect (dollar values in 2013$)	5-413
Table 5.169 Costs for Low Drag Brakes (dollar values in 2013$)	5-413
Table 5.170 Costs for A/C Controls (dollar values in 2013$)	5-413
Table 5.171 Costs for SCR-equipped Diesel Technology for Different Vehicle Classes (dollar values in 2013 $).... 5 -
            414
Table 5.172 Costs for Advanced Diesel Technology for Different Vehicle Classes (dollar values in 2013$)	5-414
Table 5.173 Costs for Powersplit HEV Technology for Different Vehicle Classes (dollar values in 2013$)	5-415
Table 5.174 Mass Reduction and Associated Costs Going From Vehicle Version 1.0 to Vehicle Version 1.2... 5-421
Table 5.175 Mass Reduction and Costs for Vehicle Components/System	5-422
Table 5.176 Cost Per Kilogram at Distinct Mass Reduction Points MR%	5-423
Table 5.177 Components Included for Different Levels of Mass Reduction	5-427
Table 5.178 Cost Per Kilogram of Mass Reduced	5-427
Table 5.179 Direct Manufacturing Costs for Different Mass Reduction Levels	5-428
Table 5.180 Indirect Cost Multipliers Used in this Analysis	5-431
Table 5.181 Warranty and Non-Warranty Portions of ICMs	5-431
Table 5.182 ICM categories and Short Term ICM Schedules for CAFE Technologies	5-432
Table 5.183 Learning Schedules by Model Year Applied to Specific CAFE Technologies	5-436
Table 5.184 Learning Schedules for Specific CAFE Technologies	5-437

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.185 Examples of Engine Technology Costs that Scale with Engine Attributes	5-440
Table 5.186 Projected MY2017 Incremental Costs for Gasoline Engine Technology	5-441
Table 5.187 Projected MY2025 Incremental Costs for Gasoline Engine Technology	5-441
Table 5.188 Projected MY2017 Absolute Costs for Gasoline Engine Technology	5-442
Table 5.189 Projected MY2025 Absolute Costs for Gasoline Engine Technology	5-442
Table 5.190 Projected MY2017 Incremental Costs for Turbo and Turbo-Downsize Technology	5-443
Table 5.191 Projected MY2025 Incremental Costs for Turbo and Turbo-Downsize Technology	5-444
Table 5.192 Projected MY2017 and MY2025 Absolute Costs for Turbo and Turbo-Downsizing Technology .. 5-445
Table 5.193 Direct Manufacturing Costs and Learning Schedules for Advanced Engine Technologies	5-446
Table 5.194 Projected MY2017 Incremental Costs for Advanced Gasoline Engine Technologies	5-446
Table 5.195 Projected MY2025 Incremental Costs for Advanced Gasoline Engine Technologies	5-446
Table 5.196 Projected MY2017 Absolute Costs for Advanced Gasoline Engine Technologies	5-447
Table 5.197 Projected MY2025 Absolute Costs for Advanced Gasoline Engine Technologies	5-447
Table 5.198 Projected MY2017 Incremental Costs for Diesel Engines by Engine Type	5-448
Table 5.199 Projected MY2025 Incremental Costs for Diesel Engines by Engine Type	5-448
Table 5.200 Projected MY2017 Absolute Costs for Diesel Engines by Engine Type	5-449
Table 5.201 Projected MY2025 Absolute Costs for Diesel Engines by Engine Type	5-449
Table 5.202 Direct Manufacturing Costs and Learning Schedules for Transmissions	5-450
Table 5.203 Projected MY2017 Incremental Costs for Transmission Technologies by Vehicle Class	5-450
Table 5.204 Projected MY2025 Incremental Costs for Transmission Technologies by Vehicle Class	5-451
Table 5.205 Projected MY2017 Absolute Costs for Transmission Technologies by Vehicle Class	5-451
Table 5.206 Projected MY2025 Absolute Costs for Transmission Technologies by Vehicle Class	5-452
Table 5.207 Direct Manufacturing Costs and Learning Schedules for Electric Vehicle and Accessory Systems by
             Vehicle Technology Class	5-453
Table 5.208 Projected MY2017 Incremental Costs for Electric  Vehicle and Accessory Systems by Vehicle Class . 5-
             454
Table 5.209 Projected MY2025 Incremental Costs for Electric  Vehicle and Accessory Systems by Vehicle Class . 5-
             454
Table 5.210 Projected MY2017 Absolute Costs for Electric Vehicle and Accessory Systems by Vehicle Class 5-455
Table 5.211 Projected MY2025 Absolute Costs for Electric Vehicle and Accessory Systems by Vehicle Class 5-455
Table 5.212 Direct Manufacturing Costs and Learning Schedules for Vehicle Technologies	5-456
Table 5.213 Projected MY2017 Incremental Costs for Vehicle  Technologies	5-456
Table 5.214 Projected MY2025 Incremental Costs for Vehicle  Technologies	5-456
Table 5.215 Projected MY2017 Absolute Costs for Vehicle Technologies	5-457
Table 5.216 Projected MY2025 Absolute Costs for Vehicle Technologies	5-457
Table 5.217 Vehicle and Powertrain Technologies Evaluated	5-476
Table 5.218 Vehicle and Powertrain Technologies Evaluated	5-502
Table 5.219 Reference Vehicle Assumptions for all Classes in  Autonomie	5-503
Table 5.220 Gear Ratio, Final Drive Information for Sample 6-Speed Automatic Transmission Vehicles	5-515
Table 5.221 Comparison of Gear Span, Final Drive and Engine Speed for Three Transmissions	5-516
Table 5.222 Progression Ratio for Numerous Vehicles with 6-speed AU	5-518
Table 5.223 Transmission Attributes	5-519
Table 5.224 Transmission Peak Efficiency	5-520
Table 5.225 Reference Battery Characteristics	5-523

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
Chapter 5:  Technology Costs, Effectiveness, and Lead-Time Assessment

5.1    Overview

   The light-duty vehicle 2017-2025 final rule analysis was based on the agencies' assessment of
technologies as of the 2012 calendar year timeframe.  This included technologies that were
currently in production at the time, or pending near term release, as well as consideration of
further developments in technologies where reliable evidence was available.  As described in
Chapter 3, the penetration of these technologies into the fleet has proceeded steadily since then.
The focus of this chapter is on the current state of technology and the likely future developments
through MY2025, an explanation of all of the underlying new technical work that has been done
to support the agencies' analyses, and a summary of the technology assumptions and inputs used.
The agencies' modeling results are presented in Chapters 12 and 13 for the GHG and CAFE
standards, respectively.

   Throughout this initial phase of the Midterm Evaluation, EPA, NHTSA, and CARB have
evaluated the state of technologies based on many sources including new vehicle certifications,
internal full vehicle simulation modeling, technical literature reviews and technical conference
information, vehicle manufacturer and supplier meetings, and the 2015 NAS report. This
collaborative effort to collect information has produced a list of technologies for this report that
builds upon that of the GHG and CAFE 2012 final rule assessments. At the same time, the CAFE
and GHG assessments were done largely independently, due in part to differences in the
agencies' statutory authorities and through independent decisions made in each agency. The
agencies all agree that independent and parallel analyses can provide complementary results (as
shown by the differing and mutually supporting analyses in sections III and IV, respectively, of
both the MY 2012-2016  standard rulemaking preamble, and the 2017-2025 standards preamble).
It is clear that the automotive industry is innovating and bringing new technology to market at a
brisk pace and neither the GHG nor the CAFE analysis reflect all of the latest and emerging
technology since the FRM.

   While the cost, effectiveness,A and implementation feasibility of individual technologies are
generally consistent with the compliance pathways projected in the FRM, some developments
were not foreseen by the agencies. Several new technologies or unforeseen application of
technologies are now under active development and some have emerged into the light-duty
vehicle market since the  LD 2017-2025 Final Rule was completed. These technologies include
the application of direct injection Atkinson Cycle engines in non-hybrids, greater penetration of
continuously variable transmissions (CVT) and greater market penetration of diesel engines. In
addition, the development of several technologies has proceeded differently than was assumed in
the FRM, including development of downsized turbo-charged engines, cylinder deactivation and
vehicle electrification.

   In general, the agencies have initially found the estimates of technology effectiveness used in
the FRM to have been robust and accurate. Through analysis of current vehicle certification,
benchmarking, literature reviews and modeling, the agencies have, in many cases,  confirmed in
A The term 'effectiveness' is used throughout this Chapter to refer both to a reduction in tailpipe CCh emissions and a
  reduction in fuel consumption. In cases where the two are not equivalent (e.g., when changing fuel type), separate
  values are presented.


                                              5-1

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
this initial analysis that the values used in the FRM are an appropriate estimate of technology
effectiveness. This is not to imply that every manufacturer that has added technology has
achieved the effectiveness estimated in the FRM. Some manufacturers have chosen to adopt
technology and use it to improve other vehicle attributes, other than solely improving vehicle
efficiency. These other attributes include 0 to 60 mph acceleration, increased cargo capacity,
increased towing capability, and/or increased vehicle size and mass. Some applications of
technology are in their first or second design iteration and we expect that each successive
iteration will improve its effectiveness. One example of this is the emerging use of integrated and
cooled exhaust manifolds and the resulting improved effectiveness from turbo-charged
downsized engines. Vehicle manufacturers have adopted many examples of technologies that
perform very well, such as the Mazda SKYACTIV-G® engine and the ZF 8-speed transmission,
and when these technologies are combined with the sole intent of improving vehicle efficiency,
our analysis shows that significant improvements from the baseline fleets are broadly achievable
using conventional powertrains.

   The agencies continue to assess technology as it becomes available and as it develops in the
market and will revisit all of the technology effectiveness estimates for later steps of the midterm
evaluation process, including the EPA's Proposed Determination and NHTSA's Notice of
Proposed Rulemaking (NPRM). For several technologies, such as CVT's and Miller cycle
engines, some ongoing projects were not completed at the time of publishing this Draft TAR;
detailed benchmarking and simulation work will continue to be performed, and will be
considered by the agencies as it becomes available. Further, there are longer-term research
efforts underway that may be valuable in informing future technology developments, even
beyond the timeframe of the 2025 standards. One such research program is the Department of
Energy's Co-Optimization of Fuels and Engines initiative, which is working to accelerate the
introduction of high-efficiency, low emissions engines and sustainable biofuels.B In addition to
these and other examples of ongoing research on advancing technologies, the agencies will be
considering new vehicle  certifications, new work with regard to technology that is done in the
public domain, and information that is shared by stakeholders in later steps of the midterm
evaluation process and CAFE rulemaking. The agencies are therefore requesting public
comments on vehicle technologies, including data on costs and effectiveness of technologies
discussed here or additional information on technologies which could be in production in the
2022-2025 timeframe or are already in production today that may have been omitted from this
Draft TAR.

   This Chapter is organized to provide a complete description of the cost, effectiveness, and
application of the technologies considered by the agencies in this technical assessment. We have
included a brief review of the technology assessment used in the FRM as well as a summary of
all the research that has been performed since the FRM to inform the Draft TAR. Finally, we
discuss how we synthesized all of the various inputs to inform the final cost, effectiveness, and
application conclusions.

   Section 5.2 presents the agencies'joint assessment of the current state of technologies and the
advancements that have occurred  since the FRM. The agencies have reexamined every
technology considered in the FRM, as well as assessing some technologies that are currently
B For more information see http://energy.gov/eere^ioenergy/co-optimization-fuels-engines.
                                              5-2

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
commercially available but did not play a significant role in the FRM analysis, as well as
emerging technology for which enough information is known that it may be included in this
Draft TAR. The categories of technologies discussed in Section 5.2 include: engines,
transmissions, electrification, aerodynamics, tires, mass reduction, and other vehicle
technologies, such as improved accessories and low drag brakes. In addition, Section 5.2.9
provides an overview of the air conditioning efficiency and leakage credits, updates on test
evaluations for the Idle Test and the AC 17 air conditioning performance test, and a summary of
the situation regarding low global warming potential (GWP) refrigerant. Section 0 concludes
with a summary of the off-cycle credit program and an overview of how off-cycle credits have
been used by manufacturers in their current compliance with the GHG program. This section
also details how off-cycle credits have been considered in the Draft TAR analysis.

   The final two sections of this  chapter are devoted to presenting the details of the approaches,
assumptions, and technology inputs used in the agencies' independent assessments; beginning in
Section 5.3 with the technology assessment that forms the basis of the analysis of the GHG
standards, followed by the technology assessment for the CAFE program in Section 5.4.

   The particular details of the technology assessment for the GHG analysis begin in Section
5.3.1 with a description of the fundamental assumptions for fuels, performance neutrality, and
cost and effectiveness measurement that underpin the technical analysis.

   Section 5.3.2 focuses on the overall costing methodologies used in the GHG analysis which
include the determination of both direct and indirect costs, as well as the application of learning
and maintenance and repair  costs. The methodologies used to develop technology costs remain
largely unchanged from the  FRM. However, all of the technology cost inputs have been
reevaluated based on any new information available since the FRM.  In some cases, the costs
used in the FRM were determined to remain the most appropriate; in other cases, cost values
have been updated, including transmissions due to updates to the teardown results used in the
FRM, and battery costs due  to updates to the model upon which the FRM's battery costs were
based. Further, we have updated the costs for 24-bar turbocharged packages to include additional
costs associated with variable geometry turbochargers, as well as updating mass reduction costs
based on teardown studies completed since the FRM.  Importantly, we have also added new
technologies that were not considered in the FRM, notably  a direct injection Atkinson Cycle
engine and a 48 Volt mild-hybrid.

   Section 5.3.3 describes the approach used for determining technology effectiveness in the
GHG analysis. Vehicle benchmarking is at the foundation of the EPA's analysis for technology
effectiveness and a description of the benchmarking testing conducted by the EPA can be found
in Section 5.3.3.1. The benchmarking data have been used largely to inform EPA's full  vehicle
simulation model, ALPHA,  and information regarding vehicle modeling is provided in  Section
5.3.3.2. EPA has also estimated the effects of adding technology to existing powertrains using
Gamma Technology's GT Power model and the results of this investigation can be found in
Section 5.4.1. Finally, EPA  continues to apply the Lumped Parameter Model (LPM) to
efficiently estimate the overall effectiveness of technology packages, and the updates to the LPM
and its application in the Draft TAR is described in Section 5.3.3.4.

   In Section 5.3.4, EPA describes the specific data and assumptions for individual technologies
that are used in the GHG analysis in this Draft TAR. Informed by all of the information on the
state of technologies described in Section 5.2, these inputs and assumptions for cost,
                                              5-3

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
effectiveness, and technology application are used in the OMEGA model determination of the
cost-minimizing compliance pathway presented in Chapter 12.

   Section 5.4 presents the approaches, methodologies, and inputs used in the technology
assessment for the CAFE analysis.

   Section 5.4.1 describes the methodologies for estimating technology costs in the CAFE
analysis, and particular cost assumptions for individual technologies.

   Section 5.4.2 provides detail onNHTSA's evaluation of technology effectiveness based on
vehicle benchmarking, engine simulation using the GT Power model and full vehicle simulation
modeling using Argonne National Laboratory's Autonomie model.

     Some of the technologies considered for this Draft TAR for which there are notable updates
from the FRM analysis are summarized below. The full discussion of these updates is provided
throughout the remaining sections of this chapter.

       •  Direct Injection Atkinson Cycle Engine
          0   In the FRM, the use of Atkinson Cycle engines was primarily considered in HEV
              applications. In the last few years,  a new generation of naturally-aspirated SI
              Atkinson Cycle engines applicable outside of HEVs have been introduced into
              light-duty vehicle applications. The most prominent application of this technology
              is the Mazda SKYACTIV-G® system. It combines direct injection, an ability to
              operate over an Atkinson Cycle with increased expansion ratio, wide-authority
              intake camshaft timing, and an optimized combustion process. This type of
              engine operation is not limited to naturally aspirated engines and when applied to
              boosted engines is referred to as "Miller Cycle," as described below.

       •  Turbocharged, Downsized Engines
          0   In the FRM, turbocharged, downsized engines were anticipated to be a prominent
              technology applied by vehicle manufacturers to improve vehicle powertrain
              efficiency.
          0   The penetration rate of turbo-downsized engines into the light-duty fleet has
              increased from 3 percent in 2008 to 16 percent in 2014.l
          °   Turbocharged, downsized engines are beginning to adopt head-integrated exhaust
              manifolds or separate, water-cooled exhaust manifolds.  These systems also use
              separate coolant loops for the head/manifold and for the engine block. The
              changes allow faster warmup, improved temperature control of critical engine
              components, further engine downspeeding, and reduce the necessity for
              commanded enrichment for component protection. The  net result is improved
              efficiency over the regulatory cycles and during real world driving. Engine
              downspeeding also has synergies with recently developed, high-gear-ratio spread
              transmissions that may result in further drive cycle efficiency improvements.

       •  Direct Injection Miller Cycle Engine
          0   This new generation of turbocharged GDI engine combines direct injection, the
              ability to operate over a Miller Cycle (boosted Atkinson Cycle) with increased
                                              5-4

-------
                        Technology Cost, Effectiveness, and Lead-Time Assessment
       expansion ratio, wide-authority intake camshaft timing, and an optimized
       combustion process.

•  Turbocharger Improvements
   0   Newer turbochargers have been developed that reduce both turbine and
       compressor inertia allowing faster turbocharger spool-up.
   0   Improvements have been made to broaden the range of compressor operation
       before encountering surge and to improve compressor efficiency at high pressure
       ratios.
   0   The introduction of head-integrated exhaust manifolds or separate, water-cooled
       exhaust manifolds reduces exhaust turbine inlet temperatures under high-load
       conditions and improves exhaust temperature control. This allows the use of less
       expensive, lower temperature materials for the turbine housing and exhaust
       turbine. Reduced turbine inlet temperatures also allow the introduction of
       turbochargers with variable nozzle turbines into SI engine applications, similar to
       those used in light-duty diesel applications.
   0   Twin-scroll turbochargers are finding broad application in turbocharged,
       downsized GDI engines. Twin-scroll turbochargers improve turbocharger spool-
       up and improve torque output at lower engine speeds, allowing further engine
       downspeeding.
   0   Turbochargers with variable nozzle turbines (VNT) are now common in light-
       duty diesel applications and are under development for gasoline spark ignition
       engines, particularly those that use cooled EGR and head-integrated exhaust
       manifolds.

•  Cylinder Deactivation
   0   Cylinder deactivation applied to engines with less than six cylinders was not
       analyzed as part of the FRM. Further developments in NVH (noise, vibration, and
       harshness) abatement, including the use of dual-mass dampening systems, has
       resulted  in the recent introduction of a 4-cylinder/2-cylinder engine into the
       European light-duty vehicle market.
   0   The development of rolling or dynamic cylinder deactivation systems allows a
       further degree of cylinder deactivation for odd-cylinder (e.g., 3-cylinder, 5-
       cylinder) inline engines than was possible with previous cylinder deactivation
       system designs.
   0   Both 3-cylinder/2-cylinder and 3-cylinder/1.5-cylinder (rolling deactivation)
       designs are at advanced stages of engine development

•  Variable Geometry Valvetrain Systems
   0   In the FRM, variable geometry valvetrain systems, including those that vary valve
       timing and/or valve lift, were anticipated in the FRM to be a major technology for
       reducing engine pumping losses.

•  Continuously Variable Transmissions (CVT)
                                       5-5

-------
                    Technology Cost, Effectiveness, and Lead-Time Assessment
0  A new generation of C VTs has been introduced into the LD market by several
   major OEMs. These new CVTs have significant improvements in the areas of
   efficiency, integration, and customer acceptance over the previous generation.
0  Early CVTs had various customer acceptance issues mainly due to lack of
   positive shift feel typical in a conventional automatic transmission. Recent
   changes to transmission control strategies include an index shift, providing the
   consumer with an experience that more closely resembles a conventional
   automatic transmission. These changes in shift strategies may or may not result in
   a small decrease in overall powertrain efficiency; however, the bulk of the
   customer acceptance issues have been addressed and CVTs have become very
   popular.

Dual Clutch Transmissions (DCT)
0  Initial implementation of DCTs, mostly in non-performance vehicles, were
   accepted in Europe but were not widely accepted in the North American market.
   Launch and shift characteristics differed from conventional automatic
   transmission performance affecting some consumer acceptance in the United
   States. However, strategies have been developed to improve overall DCT
   operational characteristics.
0  Damp Dual  Clutch Transmission (DCT)
       The Damp Clutch DCT combines the improved durability and drivability of
       the Wet Clutch DCT with the  efficiency of a Dry Clutch DCT.
0  Torque Converter Dual Clutch Transmission
       The addition of a torque converter as a launch device greatly improves
       operational characteristics and eliminates the need for complex crankshaft
       dampers and other NVH technologies. The elimination of these NVH
       technologies approximately offsets the additional cost of the torque converter.
0  HEV or Mild Hybrid
       Integrating a DCT into either HEV or low-voltage, 48V P2 drive systems
       provides improved launch assist, low-speed creep capability, and torque
       between shifts comparable to the driving characteristics of a torque-
       converter/planetary  gear-set automatic transmission.

Vehicle Electrification
0  The sales of hybrid products have been negatively impacted by lower fuel prices
   and improvements in the efficiency of conventional vehicles that are,  in many
   cases, closing the fuel economy gap between hybrid and conventional vehicles.
0  While stop-start has been in production for a considerable amount of time in
   Europe (a predominantly manual transmission market), some of the initial product
   offerings had consumer feedback  concerns. Recent vehicles introduced with stop-
   start that were specifically designed for the U.S. market, such as the Chevrolet
   Malibu, have been met with very good reviews. Indications from suppliers  are
   that further improvements, including the use of continuously engaged starters, are
   under development.
0  Low Voltage Mild Hybrid
                                   5-6

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                 A new generation of Mild Hybrid technologies has been introduced into the
                 LD market using a nominal 48 volt electrical system that features the
                 elimination of costly high voltage safety requirements and leverages the use of
                 lower cost battery technologies. An effectiveness close to that of higher-
                 voltage mild hybrids can be achieved by significantly reducing battery pack
                 weight, and by eliminating active battery pack cooling hardware and heavy 3-
                 phase AC cables.

5.2    State of Technology and Advancements Since the 2012 Final Rule

   Since the 2017-2025MY GHG standards were established in 2012, efficiency technologies
have been developed further and steadily implemented by manufacturers over a broad range of
vehicles. Many of these are key technologies that factored  prominently in the FRM analysis,
such as direct injection, turbocharging and downsizing, and higher gear count transmissions.
The goal of improving cost-effectiveness is a consistent driver of innovation, and the resulting
advancement that is occurring for even previously established technologies necessitates a re-
evaluation of cost, effectiveness, and implementation for this analysis. For example,  the light-
weight materials, aerodynamic features,  and dual-clutch transmissions applied initially to high-
performance and luxury vehicles are requiring more cost-effective implementations and different
consumer considerations for their successful adoption in mass-market vehicles.

   Other technologies that were known,  but not included previously, have continued to evolve
and are now being applied in ways that were not expected or considered at the time of the FRM
analysis. Direct injection Atkinson Cycle engines have been applied to non-hybrids successfully,
and continuously variable transmissions are contributing to high powertrain efficiencies in
applications that have been well-received by consumers and expert reviewers.

   Still other technologies have emerged since the FRM analysis which were previously thought
to be beyond the 2017-2025MY timeframe, but now appear promising or even likely  due to
further innovation and development. Mild hybrid electric vehicles with 48 volt electrical
systems are one example that have undergone substantial testing and development by multiple
suppliers, and have demonstrated significant efficiency benefits with lower complexity and
system cost compared to strong hybrid systems or higher voltage mild hybrid systems.

5.2.1   Individual Technologies and Key Developments

   The technologies considered for this Draft TAR are briefly described below. They fit
generally into four broad categories: engine, transmission,  vehicle, and electrification
technologies. A more detailed description of each technology, and the technology's costs and
effectiveness, is described in greater detail later in this section. These technologies were also
considered in the FRM unless otherwise noted.

   Types of engine technologies applied in this Draft TAR  analysis to improve fuel economy and
reduce CCh emissions include the following:

       •  Low-friction lubricants - low viscosity and advanced low friction lubricants oils are
          now available with improved performance and better lubrication.
       •  Reduction of engine friction losses - can be achieved through low-tension piston
          rings, roller cam followers, improved material coatings, more optimal thermal
          management, piston surface treatments, cylinder wall treatments and other
                                              5-7

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
          improvements in the design of engine components and subsystems that improve
          engine operation.
          Second level of low-friction lubricants and engine friction reduction - As
          technologies advance between now and the rulemaking timeframe, there will be
          further developments enabling lower viscosity and lower friction lubricants and more
          engine friction reduction technologies available, including the use of roller bearings
          for balance shaft systems and further improvements to surface treatment coatings.
          Cylinder deactivation - deactivates the intake and exhaust valves and prevents fuel
          injection into some cylinders during light-load operation.  The engine runs
          temporarily as though it were a smaller displacement engine with fewer cylinders
          which substantially reduces pumping losses.
          Variable valve timing - alters the timing or phase of the intake valve, exhaust valve,
          or both, primarily to reduce pumping losses, increase specific power, and control
          residual gases.
          Discrete variable valve lift - increases efficiency by optimizing air flow over a
          broader range of engine operation which reduces pumping losses. Accomplished by
          controlled switching between two or more cam profiles.
          Continuous variable valve lift - an electromechanically controlled system  in which
          cam period and phasing is changed as lift height is controlled.  This yields a wide
          range of performance optimization and volumetric efficiency, including enabling the
          engine to be valve throttled.
          Stoichiometric gasoline direct-injection technology - injects fuel at high pressure
          directly into the combustion chamber to improve cooling of the air/fuel charge within
          the cylinder, which allows for higher compression ratios and increased
          thermodynamic efficiency.
          Turbocharging and downsizing - increases the available airflow and specific power
          level, allowing a reduced engine size while maintaining performance.  This reduces
          pumping losses at lighter loads in comparison to a larger engine. In this Draft TAR,
          the agencies considered two levels of boosting, 18 bar brake mean effective pressure
          (BMEP) and 24 bar, as well as four levels of downsizing, from 14 to smaller 14 or 13,
          from V6 to 14 and from V8 to V6 and 14. 18 bar BMEP is applied with 33  percent
          downsizing and 24 bar BMEP is applied with 50 percent.  To achieve the same level
          of torque when downsizing the displacement of an engine by 50 percent,
          approximately double the manifold absolute pressure (2 bar) is required. Engine
          downsizing to 27 bar BMEP used in the 2017-2025 FRM was not considered in this
          Draft TAR.
          Atkinson Cycle Engines - combine a substantial increase in geometric compression
          ratio0 (in the range of 12.5 - 14:1) and alters intake valve event  timing to provide
c Geometric compression ratio is a ratio of the piston clearance volume + displacement swept volume to the
  displacement swept volume in a reciprocating piston engine. The actual effective compression ratio and
  expansion ratio must also take into account valve events governing the actual flows involved in the combustion
  process. Effective compression ratio and expansion ratios for typical Otto-cycle engines are nearly equivalent
  and governed by the chosen geometric compression ratio. Atkinson and Miller Cycle engines lower the trapped
  air or air-fuel charge volume during intake via either late intake valve closing or early intake valve closing to
  reduce effective compression ratio while simultaneously increasing effective expansion ratio. This is done by
  reducing the piston clearance volume and thus increasing the geometric compression ratio.
                                               5-8

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
       much later intake valve closing (LIVC).  This lowers the trapped air charge,
       effectively lowering actual compression ratio to reduce knock limited operation while
       maintaining the expansion ratio for improved efficiency.  Although producing lower
       torque at low engine speeds for a given displacement, this engine has  specific high
       efficiency operating points and is capable of significant CCh reductions when
       properly matched to a strong hybrid system. Electric motor/generators produce high
       torque at low speeds are thus are capable offsetting low engine speed torque
       deficiencies with Atkinson Cycle engines.
    •  Direct Injection Atkinson Cycle Engines - combine direct injection, a substantial
       increase in geometric compression ratio (in the range of 13 - 14:1), wide authority
       intake camshaft timing, variable exhaust camshaft timing and  an optimized
       combustion process enabling significant reductions in CCh as  compared to a standard
       direct injected engine. This engine is capable of changing the effective compression
       ratio (i.e., varying the degree of Atkinson operation) by varying intake valve events.
       The ability to reduce pumping losses over a large area of operation may allow
       avoidance of the additional cost of higher gear count transmissions. The Mazda
       SKYACTIV-G engine is one example of this technology. This technology was not
       considered in the FRM.
    •  Miller Cycle Engines - combine direct injection,  a substantial increase in geometric
       compression ratio relative to other boosted engines, wide authority intake camshaft
       timing, and variable exhaust camshaft timing, and an optimized combustion process
       enabling significant reductions in CCh as compared to a standard direct injected
       engine. This is essentially Atkinson Cycle with the addition of a turbocharger
       boosting system.  The addition of a turbocharger improves volumetric efficiency and
       broadens the areas of high-efficiency operation.  The ability to reduce pumping losses
       over a large area of operation may allow avoidance of the additional cost of higher
       gear count transmissions. This technology was not considered in the FRM.
    •  Exhaust-gas recirculation with boost - increases the exhaust-gas recirculation used in
       the combustion process to increase thermal efficiency and reduce pumping losses.
       Peak levels of exhaust gas recirculation approach 25 percent by volume in these
       highly boosted engines (this, in turn raises the boost requirement by approximately 25
       percent). This technology is only applied to 24 bar BMEP and Miller cycle engines
       in this Draft TAR. The 27 bar BMEP engine used in the FRM was not considered for
       this Draft TAR.
    •  Diesel Engines - have several characteristics that give superior fuel efficiency,
       including reduced pumping losses due to lack of (or greatly reduced) throttling, and a
       combustion cycle that operates at higher compression ratio and expansion ratios, with
       a very lean air/fuel mixture, than an equivalent-performance gasoline  engine. This
       technology requires additional enablers, such as use of NOx adsorption exhaust
       catalyst (NAC), selective catalytic reduction (SCR) of NOx, or a combination of both
       NAC and SCR NOx catalytic after-treatment.
Transmission technologies considered in this Draft TAR include:

    •  Improved automatic transmission controls - optimizes  shift schedule to maximize
       fuel efficiency under wide ranging conditions,  and minimizes  losses associated with
       torque converter slip through lock-up or modulation.
                                           5-9

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
       Six, seven, and eight-speed automatic transmissions - the gear ratio spacing and
       transmission ratio are optimized to enable the engine to operate in a more efficient
       operating range over a broader range of vehicle operating conditions.
       Dual clutch transmission (DCT) - are similar to a manual transmission, but the
       vehicle controls shifting and launch functions. A dual-clutch automated shift manual
       transmission uses separate clutches for even-numbered and odd-numbered gears, so
       the next expected gear is pre-selected, which allows for faster, smoother shifting.
       Continuously Variable Transmission (CVT) - uses a belt between two variable ratio
       pulleys allowing an infinite set of gear ratios to enable the engine to operate in a more
       efficient operating range over a broad range of vehicle operating conditions.
       Shift Optimization - targets engine operation at the most efficient point for a given
       power demand. The shift controller emulates a traditional Continuously Variable
       Transmission by selecting the best gear ratio for fuel economy at a given required
       vehicle power level to take full advantage of high BMEP engines.  The shift
       controller also incorporates boundary conditions to prevent undesirable operation
       such as shift busyness and NVH issues.
       Manual 6-speed transmission - offers an additional gear ratio, often with a higher
       overdrive gear ratio, than a 5-speed manual transmission.
       High Efficiency Gearbox (automatic, DCT, CVT, CVT, or manual) - continuous
       improvement in seals, bearings and clutches, super finishing of gearbox parts, and
       development in the area of lubrication, all aimed at reducing frictional and other
       parasitic load in the system for an automatic, DCT or manual type transmission.
Types of vehicle technologies applied in this Draft TAR analysis include:

    •  Low-rolling-resistance tires - have characteristics that reduce frictional losses
       associated with the energy dissipated in the deformation of the tires under load,
       thereby reducing the energy needed to move the vehicle.  There are two levels of
       rolling resistance reduction considered in this Draft TAR analysis targeting at 10
       percent and 20 percent rolling resistance reduction respectively.
    •  Low-drag and zero drag brakes - reduce the sliding friction of disc brake pads on
       rotors when the brakes are not engaged because the brake pads are pulled away from
       the rotors.
    •  Secondary axle disconnect for four-wheel drive systems - provides a torque
       distribution disconnect between front and rear axles when torque is not required for
       the non-driving axle.  This results in the reduction of associated parasitic energy
       losses.
    •  Aerodynamic drag reduction - is achieved by  changing vehicle shapes, reducing
       frontal area, sealing gaps in body panels, or adding additional components including
       side trim, air dams, underbody covers, and more aerodynamic side view mirrors.
       There are two levels of aerodynamic drag reduction considered in this Draft TAR
       analysis targeting 10 percent and 20 percent aerodynamic drag reduction respectively.
    •  Mass reduction - encompasses a variety of techniques ranging from improved design
       and better component integration to application of lighter and higher-strength
       materials. Mass reduction can lead to collateral fuel economy and GHG benefits due
                                           5-10

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
          to downsized engines and/or ancillary systems (transmission, steering, brakes,
          suspension, etc.).
   Types of electrification/accessory and hybrid technologies considered in this Draft TAR
include:

       •  Electric power steering (EPS) - An electrically-assisted steering system that has
          advantages over traditional hydraulic power steering because it replaces a
          continuously operated hydraulic pump, thereby reducing parasitic losses from the
          accessory drive.
       •  Improved accessories (IACC) - There are two levels of IACC applied in this Draft
          TAR analysis. The first level may include high efficiency alternators, electrically
          driven (i.e., on-demand) water pumps and cooling systems.  This excludes other
          electrical accessories such as electric oil pumps and electrically driven air conditioner
          compressors. The second level of IACC includes alternator regenerative braking on
          top of what are included in the first level of IACC.
       •  Air Conditioner Systems - These technologies include improved hoses, connectors
          and seals for leakage control.  They also include improved compressors, expansion
          valves, heat exchangers and the control of these components for the purposes of
          improving tailpipe CCh emissions and fuel  economy when the A/C is operating.
       •  Non-hybrid 12-volt Stop-Start - Also known as idle-stop or 12V micro hybrid and is
          the most basic system that facilitates idle-stop capability.  This system typically
          includes an enhanced performance starter and battery.
       •  Mild Hybrid- Provides idle-stop capability and launch assistance and uses a higher
          voltage battery with increased energy capacity over typical automotive batteries.  The
          higher system voltage allows the use of a smaller, more powerful electric motor and
          reduces the weight of the motor, inverter, and battery wiring harnesses. This system
          replaces a standard alternator with an enhanced power, higher voltage, higher
          efficiency belt-driven starter-alternator which can recover braking  energy while the
          vehicle slows down (regenerative braking).  An example of a 100 volt system is the
          GM Chevrolet Malibu eAssist system.  Next generation mild hybrid systems
          scheduled for production starting in 2017 include versions running at 48 volts that
          significantly reduce cost by using lower cost batteries, lower cost electrical
          components, and eliminating high voltage safety systems.
       •  P2 Hybrid- P2 hybrid is a hybrid technology that uses a transmission integrated
          electric motor placed between the engine and a gearbox or CVT, with a wet or dry
          separation clutch which is used to decouple the motor/transmission from the engine.
          In addition, a P2 Hybrid would typically be equipped with a larger electric machine
          than a mild hybrid system but smaller than  a power-split hybrid architecture.
          Disengaging the clutch allows all-electric operation and more efficient brake-energy
          recovery.  Engaging the clutch allows efficient coupling of the engine and electric
          motor and based on simulation, when combined with a DCT transmission, provides
          similar efficiency to other strong hybrid systems.
       •  Power-split Hybrid (PSHEV) -A hybrid electric drive system that replaces the
          traditional transmission with a single planetary gear-set and two motor/generators.
                                              5-11

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
          The smaller motor/generator uses the engine to either charge the battery or supply
          additional power to the drive motor. The second, more powerful motor/generator is
          permanently connected to the vehicle's final drive and always turns with the wheels,
          as well as providing regenerative braking capability.  The planetary gear-set splits
          engine power between the first motor/generator and the output shaft to either charge
          the battery or supply power to the wheels. The Power-split hybrid provides similar
          efficiency to other strong hybrid systems.
          Plug-in hybrid electric vehicles (PHEV) - Are hybrid electric vehicles with the means
          to charge their battery packs from an outside source of electricity (usually the electric
          grid). These vehicles have larger battery packs than non-plug-in hybrid electric
          vehicles with more energy storage and a greater capability to be discharged.  They
          also use a control system that allows the battery pack to be substantially depleted
          under electric-only or blended mechanical/electric operation, allowing for reduced
          fuel use during "charge depleting" operation.
          Battery electric vehicles (BEV) - Are vehicles with all-electric drive and with vehicle
          systems powered by energy-optimized batteries charged from an outside source of
          electricity (usually the electric grid). BEVs with 75 mile, 100 mile and 200 mile
          ranges have been included as potential technologies.
5.2.2   Engines: State of Technology

   Internal combustion engine improvements continue to be a major focus in improving the
overall efficiency of light-duty vehicles.  While the primary type of light-duty vehicle engine in
the United States is a gasoline fueled, spark ignition (SI), port-fuel-injection (PFI) design, it is
undergoing a significant evolution as manufacturers work to improve engine brake thermal
efficiency (BTE) from what has historically been approximately 25 percent to BTE of 37 percent
and above. This focus on improving gasoline SI engines has resulted in the adoption of
technologies such as gasoline direct injection (GDI), turbo-charging and downsizing, Atkinson
Cycle, Miller Cycle, increased valve control authority through variable valve timing and variable
valve lift, integrated exhaust manifolds, reduced friction, and cooled EGR.  Vehicle
manufacturers have more choices of technology for internal combustion engines than at any
previous time in automotive history and more control over engine operation and combustion. In
addition, manufacturers have access to improved design tools  that allow them to investigate and
simulate a wide range of technology  combinations to allow them to make the best decisions
regarding the application of technology into individual vehicles. Despite the access to improved
tools and simulation, EPA believes that manufacturers have not yet explored the entire design
space of modern powertrain architectures and that innovation will continue resulting in
improvements in efficiency that are beyond what is currently being demonstrated in the new car
fleet.

   As discussed in Chapter 3, the use of many of the major powertrain technologies analyzed in
the 2012 FRM, including engine technologies such as VVT, direct injection, turbocharging, and
cylinder deactivation have increased since the publication of the FRM and appear to be  trending
towards EPA projections of technology penetration levels from the 2017-2025 FRM analysis
(see Chapter 3). Engines equipped with GDI are projected to  achieve a 46 percent market share
in MY2015. Approximately 18 percent of new vehicles are projected to be equipped with
                                              5-12

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
turbochargers for MY2015.  Use of cylinder deactivation has grown to capture a projected 13
percent of light-duty vehicle production for MY2015. Light duty diesel vehicles are projected to
increase to a projected 1.5 percent of new vehicle production for MY2015, which is the highest
level since MY1984.  Recently introduced light-duty diesels in the U.S. include several new
pickup truck (2015 Ram 1500, 2016 Chevrolet Colorado, 2016 GMC Canyon) and SUV (2015
Jeep Grand Cherokee, 2016 Land Rover Range Rover, Mercedes GLE300 and GLE350) models.

                 100%
                  90%

                  8O%

              o   70%
              1j   60%

              £   50%
              if
              £  30%

                  20%
                   0%
                          WT   Multi-    GDI    Turbo  Cylinder  Diesel
                                 valve                    Deact.
         Figure 5.1 Light-duty Vehicle Engine Technology Penetration since the 2012 Final Rule
5.2.2.1 Overview of Engine Technologies

   Since the FRM, the agencies have continued to meet with automobile manufacturers, major
Tier 1 automotive suppliers and major automotive engineering services firms to review both
public and confidential data on the development of advanced internal combustion engines for
MY2022 and later. A considerable amount of new work has been completed both within the
agencies and within industry and academia that is available for consideration for the Draft TAR.
The agencies have completed several engine benchmarking programs that have produced
detailed engine maps.  These engine maps represent some of the best performing engines
available today and have been used in the ALPHA and Autonomie models to directly estimate
the effectiveness of modern powertrain technology being applied to a wide spectrum of vehicle
applications. In addition, industry and academia regularly publishes similar levels of detail with
regard to engine operation in the public domain, and the agencies have also used this information
to either directly inform or to compare effectiveness estimations.

   In addition to creating detailed engine maps for full vehicle simulation, the agencies
conducted proof-of-concept, applied research to investigate the potential for further engine
improvements. This includes the use of both computer-aided engineering tools and the
                                             5-13

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
development and analysis of advanced engine technologies via engine dynamometer testing.
Further details are provided in Section 5.4.

   In meetings with automobile manufacturers and Tier 1 suppliers, we learned about both
convergent and divergent engine technologies trends. In many cases, it was difficult to obtain
information on specific engine technologies beyond MY2022. Through MY2022, with few
exceptions, gasoline direct injection and VVT will be applied to most engines.  Significant
attention will be placed on reducing engine friction and accessory parasitic loads. In passenger
car and smaller light-duty truck segments, there will  be considerable diversity of engine
technologies, including turbocharged GDI engines with up to 25-bar BMEP, both turbocharged
and naturally aspirated GDI engines with external cooled EGR, and engines that combine GDI
with operation over the Atkinson Cycle and use of Atkinson Cycle outside of HEV applications.
With respect to larger, heavier vehicles, including full-size SUVs and pickup trucks with
significant towing utility, some manufacturers will be relying on naturally aspirated GDI engines
with cylinder deactivation, some will be relying more on turbocharged-downsized engines, and
others will be using a variety of engine technologies, including light-duty diesels.  Vehicle
manufacturers are at advanced stages of research with respect to:

       •  Stratified-charge, lean-burn combustion
       •  Multi-mode combustion approaches
          0  homogenous charge, compression ignition, lean-burn operation at light loads
          0   stratified-charge, lean-burn spark ignition at moderate loads
          0   stoichiometric homogenous charge, spark ignition at high loads
       •  Variable-compression ratio (VCR) engines
       •  Engines exceeding 24-bar BMEP
   While the introduction of variable compression ratio engines and highly boosted GDI engines
above 24-bar BMEP is expected within the 2022-2025 timeframe, these technologies will most
likely be introduced into relatively low-volume, high performance applications.  Manufacturers
and suppliers are finding that turbocharged engines can achieve lower CCh emissions over the
regulatory drive cycles and improved real-world fuel economy at more moderate (24 bar and
below) BMEP levels.  While there are both performance and efficiency advantages to VCR at
high BMEP levels, both Atkinson Cycle and Miller Cycle with VVT are technologies that
compete with VCR and that have a comparable ability to vary effective compression ratio but
with reduced cost and complexity.

   We also learned from manufacturers and suppliers that specific engine technologies have
synergies with other CCh-reduction technologies.  For example, measures to reduce engine
friction, particularly friction at startup, help reduce the motor torque necessary for restart in 12V
start/stop systems. GDI and electric cam phasing systems can be used for combustion assistance
of engine restart.  There are also synergies between Miller Cycle, IEM, cooled-EGR, and the use
of VNT turbochargers which are described in more detail in Section 5.2.2.7'.

   Despite recent EPA and California ARB compliance actions with respect to light-duty diesel
NOx emissions, diesel engines remain a technology for the reduction of GHG emissions from
light-duty vehicles. Advances in NOx and PM emissions control technology are bringing light-
duty diesels fully  into compliance with Federal Tier 3 and California LEV III emissions
                                             5-14

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
standards at a cost that is competitive with the cost-effectiveness other high efficiency, advanced
engine technologies.  In the FRM, diesel powertrains were not expected to be a significant
technology for improving vehicle efficiency, however, since then many new light-duty vehicles
have been introduced to the U.S. market with diesel engines, including the Ram 1500 full-size
pickup truck, the Chevrolet Colorado mid-size pickup truck, the Jeep Grand Cherokee SUV, and
the Chevrolet Cruze.  In addition, diesel engines are continuing to evolve using technologies
similar to those being introduced in new light-duty gasoline engines and heavy-duty diesel truck
engines, including the use of advanced friction reduction measures, increased turbocharger
boosting and engine downsizing, engine "downspeeding," the use of advanced cooled EGR
systems, improved integration of charge air cooling into the air intake system, and improved
integration of exhaust emissions control systems for criteria pollutant control.  The best BTE of
advanced diesel engines under development for light duty applications is now 46 percent and
thus is approaching that of heavy-duty diesel truck engines.2

   In  addition to a reevaluation all the cost  and effectiveness values of the technologies that
were considered in the FRM, this assessment includes evaluations of technologies where
substantial new information has emerged since the FRM, including Atkinson and Miller cycle
engines, and application of cylinder deactivation operation to 3-cylinder, 4-cylinder, and
turbocharged engines.

5.2.2.2 Sources of Engine Effectiveness Data

   In addition to the sources of engine CCh  effectiveness data used  in the 2017-2025 LD GHG
FRM, the agencies also used engine data from a wide range of sources to update engine
effectiveness for this assessment:

       •  Newly available, public data (e.g., peer-reviewed journals, peer-reviewed technical
          papers, conference proceedings)
       •  Data directly acquired by EPA via engine dynamometer testing at EPA-NVFEL or at
          contract laboratories
       •  Benchmarking and simulation modeling of current and future engine configurations
       •  Confidential data from OEMs, Tier 1 suppliers, and major automotive engineering
          services firms
       •  Data from the U.S. Department of Energy Vehicle Technologies Program
   A considerable amount of brake-specific fuel consumption (BSFC), brake-thermal efficiency
(BTE) and chassis-dynamometer drive cycle fuel consumption data for advanced powertrains has
been published in journals, technical papers and conference proceedings since the publication of
the 2012 FRM. In some cases, published data includes detailed engine maps of BSFC and/or
BTE over a wide area of engine operation. In addition, these publications provide a great deal of
information regarding the specific design changes made to an engine which allow the engine to
operate at an improved BSFC and vehicles to operate with improved fuel consumption.  These
design details often include changes to engine friction, changes to valvetrain and valve control,
combustion chamber design and combustion control, boosting components and boosting control,
and exhaust system modifications.  This information provides the agency an indication of which
technologies to investigate in more detail and offer the opportunity  to correlate testing and
simulation results against  currently available and future designs.
                                             5-15

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Since 2012, many examples of advanced engine technologies have gone into production for
the U.S., European and Japanese markets. EPA has acquired many vehicles for chassis
dynamometer testing and has developed a methodology for conducting detailed engine
dynamometer testing of engines and engine/transmission combinations. Engine dynamometer
testing was conducted both at the EPA-NVFEL facility in Ann Arbor, MI and at other test
facilities under contract with EPA. Engine dynamometer testing of production engines outside
of the vehicle chassis required the use of a vehicle-to-engine (or vehicle-to-engine/transmission)
wiring tether and simulated vehicle feedback signals in order to allow use of the vehicle
manufacturer's engine management system and calibrated control parameters.  NHTSA
conducted engine dynamometer testing of light-duty truck engines at Southwest Research
Institute. In addition to fuel consumption and regulated emissions, many of the engines were
also instrumented with piezo-electric cylinder pressure transducers and crankshaft position
sensors to allow calculation of the apparent rate of heat release and combustion phasing.
Engines with camshaft-phasing were also equipped with camshaft position sensors to allow
monitoring of the timing of valve events.  Engine dynamometer testing also incorporated
hardware-in-the-loop simulation of drive cycles so that vehicle packages with varying
transmission configurations and road-loads could be evaluated.

   While the confidential data provided by vehicle manufacturers, suppliers and engineering
firms cannot be published in the Draft TAR, these sources of data were important as they
allowed the EPA to perform quality and rationality checks against the data that we are making
publically available. In each  case where a specific technology was benchmarked, EPA met with
the vehicle manufacturer.  In  cases where expected combinations of future engine technologies
were not available for testing from current production vehicles, a combination of proof-of-
concept engine dynamometer testing and engine and vehicle CAE simulations were used to
determine drive cycle effectiveness. For example, use of cooled EGR and an increased
geometric compression ratio was modeled using Gamma Technologies GT-Power simulations of
combustion and gas dynamics with subsequent engine dynamometer validation conducted using
a prototype engine  management system, a developmental external low-pressure cooled EGR
system,  and a developmental  dual-coil offset ignition system.  Finally, several of these
benchmarking activities were the subject of technical papers published by SAE and included a
peer review of the results as part of the publication process.

5.2.2.3 Low Friction Lubricants (LUB)

   One of the most basic methods of reducing fuel consumption in gasoline engines is the use of
lower viscosity engine lubricants. More advanced multi-viscosity engine oils are available today
with improved performance in a wider temperature band and with better lubricating properties.
This can be accomplished by  changes to the oil base stock (e.g., switching engine lubricants from
a Group I base oils to lower-friction, lower viscosity Group III synthetic) and through changes to
lubricant additive packages (e.g., friction modifiers and viscosity improvers).  The use of 5W-30
motor oil is now widespread and auto manufacturers are introducing the use of even lower
viscosity oils, such as 5W-20 and OW-20, to improve cold-flow properties and reduce cold start
friction. However, in some cases, changes to the crankshaft, rod and main bearings and changes
to the mechanical tolerances of engine components may be required. In all cases, durability
testing is required to ensure that durability is not compromised. The shift to lower viscosity and
lower friction lubricants also  improve the effectiveness of valvetrain technologies such as
cylinder deactivation, which rely on a minimum oil temperature (viscosity) for operation.
                                             5-16

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.2.2.4 Engine Friction Reduction (EFR1, EFR2)

   In addition to low friction lubricants, manufacturers can also reduce friction and improve fuel
consumption by improving the design of engine components and subsystems. Approximately 10
percent of the energy consumed by a vehicle is lost to friction, and just over half is due to
frictional losses within the engine. Examples include improvements in low-tension piston rings,
piston skirt design, roller cam followers, improved crankshaft design and bearings, material
coatings, material substitution, more  optimal thermal management, and piston and cylinder
surface treatments. Additionally, as computer-aided modeling software continues to improve,
more opportunities for evolutionary friction reductions may become available.

   All reciprocating and rotating components in the engine are potential candidates for friction
reduction, and minute improvements in several components can add up to a measurable fuel
economy improvement.

5.2.2.5 Cylinder Deactivation (DEAC)

   In conventional spark-ignited engines throttling the airflow controls engine torque output. At
partial loads, efficiency can be improved by using cylinder deactivation instead of throttling.
Cylinder deactivation (DEAC) can improve engine efficiency by disabling or deactivating
cylinders when the load is significantly less than the engine's total torque capability - the valves
are kept closed, and no fuel is injected - as a result, the trapped air within the deactivated
cylinders is simply compressed and expanded as an air spring,  with reduced friction and heat
losses.  The active cylinders combust at additional loads to compensate for the deactivated
cylinders. Pumping losses are significantly reduced as long as the engine is operated in this
"part-cylinder" mode.

   Cylinder deactivation control strategy relies on setting maximum manifold absolute pressures
or predicted torque within which it can deactivate the cylinders. Noise and vibration issues
reduce the operating range to which cylinder deactivation is allowed, although manufacturers
continue exploring vehicle changes that enable increasing the amount of time that cylinder
deactivation might be suitable. Some manufacturers have adopted active engine mounts and/or
active noise cancellations systems to address NVH concerns and to allow a greater operating
range of activation.

5.2.2.6 Variable Valve Timing (VVT) Systems

   Variable valve timing (VVT) is a family of valve-train designs that alter the timing of the
intake valve, exhaust valve, or both, primarily to reduce pumping  losses, increase specific power,
and control the level of residual gases in the cylinder. VVT reduces pumping losses when the
engine is lightly loaded by controlling valve timing closer to an optimum needed to sustain
horsepower and torque. VVT can also improve volumetric efficiency at higher engine speeds
and loads. Additionally, VVT can be used to alter (and optimize) the effective compression ratio
where it is advantageous for certain engine operating modes (e.g., in the Atkinson Cycle).

   VVT has now become a widely adopted technology. In MY2015, more than 98 percent of
light-duty vehicles sold in the U.S. are projected to use some form of VVT. The three major
types of VVT are  listed in the sub-sections below.
                                             5-17

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Each of the three implementations of WT uses a cam phaser to adjust the camshaft angular
position relative to the crankshaft position, referred to as "camshaft phasing." The phase
adjustment results in changes to the pumping work required by the engine to accomplish the gas
exchange process. The majority of current cam phaser applications use hydraulically-actuated
units, powered by engine oil pressure and managed by a solenoid that controls the oil pressure
supplied to the phaser.  Electric cam phasing allows a wider range of camshaft phasing, faster
time-to-position, and allows adjustment of camshaft phasing under conditions that can be
challenging for hydraulic systems, for example, during and immediately after engine startup.

5.2.2.6.1      Intake Cam Phasing (ICP)

   Valvetrains with ICP can modify the timing of the inlet valves by phasing the intake camshaft
while the exhaust valve timing remains fixed. This requires the addition of a  cam phaser on each
bank of intake valves on the engine. An in-line 4-cylinder engine has one bank of intake valves,
while V-configured engines have two banks of intake valves.

5.2.2.6.2      Cowled Cam Phasing (CCP)

   Valvetrains with coupled (or coordinated) cam phasing can modify the timing of both the inlet
valves and the exhaust valves an equal amount by phasing the camshaft of a single overhead cam
(SOHC) engine or a cam-in-block, overhead valve (OHV) engine. For overhead cam engines,
this requires the addition of a cam phaser on each bank of the engine. Thus, an  in-line 4-cylinder
engine has one cam phaser, while SOHC V-engines have two cam phasers. For overhead valve
(OHV) engines, which have only one camshaft to actuate both inlet and exhaust valves, CCP is
the only VVT implementation option available and requires only one cam phaser.

5.2.2.6.3      Dual Cam Phasins (DCP)

   The most flexible VVT design is dual (independent) cam phasing, where the intake and
exhaust valve opening and closing events are controlled independently. This option allows the
option of controlling valve overlap, which can be used as an internal EGR strategy.  At low
engine loads, DCP creates a reduction in pumping losses, resulting in improved fuel
consumption/reduced CO2 emissions.  Increased internal EGR also results in lower engine-out
NOx emissions. The amount by which fuel consumption is improved and CO2 emissions  are
reduced depends on the residual tolerance of the combustion system and on the  combustion
phasing achieved. Additional improvements are observed at idle, where low valve overlap could
result in improved combustion stability, potentially reducing idle fuel consumption.

5.2.2.6.4      Variable Valve Lift (WL)

   Controlling the lift of the valves provides a potential for further efficiency  improvements.  By
optimizing the valve-lift profile for specific engine operating regions, the pumping losses  can be
reduced by reducing the amount of throttling required to produce the desired engine power
output.  By moving the throttling losses further downstream of the throttle valve, the heat
transfer losses that occur from the throttling process are directed into the fresh charge-air mixture
just prior to compression, delaying the onset of knock-limited combustion. Variable valve lift
control can also be used to induce in-cylinder mixture motion, which improves fuel-air mixing
and can result in improved thermodynamic efficiency.  Variable valve lift control can also
potentially reduce overall valvetrain friction.  At the  same time, such systems may incur
increased parasitic losses associated with their actuation mechanisms.  A number of
                                             5-18

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
manufacturers have already implemented VVL into all (BMW) or portions of their fleets
(Toyota, Honda, and GM), but overall this technology is still available for application to most
vehicles.  There are two major classifications of variable valve lift, discrete variable valve lift
(DVVL) and continuous variable valve lift (CVVL).

   DVVL systems allow the selection between two or three discrete cam profiles by means of a
hydraulically-actuated mechanical system. By optimizing the cam profile for specific engine
operating regions, the pumping losses can be reduced by reducing the amount of throttling
required to produce the desired engine power output. This increases the efficiency of the engine.
These cam profiles may consist of a low and a high-lift lobe or other combinations of cam
profiles, and may also include an inert or blank lobe to incorporate cylinder deactivation (in the
case of a 3-step DVVL system). DVVL is normally applied together with VVT control. DVVL
is also known as Cam Profile Switching  (CPS). DVVL is a mature technology with low
technical risk.

   In CVVL systems, valve lift is varied by  means of a mechanical linkage, driven by an actuator
controlled by the engine control unit. The valve opening  and phasing vary as the lift is changed
and the relation depends on the geometry of the mechanical system. BMW has considerable
production experience with CVVL systems  and has versions of its "Valvetronic" CVVL system
since 2001. CVVL allows the airflow into the engine to be regulated by means of intake valve
opening reduction, which improves engine efficiency by reducing pumping losses from throttling
the intake system further upstream as with a conventionally throttled engine. CVVL provides
greater effectiveness than DVVL, since it can be fully optimized for all engine speeds and loads,
and is not limited to a two or three step compromise. There may also be a small reduction in
valvetrain friction when operating at low valve lift, resulting in improved low load fuel
consumption for cam phase control with variable valve lift as compared to cam phase control
only. Most of the fuel economy effectiveness is achieved with variable valve lift on the intake
valves only.  CVVL is only applicable to double overhead cam (DOHC) engines.

5.2.2.7 GDI, Turbocharging, Downsizing and Cylinder Deactivation

   Between 2010 and 2015, automotive manufacturers have been adopting advanced powertrain
technologies in response to GHG and CAFE standards (see 3.2 Technology Penetrations). Just
over 45 percent of MY2015 light-duty vehicles in U.S. were equipped with gasoline direct
injection (GDI) and approximately 18 percent of MY2015 light-duty vehicles were turbocharged.
Nearly all vehicles using turbocharged spark-ignition engines also used GDI to improve
suppression of knocking combustion.  GDI provides direct cooling of the in-cylinder charge via
in-cylinder fuel vaporization.3 Use of GDI allows  an increase of compression ratio of
approximately 0.5 to 1.5 points relative to naturally aspirated or turbocharged engines using port-
fuel-injection (e.g., an increase from 9.9:1 for the 5.3L PFI GM Vortec 5300 to 11:1 for the 5.3L
GDI GM EcotecS with similar 87 AKI gasoline octane requirements).

   Figure 5.2 shows a comparison of brake thermal efficiency (BTE) versus engine speed  and
load between a high-volume, MY2008 2.4L 14 engine equipped with PFI and a MY2013 GM
Ecotec™ 2.5L 14 equipped with GDI.  The GDI engine has a significantly higher compression
ratio, (11.3:1 vs 9.6:1), higher efficiency throughout its range of operation, and achieves higher
BMEP levels (approximately 12.5 bar vs 11.3 bar), allowing a significant increase in power per
displacement. The incremental effectiveness at approximately 2-bar BMEP and 2000 rpm was
                                             5-19

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
17 percent but varied from approximately 3 percent to approximately 11 percent at other speed
and load points of importance for the regulatory drive cycles.
  12
  10
                                                12
                                                10
                                               —6
                                               a.
                                        15 kw
                                        7.5 kW
                          15 kW
                          7.5 kW
       1000
              2000    3000    4000   5000
                  Engine Speed (rpm)
                                      6000
                                                      1000
2000   3000   4000   5000
    Engine Speed (rpm)
                                                                                    6000
  Figure 5.2 Comparison of BTE for A Representative MY2008 2.4L 14 NA DOHC PFI 4-valve/cyl. Engine
   with Intake Cam Phasing (Left)D and a GM Ecotec 2.5L NA GDI Engine with Dual Camshaft Phasing
                                         (Right).E
Area of Operation > 34% BTE is Shown in Light Green. Area of Operation >35% BTE is Shown in Dark Green.

   Toyota's D-4S system combines GDI and PFI systems, with two injectors per cylinder (one
directly in-cylinder and one immediately upstream of the intake port).4'5'6 As of 2015, all Toyota
vehicles in the U.S. with GDI appear to be using a variation of the D-4S dual GDI/PFI fuel
injection system. This system increases peakBMEP, provides additional flexibility with respect
to calibration of the EMS for improved cold-start emissions and offers an efficiency
improvement over GDI alone. Based on certification data and EPA confirmatory test data,
Toyota vehicles using engines equipped with the D4S system have relatively low PM emissions
over the FTP75 cycle that are roughly comparable to PFI-equipped vehicles (<0.60 mg/mi).7  A
comparison of the Toyota 2GR-FSE engine is shown compared to a 3.5L PFI engine in Figure
5.3. The 2GR-FSE achieves a very high BMEP for a naturally aspirated engine (13.7 bar).
Although both engines have comparable displacement, they are not directly comparable because
the higher BMEP attained by the 2GR-FSE would allow further engine downsizing for a similar
application, with potential for further improvement in BTE at light load relative to the 3.5L PFI
engine.  The area greater than 34 percent BTE is significantly larger for the Toyota 2GR-FSE
due to a combination of factors, including a higher compression ratio enabled by GDI and
reduced pumping losses through use of a dual camshaft phasing system that enables internal
EGR at light loads.
D Based on engine dynamometer test data provided to EPA as part of "Light Duty Vehicle Complex Systems
  Simulation," EPA Contract No. EP-W-07-064, work assignment 2-2, with PQA and Ricardo.
E Based on EPA engine dynamometer test data.
                                              5-20

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
 12


 M

  8
        10CO
              20C-3
                    3000   4000   500D
                    Engine Speed irpm|
                                     6000
1000   2000   3000    4000   6000   8000
           Engln«Sp*»d|rprn)
  Figure 5.3 Comparison of BTE for A Representative MY2010 3.5L V6 NA PFI 4-valve/cyl. EngineF (Left)
            and a Toyota 2GR-FSE GDI/PFI Engine with Dual Camshaft Phasing6 (Right).
Note: Area of Operation > 34% BTE is Shown in Light Green.
   The recently redesigned Ford turbocharged 3.5L "EcoBoost™" engine in the 2017 Ford F150
also uses a dual GDI/PFI injection system to increase power, reduce emissions, and improve
efficiency,8 but other engines in Ford's EcoBoost lineup use GDI. In MY2015, Ford offered a
version of the EcoBoost turbocharged GDI engines as standard or optional engines in nearly all
of models of light-duty cars and trucks. Ford's world-wide production of EcoBoost engines
exceeded 200,000 units per month during CY2015.9

   Approximately 13 percent of MY2015 light-duty vehicles used cylinder deactivation,
primarily in light-duty truck applications. In MY2015, General Motors introduced their
"EcotecS" line of OHV V6 and V8 engines across their entire lineup of light-duty pickups and
truck-based SUVs.  These engines are equipped with GDI, coupled-cam-phasing, and cylinder
deactivation.  Both the V6 and V8 EcoTecS engines are capable of operation on 4-cylinders
under light-load conditions. Application of GDI has synergies with cylinder deactivation. The
higher BMEP achievable with GDI also increases the BMEP achievable once cylinders have
been deactivated, thus increasing the range of operation where cylinder deactivation is enabled.

   Cylinder deactivation operates the remaining, firing cylinders at higher BMEP under light
load conditions. This moves operation of the remaining cylinders to an area of engine operation
with less throttling and thus lower pumping losses (Figure 5.4) and reduced BSFC.
F Based on engine dynamometer test data provided to EPA as part of "Light Duty Vehicle Complex Systems
  Simulation," EPA Contract No. EP-W-07-064, work assignment 2-2, with PQA and Ricardo.
G Based on EPA engine dynamometer test data.
                                              5-21

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                                           v	>  BSFC, g/kW-hr
                                                                 Approximate road load
  12
  10
 LU
 ^
 CD
            Complete Engine Operation
           50% Cylinder Deactivation
12 -
     1000   2000  3000   4000   5000   6000
                Engine Speed (rpm)
   1000   2000   3000  4000   5000   6000
              Engine Speed (rpm)
   Figure 5.4 Graphical Representation Showing How Cylinder Deactivation Moves Engine Operation to
  Regions of Operation with Improved Fuel Consumption over the UDDS Regulatory Drive Cycle (shaded
                                          area).
   In the 2017-2025 LD GHG FRM, EPA limited its analysis of cylinder deactivation to engines
with six or more cylinders.  At the time, there were concerns that application of cylinder
deactivation to 3 or 4-cylinder engines would result in unacceptable NVH. Since 2012,
improvements in crankshaft dampening systems have extended the application of cylinder
deactivation to four cylinder engines. Volkswagen introduced their 1.4L TSIEA 211
turbocharged GDI engine with "active cylinder management" in Europe for MY2013.10 This
engine is the first production application of cylinder deactivation to an 14 engine and can
deactivate 2 cylinders via cam-shifting under light load conditions. VW recently introduced a
Miller Cycle variant of the same EA211 engine family with cylinder deactivation.11 Schaeffler
has developed a dynamic cylinder deactivation system for 13 and 15 engines that alternates or
"rolls" the deactivated cylinders.  This system  allows all cylinders to be deactivated after every
ignition cycle and reactivated during the next cycle. Cylinder deactivation thus alternates within
a single deactivation phase and not each time a new deactivation mode is introduced.  The net
result is that engines with an odd number of cylinders can operate, on average, with half their
cylinder displacement (i.e.,  13 can drop to 1.5 cylinders on average or an 15 can drop to 2.5
cylinders on average). Ford and Schaeffler investigated both rolling cylinder deactivation and a
system to deactivate one cylinder with Ford's EcoBoost l.OL 13 engine and found that, with
appropriate vibrational dampening, either strategy could be implemented with no NVH
deterioration and with 3 percent or greater improvement in both real-world and EU drive cycle
fuel economy.12  Tula Technology has demonstrated a system with the capability of deactivating
                                              5-22

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
any cylinder that they refer to as "Dynamic Skip Fire."13 Tula found a combined-cycle fuel
economy improvement of approximately 14 percent for an unspecified vehicle equipped with a
6.2L PFI V8 and approximately 6 percent for an application equipped with the GM Active Fuel
Management 4/8 cylinder deactivation system.  It should be noted that engines with more
opportunity for pumping loss reduction over the regulatory drive cycles (e.g., larger
displacement, naturally aspirated, PFI) generally have higher CCh effectiveness when equipped
with cylinder deactivation.

   Many automotive manufacturers have launched a third or fourth generation of GDI engines
since their initial introduction in the U.S. in 2007. Turbocharged, GDI engines are in now in
volume production at between 21-bar and 25-bar BMEP.  Most recent turbocharged engine
designs now use head-integrated, water-cooled exhaust manifolds and coolant loops that separate
the cooling circuits between the engine block and the head/exhaust manifold(s).  Head-integrated
exhaust manifolds (IBM) are described further in the section on thermal management in 5.2.2.11.
The use of IBM was assumed within the EPA analysis of 27-bar BMEP turbocharged GDI
engines for the FRM.  The benefits, including increased ability to downspeed the engine without
pre-ignition and the potential for cost savings in the design of the turbocharger turbine housing
appear to extend to lower BMEP-level turbocharged GDI engines and will likely be incorporated
into many future turbocharged light-duty vehicle applications. The application of IBM's does
effect cooling system design and manufacturers will be required to provide sufficient cooling
system capacity if they adopt this technology.

   The 2.7L Ford Ecoboost engine was introduced in the MY2015 Ford F150. This engines uses
one turbocharger per bank,  IEM and dual camshaft phasing. Peak BMEP is approximately 24-bar
and the maximum towing capacity of the F150 equipped with this engine is 13,300 Ibs. when
used with a 3.73:1 final drive ratio in the 2016 Ford F150.  Figure 5.5 shows a comparison of
BMEP and torque vs. engine speed and BTE between a conventional MY2010 5.4L OHC V8
light-duty pickup truck engine and the MY 2015 2.7L Ford Ecoboost engine. This comparison
thus represents 50 percent engine downsizing using turbocharging and GDI. The 2.7L Ecoboost
engine has bother higher peak torque and power, higher peak BTE, and approximately double the
area above 34 percent BTE.
                                             5-23

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
 12
 10
  . 350
— 6.
                                               24
                                  30 kW
                                  15kW
        1000
              2000   3000   4000   5000
                  Engine Speed (rpm)
                        30 kW
                        15kW
                                                      1000
2000   3000   4000   5000
    Engine Speed (rpm)
  Figure 5.5 Comparison of BTE for A Representative MY2010 5.4L V8 NA PFI 3-valve/cyl. Engine11 (Left)
     and a Ford 2.7L V6 Ecoboost Turbocharged, GDI Engine With Dual Camshaft Phasing1 (Right).
Note: Area of Operation > 35% BTE is Shown in Green.
   Figure  5.6 shows maps of BMEP and torque vs. engine speed and BTE for a representative
MY2010 2.4L PFI engine with intake camshaft phasing and a MY2012 l.OL Ford EcoBoost
turbocharged, GDI, engine with an integrated exhaust manifold (IEM) and dual camshaft
phasing.14 The l.OL EcoBoost engine features turbocharging to a peak BMEP of 25-bar, GDI
with center-mounted, spray-guided injection, a cylinder-head integrated exhaust manifold, and
dual camshaft phasing. While not a direct comparison for purposes of engine downsizing (the
l.OL EcoBoost is more comparable to a 1.8 - 2.0L NA PFI engine based on torque
characteristics and rated power), this comparison of BTE does demonstrate the manner that
turbocharging and downsizing can be used to expand regions of high thermal efficiency to cover
a larger portion of engine operation.  For example, the EcoBoost engine exceeds 30 percent BTE
above 6-bar BMEP/50 N-m torque over most of the engine's range of engine speeds while the
area above 30 percent BTE for the NA PFI engine is considerably  smaller.
H Based on engine dynamometer test data provided to EPA as part of "Light Duty Vehicle Complex Systems
  Simulation," EPA Contract No. EP-W-07-064, work assignment 2-2, with PQA and Ricardo.
1 Based on EPA engine dynamometer test data.
                                              5-24

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
 n
 ii
I.
       1000
             2000   3000    4000    5000
                  Engine Speed |rpm)
                                      6000
                                                                                      1&UI
                                                                                      - >
1000   2000   3000    4000    6000
           Engine Speed |rpm|
                                                                                    6C'OC>
 Figure 5.6 Comparison of BTE for A Representative MY2010 2.4L NA PFI EngineJ (Left) and A Modern,
                       l.OL Turbocharged, Downsized GDI EngineK (Right).
Note: Area of Operation > 34% BTE is Shown in Light Green.
   A comparison of the same 2.4L PFI engine with a more recent, MY2017 Honda 1.5L
Turbocharged GDI engine with IBM is shown in Figure 5.7.15'16  The torque characteristics of the
Honda engine are a closer match to the 2.4L PFI engine and the Honda engine represents
approximately 37 percent downsizing relative to the 2.4L PFI engine due to turbocharging and
includes other improvements  (friction reduction, dual cam phasing, higher rates of internal
EGR). The Honda 1.5L turbocharged GDI engine has significantly improved efficiency when
comparing BTE across 20 speed and load points of significance for the regulatory drive cycles
(1500 -2500 rpm and 2-bar to 8-bar BMEP as referenced to the 2.41 ENGINE). The BTE of the
Honda 1.5L turbocharged  engine showed an incremental effectiveness of 6 percent to 30 percent
across this entire range of operation. The difference was more pronounced at lighter loads.
Incremental effectiveness was 16 percent to 30 percent below 6-bar BMEP relative to the 2.4L
engine (~112 N-m of torque).
1 Based on engine dynamometer test data provided to EPA as part of "Light Duty Vehicle Complex Systems
  Simulation," EPA Contract No. EP-W-07-064, work assignment 2-2, with PQA and Ricardo.
K Adapted from Ernst et al. 2011.14
                                              5-25

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
 12
                                       15 kw
                                       7.5 kW
       1000
             2000   3000   4000   5000
                 Engine Speed (rpm)
                                     6000
                                                     1000
                                                          2000
3000   4000   5000
    Engine Speed (rpm)
                                                                                   6000
  Figure 5.7 Comparison of BTE for A Representative MY2010 2.4L NA PFI EngineJ (Left) and A Modern,
                       1.5L Turbocharged, Downsized GDI EngineL (Right).
Note: Area of Operation > 34% BTE is Shown in Light Green. Area of Operation >35% BTE is Shown in Dark
Green. BTE Was Also Compared Across 20 Operational Points of Significance for Regulatory Drive Cycles
between 1500 and 2500 RPM.
   Recent turbocharger improvements have included use of lower-mass, lower inertia
components and lower friction ball bearings to reduce turbocharger lag and enable higher peak
rotational speeds.  Improvements have also been made to turbocharger compressor designs to
improve compressor efficiency and to expand the limits of compressor operation by improving
surge characteristics (see Figure 5.8).
L Adapted from Wada et al. 2016 and Nakano et al 2016.1
                                               5-26

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                             Original
                            Surge Line
                                    60.109
                               0.02
                                     0.04   0.06
                                    Volume (tow rate *,
0.08    0,10    0.12 (nflte]
 Figure 5.8 Typical Turbocharger Compressor Map Showing How Pressure And Flow Characteristics Can
 Be Matched Over a Broader Range of Engine Operation Via Surge Improvement and Higher Operational
                                          Speed.

   Turbochargers with variable nozzle turbines (VNT) use moveable vanes within the
turbocharger to allow adjustment of the effective exhaust turbine aspect ratio, allowing the
operation of the turbocharger to be better matched across the entire speed and load range of an
engine. VNT turbochargers are commonly used in modern light-duty and heavy-duty diesel
engines.  The use of head-integrated exhaust manifolds (IBM) and split-coolant loops within the
engine and the use of cooled EGR (Sections 5.2.2.8 and 5.2.2.11) can reduce peak exhaust
temperatures sufficiently to allow lower cost implementation of VNT turbochargers in spark
ignition engines.  There are also synergies between the application of VNT and Miller cycle
(increased low-speed torque, improved torque response).11
  Figure 5.9 Cross Sectional View of a Honeywell VNT Turbocharger. The Moveable Turbine Vanes And
                          Servo Linkage Are Highlighted In Light Red.
                                               5-27

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
5.2.2.8 EGR

   Exhaust gas recirculation (EGR) is a broad term used for systems that control and vary the
amount of inert, residual exhaust gases left in cylinder during combustion.  EGR can improve
efficiency at part-load by reducing pumping losses due to engine throttling. EGR also reduces
combustion temperatures and thus reduces NOx formation. The use of cEGR can reduce
knocking combustion, thus allowing compression ratio and/or turbocharger boost pressure to be
increased or spark timing to be advanced.  EGR also slows the rate of combustion, so its use is
often accompanied by other changes to the engine (e.g., inducing charge motion and turbulent
combustion) to shorten combustion duration and allow improved combustion phasing. Internal
EGR uses changes in independent cam-phasing to vary the overlap between intake and exhaust
valve timing events, thus changing the amount of residual gases trapped in cylinder after cylinder
scavenging. External EGR recirculates exhaust gases downstream of the exhaust valve back into
the air induction system. With turbocharged engines, there are variants of external EGR that use
a low pressure loop, a high pressure loop or combinations of the two system types (see Figure
5.10). External  EGR systems can also incorporate a heat-exchanger to lower the temperature of
the recirculated  exhaust gases (e.g., cooled EGR or cEGR), improving both volumetric efficiency
and enabling higher rates of EGR. Nearly all light-duty diesel engines are equipped with cEGR
as part of their NOx emission control system. Some diesel applications also use relatively large
amounts (>25 percent) of cEGR at light- to part-load conditions to enable dilute low-temperature
combustion (see Section 5.2.2.11 for a more detailed description of light-duty diesel
technologies). Research is also underway to apply similar forms of low-temperature combustion
using high EGR rates to gasoline engine applications. This includes lean-homogenous
compression auto ignition (see Section 5.2.2.14) and other homogenous charge compression
ignition concepts (see Section 5.2.2.11).

   The use of cEGR was analyzed as part of EPA's technology packages for post-2017 light-
duty vehicles with engines at 24-bar BMEP, primarily as a means to prevent pre-ignition at the
high turbocharger boost levels needed at 24-bar BMEP and above. The analysis did take into
account efficiency benefits from the use of cEGR with turbocharged engines due primarily to
part-load reductions in pumping losses and the reduction or elimination of commanded fuel
enrichment under high-load conditions.

   Prior to 2012, there were no examples of production vehicles equipped with turbocharged
GDI engines using cEGR.  The PSA 1.2L EB PureTech Turbo engine was recently launched in
the MY2014 Peugeot 308 in Europe as the first high-volume production application  of cEGR on
a turbocharged GDI engine. This engine has over 24-bar BMEP and also operates using Miller
Cycle (see Section 5.2.2.10 for a more detailed description of Miller-Cycle). The MY2016
Mazda CX-9 2.5L SKYACTIV Turbo engine similarly combines the use of Miller Cycle with
cEGR.
                                             5-28

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
          Low Pressure Loop cEGR
High Pressure Loop EGR
         Charge Air
          Cooler
                                      Exhaust
                                      Catalyst
            Airflow
                 EGR
                Coole
  Figure 5.10 A Functional Schematic Example of a Turbocharged Engine Using Two Variants of External
                                          EGR
Note: The Schematic On The Left Shows The Details Of A Low Pressure Loop (Post-Turbine To Pre-Compressor)
Cegr System. The Schematic Inset on the Right Shows High Pressure Loop (Pre-Turbine to Post-Compressor) EGR.17
In The FRM Analysis, Some TDS24 Packages And All TDS27 Packages Used Dual-Loop (Both High And Low
Pressure) EGR.

5.2.2.9 Atkinson Cycle

   Typical 4-cycle internal combustion engines have an effective compression ratio and effective
expansion ratio that are approximately equivalent. Current and past production Atkinson Cycle
engines use changes in valve timing (e.g., late-intake-valve-closing or LIVC) to reduce the
effective compression ratio while maintaining the expansion ratio (see Figure 5.11 and Figure
5.12).  This approach allows a reduction in top-dead-center (TDC) clearance ratio (e.g., increase
in "mechanical" or "physical" compression ratio) to increase the effective expansion ratio
without increasing the effective compression ratio to a point that knock-limited operation is
encountered. Increasing the expansion ratio in this manner improves thermal efficiency but also
lowers peak brake-mean-effective-pressure (BMEP), particularly at lower engine speeds.M
Depending on how it is implemented, some Atkinson Cycle engines  may also have sufficient
cam-phasing authority to widely vary effective compression ratio and can use this variation as a
M BMEP is defined as torque normalized by cylinder displacement. It allows for emissions and efficiency
  comparisons between engines of different displacement.
                                              5-29

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
means of load control without use of the standard throttle, resulting in additional pumping loss
reductions.
                     12
                     10
                    |8
                    •5
                    >
                        Otto-cycle and LIVC Atkinson/Miller Cycle Valve Events
• Exhaust Valve
1 Intake Valve
 LIVC Intake Valve
                       0          180         360          540
                                      Crank Angle (degrees)

  Figure 5.11  Comparison of the Timing of Valve Events for Otto-Cycle (black and orange lines) and LIVC
                 Implementations of Atkinson- Or Miller-Cycle (black and green lines).
                                                 5-30

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                    3
                    (A
                    (A
                    £
                   a.
                    !_
                    0)
                   •c
                    c
                   ">
                   O
                                                    Otto Cycle P-V
                                                    Atkinson Cycle P-V
Compress!
   jssion
Stroke
         LIVC Atkinson Cycle
            Pumping Loss
                                                                Start of
                                                              Compression
          Cylinder Volume
                             1 Otto Cycle
                            Pumping Loss
 Figure 5.12 Diagrams Of Cylinder Pressure Vs. Cylinder Volume For A Conventional Otto-Cycle SI Engine
    (orange line) Compared To A LIVC Implementation of Atkinson Cycle (green line) Highlighting the
                               Reduction in Pumping Losses.
   Prior to 2012, the use of naturally-aspirated Atkinson Cycle engines has been limited to HEV
and PHEV applications where the electric machine could be used to boost torque output,
particularly at low engine speeds. Because of this, EPA's analyses for the FRM did not include
the use of Atkinson Cycle outside of HEV and PHEV applications. Nearly all HEV/PHEV
applications in the U.S. use Atkinson Cycle, including the Honda Insight, Toyota Prius, Toyota
Camry Hybrid, Lexus 400h, Hyundai Sonata Hybrid and Chevrolet Volt. The Toyota 2ZR-FXE
used in the third-generation Toyota Prius and Lexus 200h uses a combination of LIVC Atkinson
Cycle, cooled EGR, and port-fuel-injection (PFI) to achieve a peak BTE of 38.5 percent, the
highest BTE achieved to date for a production spark-ignition engine.  Further refinements to this
engine, including increased tumble to increase both the speed of combustion and EGR tolerance,
have resulted in peak BTE of 40 percent.18

   Since 2012, Atkinson Cycle engines have been introduced into non-hybrid applications.
These applications use camshaft-phasing with a high  degree of authority together with either
GDI (e.g., Mazda SKYACTIV-G 1.5L, 2.0L and 2.5L engines, Toyota 2GR-FKS engine), PFI
(MY2017 Hyundai Elantra "Nu" 2.0-liter PFI Atkinson) or a combination of PFI with cooled
EGR (Toyota 1NR-FKE and 2NR-FKE engines).  As of MY2017, all of Mazda's engines for the
U.S. market are either Atkinson Cycle or Miller Cycle (boosted Atkinson). Toyota's 2GR-FKS
engine became an optional engine offered in the Toyota Tacoma pickup truck beginning in
                                            5-31

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
MY2016.  The Tacoma is currently the mid-size pickup truck segment sales leader in the U.S.
The Toyota Tacoma equipped with the 2GR-FKS Atkinson Cycle engine has an SAE J2807 tow
rating of 6,800 pounds. The Hyundai  "Nu" 2.0-liter PFI Atkinson Cycle engine is the base
engine offering in the Hyundai Elantra. The Hyundai Elantra is currently within the top 5 in
sales within the compact car segment in the U.S.

   The effective compression ratio of Atkinson Cycle engines can be varied using camshaft
phasing to increase BMEP and the use of GDI (Mazda) or cEGR (Toyota) are used, in part, for
knock mitigation.  These engines from Mazda and Toyota also incorporate other improvements,
such as friction reduction from valvetrain and piston design enhancements. The Toyota 1NR-
FKE 1.3L 13 and 2NR-FKE 1.5L 14 engines achieve a peak BTE of 38 percent, very close to the
BTE achieved with the 2ZR-FXE engine used in the Toyota Prius.18'19 EPA testing of 2.0L and
2.5L variants of the Mazda SKYACTIV-G engine achieved peak BTE of 37 percent while using
either 88AKI (91 RON) or 92 AKI (96 RON) fuel. More important from a standpoint of drive-
cycle fuel economy and CO2 emissions was the very large "island" of more than 32 percent BTE
(Figure 5.13) which, depending on the transmission and road load, would cover most operation
over the UDDS and HwFET regulatory drive cycles depending on the specific vehicle
application (e.g., road loads, final drive, gear-ratio spread).  In the case of the Mazda
SKYACTIV-G engines, the use of GDI and cam-phasing resulted in increased BMEP and rated
power relative to the previous PFI, non-Atkinson versions of this engine and allowed a small
degree of engine downsizing (e.g., replacement of the previous 2.5L PFI engine with the 2.0
SKYACTIV-G) on some Mazda platforms with equal or improved performance. In the case of
the Toyota 1NR-FKE, the use of cEGR and cam-phasing allowed BMEP to be maintained
relative to peak BMEP of the Non-Atkinson Cycle engine it replaced and allowed the use of a
lower cost PFI fuel system. Both the Mazda and Toyota Atkinson Cycle engines use electro-
mechanical systems for camshaft phasing on the intake camshaft.
  12
  10
 r6
 I
 m
  4
 200

 180

 160


 140

 120

 100

f
Z^
«L>

I
>-40


 20

  0
135 kW


120 kW


105 kW


90 kW


75kW


iO kW


45 kW


30 kW


15 kW

7.5 kW
                                              12
                                              10
       1000
              2000    3000    4000   5000
                  Engine Speed (rpm)
                                      6000
                                                    1000
                                                       2000    3000   4000    5000
                                                           Engine Speed (rpm)
                                                                                   6000
  Figure 5.13 Comparison of BTE for a Representative MY2010 2.4L NA PFI EngineN (left) and a 2.5L NA
                    GDI LIVC Atkinson Cycle Engine (right) tested by EPA.0'20
N Based upon engine dynamometer data provided to EPA under a contract with PQA and Ricardo, "Light Duty
  Vehicle Complex Systems Simulation" EPA Contract No. EP-W-07-064, work assignment 2-2.
0 Derived from EPA engine dynamometer data first presented by Lee et al. 2016.20
                                             5-32

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   A recent benchmarking analysis by EPA of a 2014 Mazda SKYACTIV-G naturally aspirated
(NA) gasoline direct injection (GDI) engine showed a peak BTE of approximately 37 percent,
relatively high for SI engines.0'21  This was in part due to an ability to use late-intake-valve-
closing (LIVC) Atkinson-cycle operation to decouple the knock-limited effective CR from the
expansion ratio available from a very high 13:1 geometric CR. The Mazda SKYACTIV-G is
one of the first implementations of a naturally-aspirated, LIVC Atkinson-cycle engine in U.S.
automotive applications outside of hybrid electric vehicles (HEV) and also appears to be the first
Atkinson-cycle engine to use GDI. Port-fuel-injected (PFI) Atkinson-cycle engines have been
used in hybrid electric vehicle applications in the U.S. for over a decade. PFI/Atkinson-cycle
engines have demonstrated peak BTE of approximately 39 percent in the 2015 Honda Accord
HEV and 40 percent in the 2016 Toyota Prius HEV. While NA/Atkinson-cycle engines can
achieve comparable or better peak BTE in comparison with downsized, highly boosted,
turbocharged GDI engines like the Ricardo EGRB configuration, modern turbocharged GDI
engines often have relatively high BTE across a broader range of engine speed and torque as well
as improved BTE and fuel consumption at light loads, as shown in Figure 5.14. Based on EPA's
initial engineering analysis of the Mazda SKYACTIV-G engine, it appeared that another
reasonable, alternative technological path to both high peak BTE and a broad range of operation
with high BTE might be possible through the application of cooled-EGR (cEGR), a higher
compression ratio, and cylinder deactivation to a naturally-aspirated GDI/Atkinson-cycle engine
like the SKYACTIV-G.
 12
 10
                                                                                    10 kW
                                                                                    7.5kW
       1000
              2000   3000   4000   5000
                  Engine Speed (rpm)
                                      6000
                                                   1000
                                            2000   3000   4000   5000
                                                Engine Speed (rpm)
                                                                                  6000
Figure 5.14 A Comparison of BSFC Maps Measured For The 2.0L 13:1CR SKYACTIV-G Engine0 (left) and
                   Modeled For A l.OL Ricardo "EGRB Configuration'^ (right).
5.2.2.10
Miller Cycle
   Like Atkinson Cycle, Miller Cycle engines use changes in valve timing to reduce the effective
compression ratio while maintaining the expansion ratio. Automakers have investigated both
early intake valve closing (EIVC) and LIVC variants. There is some disagreement over the
application of the terms Atkinson or Miller Cycle to EIVC and LIVC valve event timing and
sometimes the terms are used interchangeably. For the purpose of EPA's analyses, Miller Cycle
is a variant of Atkinson cycle with intake manifold pressure boosted by a either a turbocharger
                                             5-33

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
and/or a mechanically or electrically driven supercharger. It is simply an extension of Atkinson
Cycle to boosted engines. The first production vehicle offered using Miller Cycle was the
MY1995 Mazda Millenia S, which used the KJ-ZEM 2.3L PFI engine with a crankshaft-driven
Lysholm compressor for supercharging. Until recently, no Miller Cycle gasoline SI engines
were in mass production after 2003, and Miller Cycle was not evaluated as a potential gasoline
engine technology as part of the 2017-2025 GHGFRM.

   As with Atkinson Cycle engines, the use of GDI and camshaft-phasing with a high degree of
authority have significant synergies with Miller Cycle. Modern turbocharger and after cooler
systems allow Miller Cycle engines to attain BMEP levels approaching those of other modern,
downsized, turbocharged GDI engines. The 1.2L 13 PSA "EB PureTech Turbo" Miller engine
recently launched in Europe, N. Africa and S. America in the MY2014 Peugeot 30822. In
addition to Miller Cycle, the engine also uses cEGR. This engine has a maximum BMEP of 24-
bar and is similar in many respects to the Ford l.OL 13 EcoBoost but achieves 35 percent BTE
over a slightly broader area of operation vs. 34 percent BTE for the EcoBoost (see Figure 5.15).
   24

   20
  £.16
  Q_
  LU
  CO
112

  8

  4
                                        -200
                                        -150
                                        -100
                                        - 50
      1000 2000  3000  4000  5000  6000
               Engine Speed (rpm)
                                               CO
                                                                                     -240
                                                                                     -180
                                                                                     -120
                                                                                     - 50
                                                 1000  2000  3000  4000  5000  6000
                                                          Engine Speed (rpm)
             Figure 5.15 Comparison of BTE for Downsized, Turbocharged GDI Engines.
Note: Ford l.OL EcoBoost Engine Is On The Left And A 1.2L Miller Cycle PSA EB Puretech Engine Is On The
Right.  A More Detailed BTE Map Is Not Yet Available For The PSA Engine.
   In MY2016, VW will be launching a Miller Cycle variant of the 2.0L EA888 turbocharged
GDI engine in the U.S.  The VW implementation of Miller Cycle has a second Miller Cycle cam
profile and uses camshaft lobe switching on the intake cam to go into and out of an EIVC version
of Miller Cycle.23'24 The peak BTE of 37 percent is higher than that of the PSA Miller cycle
engine, in part due to a higher  expansion ratio (11.7:1 for the VW engine vs. 10.5:1 for the PSA
engine).  Like the PSA engine, the VW uses high-pressure cEGR.  Peak BTE is comparable to
the Mazda SKYACTIV-G engines but is available over a broader range of speed and load
conditions. Both Atkinson and Miller Cycle engines show broad areas of operation at greater
than 32 percent BTE. Figure 5.16 shows a comparison between a MY2010 3.5L NA PFI DOHC
V6 and the VW 2.0L EA888 Miller Cycle engine with comparable torque delivery. The area  of
operation at greater than 32 percent BTE is approximately double for the Miller Cycle  engine
                                             5-34

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
relative to the DOHC PFI engine. BTE is improved by approximately 40 percent at light load for
the Miller Cycle engine and peak BTE is improved approximately 6 percent.
 12
 10
         1000
               2000
                     3000    4000    5000
                      Engine Speed (rpm)
                                        6000
500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500
              Engine Speed (rpm)
   Figure 5.16  Comparison of BTE for A Representative MY2010 3.5L NA PFI V6 Enginep (Left) And A
                         Downsized 2.0L 14 Miller Cycle Engine0 (Right).
Note: The Light Green Area Shows Regions of >34% BTE. The Dark Green Area Shows a Region >35% BTE.
   Since VW has published detailed data for both Miller Cycle and a turbocharged GDI (non-
Miller) variants of the EA888 series of engines, a more direct comparison between turbocharged,
downsized GDI and Miller Cycle engines is possible.  Figure 5.17 shows BTE for both variants
of the 2.0L 14 VW EA888 engine. When comparing BTE at comparable BMEP, there is a 6-10
percent incremental improvement for the Miller Cycle engine relative to the turbocharged GDI
engine over a broad area of operation from 1500-2500 rpm and from 2-bar to 12-bar BMEP (i.e.,
below  55 - 60 percent of peak BMEP - areas of importance for the regulatory drive cycles).R
Comparing BTE of the 2.0 Miller cycle variant to the smaller displacement, 1.8L version of the
same engine family (similar 22-bar BMEP to the 2.0L turbocharged GDI, but equivalent torque
to the 2.0L Miller Cycle engine) lowers the incremental effectiveness for Miller Cycle to
approximately 4-7 percent relative to  a turbocharged GDI engine and comparable partial load
operation from 1500-2500 rpm.  Confidential business information from a Tier 1 automotive
supplier provided an estimate of approximately 5 percent CCh combined-cycle incremental
benefit for Miller Cycle relative to a 24-bar BMEP turbocharged, downsized engine and a loss of
approximately 8-12 percent peak BMEP due to reduced volumetric efficiency for Miller Cycle.
This is consistent relative to the data published by VW.
p Based upon engine dynamometer data provided to EPA under a contract with PQA and Ricardo, "Light Duty
  Vehicle Complex Systems Simulation" EPA Contract No. EP-W-07-064, work assignment 2-2.
Q Adapted from Wurms et al. 2015.538
R Note that VW did not significantly change the turbocharging system when applying Miller Cycle to this engine
  family, so the Miller Cycle variant has a peak BMEP of 20-bar instead of 22-bar due to the reduced induction
  from LIVC. Turbocharger improvements (e.g., higher pressure ratio and different flow characteristics) would be
  necessary to maintain the 2.0L Miller Cycle engine at 22-bar BMEP, thus comparisons in this case are limited to
  20-bar BMEP and below.
                                               5-35

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
      50010001500 2000 2500 3000 3500 4000 4500 50005500
                    Engine Speed (rpm|
500 1000 1500 20002500 5000 J500 4000 1500 5000 5500
              Engine Spaed irpn»|
Figure 5.17 Comparison of BTE for 2015 Turbocharged, Downsized GDI (left) and 2017 Miller Cycle (right)
                     variants of the same engine family, the 2.0L VW EA888.Q
Note: Green area shows region of high (35%) BTE.
5.2.2.11       Light-duty Diesel Engines

   Diesel engines have characteristics that differ from gasoline spark ignition (SI) engines and
allow improved fuel efficiency, particularly at part-load conditions. These include reduced
pumping losses due to lack of (or greatly reduced) throttling, and a combustion cycle that
operates at a higher compression ratio and at very lean air/fuel ratio when compared with an
equivalent-performance gasoline engine.  Operating with a lean-of-stoichiometric air/fuel ratio
poses challenges with respect to NOx control, requiring either a NOx adsorption catalyst (NAC),
urea or ammonia-based selective catalytic reduction (SCR) or some combination of NAC and
SCR in order to meet Federal Tier 3 and California LEV III NOx emissions standards.
Beginning with Federal Tier 2 emission standards, it has also been necessary to equip light-duty
diesels with catalyzed diesel particulate filters (CDPFs) in order to comply with light duty PM
emission standards.

   Detailed analysis of the vehicle simulation results used within the FRM uncovered some
shortcomings within the MSC EASY5 vehicle simulations used as light-duty diesel vehicle GHG
effectiveness inputs into the Ricardo Surface Response Model. The modeled light-duty diesel
technology packages did not operate in the most efficient regions of engine operation.  This may
have been in part due to inconsistencies in the application of the  optimized shift strategy and in
part due to an oversight that resulted in the apparent oversizing of light-duty diesel engine
displacements.  For example, plotting the average engine speed and load operating points over
the regulatory drive cycles for the MSC EASY5 diesel simulations on top of the diesel engine
maps showed that there was significant potential for improvement in the choice of selected gear.
As a result, additional analyses using the ALPHA vehicle simulation model have been conducted
for light-duty diesel engine technology packages in order to update GHG effectiveness from
these packages.

   Light-duty diesel engines have also evolved considerably over the last five years, particularly
in Europe.  Modern light-duty diesel engine designs appear to be following similar trends to
those of turbocharged GDI engines and, in some cases, heavy-duty diesel engine designs,
including:
                                              5-36

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
       1) Engine downsizing (increased peak BMEP)
       2) Engine down-speeding
       3) Advanced friction reduction measures
       4) Reduced parasitics
       5) Improved thermal management
       6) Use of a combination of both low- and high-pressure-loop cooled EGR
       7) Advanced turbocharging, including the use of VNT and sequential turbocharging
       8) Incorporation of highly-integrated exhaust catalyst systems with high NOx and PM
          removal efficiencies
       9) Adoption high-pressure common rail fuel injection systems with higher injection
          pressures and increased capability (i.e., multiple injections per firing cycle)
   The highest BMEP engines currently in mass-production for high-volume light-duty vehicle
applications are all diesel engines. MY2016-2017 light-duty diesel engines are available from
Honda, BMW and Mercedes Benz in the EU with approximately 26-bar to 29-bar BMEP and
peak cylinder pressures at or above 200-bar. 25>26>27  The light-duty diesel technology packages
used in the FRM analyses relied on engine data with peak BMEP in the range of 18 - 20 bar.
These were engine configurations using single-stage turbocharging with electronic wastegate
control, high-pressure or low-pressure (single-loop) cooled EGR, and common-rail fuel injection
with a 1800 bar peak pressure.  The cost analysis in the FRM for advanced light-duty diesel
vehicles assumed use of using a DOC+DPF+SCR system for meeting emissions standards for
criteria pollutants.

   In response to EPA Heavy Duty GHG emissions standards, large Class 8 heavy-duty truck
engine designs have exceeded 50  percent BTE.28'29  Despite their inherent differences, there now
appears to be a significant transfer of technology from heavy-duty diesel engines to much
smaller bore,  higher speed light-duty diesel engines underway, particularly for engines with high
BMEP. Use of CAE tools to design complex, stepped-geometry steel piston crowns and the use
of carefully designed piston oil-cooling galleries result in remarkably similar approaches when
comparing recent approaches to heavy-duty truck piston designs to recent light-duty diesel
engine piston designs such as that of the Mercedes-Benz OM654.28'30 The Mercedes-Benz
OM654 engine incorporates other design elements that are similar to current heavy-duty diesel
engine designs, including driving the camshaft and  some auxiliaries off of the rear of the engine,
the use of a high pressure common rail (HPCR) fuel injection systems with 2050 bar peak
pressure and the use of a VNT turbocharger.  BMW's B57 light-duty diesel engine used in the
MY2017 BMW 730d and 740d uses an HPCR fuel  injection system currently with 2500 bar peak
pressure and with capability to expand peak pressures to 3000-bar. Driving injection pressures
higher allows more flexibility for use of multiple injections and allows better optimization of
combustion phasing.  Modern, high BMEP light-duty diesel engines using conventional
diffusional combustion are capable of peak BTE of approximately 42 percent (see Figure 5.18).31
                                             5-37

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                                      20 kW
                                      10 kW
        1000   2000   3000   4000   5000  6000
                   Engine Speed (rpm)
                                              Si10
                                              i 8
                                                   350
                                                   300
                                                   250
                                                   200
                                                  — 150
                                                  !100
                                                    50
1000   1500   2000   2500   3000  3500  4000
            Engine Speed (rpm)
  Figure 5.18 Comparison Of BTE For A Downsized SI 2.0L 14 Miller Cycle Engine (Left)8 And A 1.7L 14
  Turbocharged Diesel Engine With HPCR, Low And High Pressure Loop Cegr, And VNT Turbocharger
                                        (Right)T.
Note: Green area shows region of high (35%) BTE.
   Advanced turbocharging and cooled EGR systems allow higher rates of EGR to be driven
and, when combined with more capable, higher pressure (2000-3000 bar) HPCR systems can
allow a degree of operation at light loads using pre-mixed charge compression ignition (PCCI) or
other low-temperature modes of combustion with inherently low NOx and PM emissions and
reduced thermal losses over a broader area of engine operation. Cummins "Light-duty Efficient,
Clean Combustion" engine development program for the U.S. DOE used mixed-mode, part-load
PCCI/high-load diffusional combustion approach and achieved a 20 percent improvement in
uncorrected city-cycle fuel economy (e.g., from 20.3 mpg to 24.5 mpg) when compared to a
more conventional diesel in a 5000 Ib. inertial test weight SUV at Tier 2, Bin 5 emissions levels.
Peak BTE for the PCCI combustion mode was approximately 46 percent compared with 42
percent peak BTE for conventional diffusional diesel combustion. Cummins developed a similar
dual-mode combustion approach as part of the Advanced Technology Powertrains for Light-
Duty (ATP-LD) and the Advanced Technology Light Automotive Systems (ATLAS) engine
development programs for the U.S. DOE.32'33  The engines developed as part of this program
combined dual-mode PCCI/diffusional combustion together with further improvements to the
turbocharger and charge air cooler systems, improved integration of the catalytic CDPF and
urea-SCR systems and addition of aNAC system for storage  of cold-start NOx emissions.
Developmental engines and emissions control  systems were integrated into Nissan Titan full-size
2-wheel-drive pickup trucks and achieved emissions consistent with Tier 3 Bin 30 compliance
and 21.8/34.3/26.0 City/Highway/Combined (uncorrected) fuel economy at a 5500 Ib. inertial
test weight. A similar engine used in the mid-size Nissan Frontier 4-wheel drive pickup at
reduced peak BMEP (21.3 bar vs. 23.4 bar in the Titan demonstration) achieved a 35 percent
s Adapted from Wurms et al. 2015.538
T Adapted From Busch Et Al. 2015.31
                                             5-38

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
combined cycle fuel economy improvement relative to the MY2015 4.0L PFI V6 Nissan
Frontier.34
5.2.2.12      Thermal Management
   Most recent turbocharged engine designs now use head-integrated, water-cooled exhaust
manifolds and coolant loops that separate the cooling circuits between the engine block and the
head/exhaust manifold(s) (Figure 5.19). Examples include the head-integrated exhaust
manifolds (IBM) and split-coolant loops used with the Ford l.OL 13, 1.5L 14, 2.0L 14 and 2.7L
V6 EcoBoost engines, the 2.0L VW EA888 engine, the GM EcoTec SGE l.OL 3-cylinder and
1.4L 4 cylinder engines, and the PSA 1.2L EB  PureTech Turbo.  The use of IEM and split-
coolant-loops is now also migrating to some naturally aspirated GDI and PFI engines, including
the GM 3.6L V6 LFX and EcoTec 1.5L engines and the l.OL 3-cylinder Toyota  1KR-FE
ESTEC.  These types of thermal management systems were included in the FRM analysis of
turbocharged GDI engines at BMEP levels  of 24-bar and above but were not considered for
turbocharged engines at lower BMEP levels or for naturally aspirated engines. Benefits include:

       •   Improved under-hood thermal management (reduced radiant heat-load)
       •   Reduced thermal gradients across the cylinder head
       •   Reduction in combustion chamber hot  spots  that can serve as pre-ignition sources
       •   Improved knock limited operation
       •   Reduce or eliminate enrichment required for component protection, particularly at
          low-speed/high-load conditions
          0  Enable additional engine "down-speeding" without encountering  enrichment
       •   Improved control of turbine inlet temperature (turbocharged engines only)
          0  Enable use of lower-cost materials  turbine and turbine housing materials
          0  Enable use of variable-geometry turbines similar to light-duty diesel applications
       •   Improved catalyst durability
       •   Shorter time to catalyst light-off after cold-start
       •   Improved coolant warmup after cold start
       •   Reduced noise
       •   Lower cost and parts count
          0  Improved durability (fewer gaskets to fail)
       •   Reduced weight (savings of approximately  1 kg/cylinder)
                                             5-39

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
         Figure 5.19 Exhaust Manifold Integrated Into a Single Casting with the Cylinder Head
5.2.2.13       Reduction of Friction and Other Mechanical Losses

   In urban driving, approximately 60 percent of engine losses are due to mechanical losses,
including engine friction.35 Piston and cylinder friction from the piston rings and piston skirts
account for 35 percent or more of engine friction in modern light-duty gasoline engines and
approximately 50 percent of engine friction in modern light-duty diesels engines.35'36'37 The
remaining frictional losses are primarily due to crankshaft, connecting rod, valvetrain and
balance shaft friction.  Piston skirt friction accounts for approximately 30 percent of piston
friction. Molybdenum disulfide (MoS2) and Diamond-like carbon (DLC) piston  skirt coatings
have demonstrated part-load engine friction reductions of approximately 16 percent and 20
percent, respectively.36 Improvements in cylinder bore surface treatments such as plasma
coatings26'27'38 and laser roughening39 have also been introduced in recent engine designs to
reduce engine friction and improve cylinder bore wear characteristics.

   Offsetting the crankshaft from the bore centerline, sometimes referred to as a desaxe cylinder
arrangement, can be used to reduce side forces on the piston and piston rings during the power
stroke, reducing friction piston/liner  friction  and reducing component wear.40 For example, the
2ZR-FXE engine used in the 2009-2015 Toyota Prius and the 2ZR-FE engine in the 2009-2016
Toyota Corolla have the  crankshaft centerline shifted 8 mm towards the intake side of the engine
to reduce friction.41

   Schaeffler has developed roller bearings that can be applied to the first and last crankshaft
main bearings without the added complexity of using built crankshafts or split main bearings to
reduce crankshaft friction and increase front journal load bearing capability when used with
higher power PO mild hybrid systems.  Roller bearing balance shafts for 3- and 4-cylinder
                                              5-40

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
engines have also been developed by Schaeffler, BMW and others that can reduce balance shaft
friction by approximately 50 percent.

   In addition to reducing engine mechanical losses, engine friction reduction also improves
engine restart when combined with stop/start systems. Reducing engine friction can also allow
additional engine downspeeding while maintaining idle and off-idle engine NVH characteristics.

   Hyundai and Delphi used a MY2011 2.4L 4-cylinder GDI engine to demonstrate a combined-
cycle fuel economy improvement of 4 percent by using a combination of a MoS2 piston skirt
coating, CrN physical vapor-deposition coated piston rings, low tension oil control rings and
engine downspeeding.42  They also achieved a further 2.9 percent combined-cycle fuel economy
improvement through use of a 2-stage variable displacement oil pump.

5.2.2.14      Potential Longer- Term Engine Technologies

   In addition to the engine technologies considered for this Draft TAR, and discussed above,
there are many other engine technology development efforts underway that may be fruitful in the
longer-term. While introduction of engines using these combustion concepts may occur prior to
2025, EPA and NHTSA do not expect significant penetration of these technologies into the light-
duty vehicle fleet in the 2022-2025 timeframe.

   Homogenous charge compression ignition (HCCI), gasoline compression ignition and other
dilute, low-temperature compression ignition gasoline combustion concepts are topics of
considerable automotive research  and development due to the potential for additional pumping
loss improvements at light and partial load conditions and reduced thermal losses.  Challenges
remain with respect to combustion control, combustion timing, and, in some cases, compliance
with Federal Tier 3 and California LEV3 NMOG+NOx standards.

   Engines using variable compression ratio (VCR) appear to be at a production-intent stage of
development, but also appear to be targeted primarily towards limited production, high
performance and very high BMEP (27-30 bar) applications. At lower BMEP levels, other
concepts (e.g., Atkinson Cycle for NA applications, Miller Cycle for boosted applications)
provide a similar means to vary effective compression ratio for knock mitigation with reduced
cost and complexity with some tradeoffs with respect to volumetric efficiency.

   One vehicle manufacturer recently entered production with a water injection system for knock
mitigation.  Injection of water and water/methanol or water/ethanol mixtures into the intake
systems of turbocharged and/or mechanically supercharged engines for knock mitigation is not a
new concept.  Aircraft engines predating World War II and some of the first turbocharged
automobile applications for the U.S. market in the 1960's used such systems for knock
mitigation.  Water injection systems compete with other means of knock mitigation (EGR,
Atkinson Cycle, Miller Cycle, and lEM/split-cooling) that do not require fluid replenishment.
Current and near term applications appear to be limited to low-volume production,  high
performance vehicles.

   The DOE Co-Optimization of Fuels and Engines (Co-Optima) initiative aims to improve
near-term efficiency of spark-ignition (SI) and compression ignition engines through the
identification of fuel properties and design parameters of existing base engines that maximize
performance.
                                             5-41

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   According to DOE, Co-Optima is a first-of-its-kind effort brings together multiple DOE
offices, national laboratories, and industry stakeholders to simultaneously conduct tandem fuel
and engine R&D and deployment assessment in order to maximize energy savings and on-road
vehicle performance, while also reducing long-term transportation-related petroleum
consumption and GHG emissions. Two parallel research tracks focus on: 1) improving near-term
efficiency of spark-ignition (SI) engines through the identification of fuel properties and design
parameters of existing base engines that maximize performance.  The efficiency target represents
a 15% fuel economy improvement over state-of-the-art, future light-duty SI engines with a
market introduction target of 2025; and 2) simultaneous testing of new fuels with existing CI
engines (as well as advanced compression ignition [ACI] combustion technologies as they are
developed) to enable a longer-term, higher-impact series of synergistic solutions. The fuel
economy target represents a 20% improvement over state-of-the-art, future light-duty SI engines
with a market introduction target of 2030. By using low-carbon fuels, such as biofuels, GHGs
and petroleum consumption can be further reduced. EPA and NHTSA will continue to closely
follow the Co-Optima program and provide input to DOE, including through EPA's technical
representative on the Co-Optima External Advisory Board, as this program has the potential to
provide meaningful data and ideas for GHG and fuel consumption reductions in the light-duty
vehicle fleet for 2026 and beyond.

5.2.3  Transmissions: State of Technology

5.2.3.1 Background

   The function of a transmission system is to reduce the relatively high engine speed and
increase the torque, so that the power output of the engine can be coupled to the wheels. The
complete drivetrain includes a differential (integral to the transmission on front-wheel-drive
vehicles; separate on rear-wheel-drive vehicles) which provides further speed reduction, and
often a hydraulic torque converter which provides significant torque multiplication at low speed
conditions. The complete drivetrain - torque converter, transmission,  and differential - is
designed as a set to best match the power available from the engine to that required to propel the
vehicle.
                                             5-42

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
 100%n
  75% -
  50%-
  25%-
      1980   1985  1990   1995   2000   2005  2010  2015  1980  1985   1990   1995   2000  2005  2010   2015
                        Car
Truck
                                             Transmission
                                             Continuously Variable
                                             Manual
                                             Other
                                                        Lockup?
                                                              Number of Gear:
                Figure 5.20 Transmission Technology Production Share, 1980 - 201443

   Different transmission architectures are available for use in light duty vehicles. Conventional
automatic transmissions (ATs) are the most popular type, and still dominate the light-duty fleet,
as seen in Figure 5.20. Manual transmissions (MTs), although less popular than in the past, are
also still part of the fleet.  Both ATs and MTs have, among other improvements, seen an increase
in the number of gears employed. Figure 5.20 shows the recent gains in six, seven, and eight-
speed transmissions in both the car and light truck segment. Two other transmission types have
also seen an increase in market share. These are dual-clutch transmissions (DCTs), which have
significantly lower parasitic losses than ATs, and continuously variable transmissions (CVTs),
which can vary their ratio to target any place within their overall spread. Each of these four
types of transmissions is discussed in more detail in the sections below.

5.2.3.2 Transmissions: Summary of State of Technology and Changes since the FRM

   In the analysis conducted for the 2017-2025MY FRM, the agencies estimated that DCT
transmissions would be very effective in reducing fuel consumption and CCh emissions, less
expensive than current automatic transmissions, and thus a highly likely pathway used by
manufacturers to comply with the regulation. However, DCTs thus far, have been used in only a
small portion of the  fleet as some OEMs have reported in meetings with the agencies have
indicated and some vehicle owners have cited drivability concerns for DCT.44  On the other
hand, the 2017-2025MY FRM analysis also  predicted a low effectiveness associated with CVTs
(due to the high internal losses and small ratio spans of CVTs in the fleet at that time), and thus
CVTs were not included in the FRM fleet modeling.  However, internal losses in current CVTs
have been much reduced and ratio spans have increased from their predecessors, leading to
increased effectiveness and further adoption rates in the fleet, particularly in the smaller car
segments. The new CVT's also tend to give  the best effectiveness for their cost.

   Again in the 2017-2025MY FRM, the agencies estimated that step transmissions with higher
numbers of gears (e.g., AT8s) would be slowly phased into the fleet.  However, AT8s have been
                                              5-43

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
"pulled ahead," appearing in substantial numbers even before 2015MY.  In addition,
manufacturers have introduced (and/or have plans to introduce) transmissions with even higher
numbers of gears (e.g., AT9s and AT 10s), a technology that was not considered in the 2017-
2025MY FRM.

   Thus, as highlights of transmission technology analysis in this Draft TAR, (a) the technology
packages and vehicle classes where DCTs are applicable have been re-evaluated to reflect
manufacturer's current choices, (b) the effectiveness of CVTs has been re-examined and
increased to reflect current vintage CVTs and their use in the fleet, and (c) nine and ten-speed
transmissions were considered (since they are or will be in the fleet) when determining the
effectiveness of future transmissions in the fleet.

5.2.3.3 Sources of Transmission Effectiveness Data

   In addition to the sources of transmission effectiveness data cited in the 2017-2025 LD FRM,
the agencies also used data from a wider range of available sources to update and refine
transmission effectiveness for this analysis. These sources included:

       1)  Peer-reviewed journals, peer-reviewed technical papers, and conference proceedings
          presenting research and development findings
       2)  Data obtained from transmission and vehicle testing programs, carried out at EPA-
          NVFEL, ANL, and other contract laboratories
       3)  Modeling results from simulation of current and future transmission configurations
       4)  Confidential data obtained from OEMs and suppliers on transmission efficiency
   For transmission testing programs, EPA contracted with FEV Engine Technologies to test
specific transmissions in a transmission component test stand.  The testing program was
primarily designed to determine transmission efficiency and torque loss over a range of input
speeds, input loads, and temperatures.  In addition, other driveline parameters, such as
transmission rotational inertia and torque converter K-factor were characterized. Two automatic
transmissions have been characterized in this test program, which is still on-going.  Torque loss
maps were generated for both a six-speed 6T40 GM automatic transmission and an eight-speed
845RE FCA automatic transmission (see Figure 5.21).
                                             5-44

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                        12
                        10
                      £
                      •z.
	 Gear 1
— Gear 3
	 GearB
	 Gear 7
— Gear 2
Gear 4
	 Gear 6
	 GearS
                                  50
                                         100      150
                                         Input Torque (Nm)
                                                         200
                                                                 250
                      100%
                      95%
                     •3 90%
                                  50
                                         100     150
                                         Input Torque (Nm)
                                                         200
                                                                 250
 Figure 5.21 Average Torque Losses (Left) And Efficiency (Right) In Each Gear For An Eight-Speed 845RE
Transmission From A Ram, Tested At 100 °C And With Line Pressures Matching Those Measured In-Use In
The Vehicle. Torque Losses Were Averaged Over 1000 Rpm - 2500 Rpm. This Transmission Is A Clone of the
                                        ZF 8HP45.

   In addition to contracting to test specific transmission, EPA has obtained torque loss maps
and/or operational strategies for current generation transmissions from manufacturers and
suppliers. These maps are CBI, but have been used to inform EPA on the effectiveness of
transmissions currently on the market. Maps obtained from manufacturers and suppliers include
examples of both CVTs and DCTs.

   To characterize transmission and torque converter operation strategies, EPA has also
performed multiple chassis dynamometer tests of current-generation vehicles equipped  with a
range of transmission technologies. The transmission gear and torque converter state (as well as
other vehicle parameters) were recorded over the FTP, HWFET, and US06  cycles.  The recorded
data were used to determine the drive strategy for the  engine-transmission pair in the vehicle.

   The transmission losses and shifting strategy were  used as modeling inputs to EPA's full-
vehicle ALPHA model.45 The shifting strategy was parameterized to allow sufficient flexibility
to maintain reasonable shift strategies while changing other vehicle attributes.46

   EPA also performed a study using chassis dynamometer testing to determine effectiveness  of
transmissions. In particular, two Dodge Chargers, one with a five-speed transmission and one
with an eight-speed transmission, were tested on the dynamometer.  Other than the transmission,
these vehicles had identical powertrains, and so provided an ideal opportunity to test the effect of
different transmissions in the vehicle.47 Multiple repetitions of the FTP and HWFET, cycles
                                              5-45

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
were run, with the result that the Charger equipped with the eight-speed transmission exhibited
on average a 6.5 percent reduction in fuel consumption over the five-speed Charger on the
combined FTP/HWFET cycle.  The eight-speed Charger also exhibited an increase in
acceleration performance, according to tests by Car and Driver, with, for example, a 0.5 second
improvement in 0-60 time.48'49

   NHTSA has leveraged work performed over the past 15 years by Argonne National
Laboratory with Autonomie under funding from the U.S. Department of Energy.  Leveraging
vehicle test data for a large number of vehicles measured at Argonne's Advanced Powertrain
Research Facility (APRF), shifting algorithms were developed and validated for multiple
transmission technologies (i.e., automatic, CVT, DCT) and gear number (i.e., 6 vs 8 speeds).50
Detailed instrumentation was also critical in developing component models and controls for
advanced transmissions such as Dual Clutch.51 While specific transmission gear ratios and
shifting algorithms were used during the validation process, a different approach was used to
design the transmission gear ratios to properly quantify the effectiveness of the technology.
Argonne used an algorithm published by Naunheimer along with a range of constraints to design
their transmission gear ratios.52 A set of efficiencies for each gear was selected to represent
today's leading technologies across all transmission types to ensure proper comparison.
Calibration of the shifting algorithms was performed within a set of constraints to ensure proper
driving quality. The constraints were defined based on vehicle test data.

5.2.3.4 Sources of GHG Emission Improvements: Reduction in Parasitic Losses, Engine
Operation, and Powertrain System Design

   The design of the transmission system can affect vehicle GHG emissions in two ways.  First,
reducing the energy losses within the transmission (and/or torque converter) reduces the energy
required from the engine, which also reduces GHG emissions. Reducing transmission losses can
be accomplished by increasing gearing efficiency, reducing parasitic losses, altering the torque
converter lockup strategy, or other means. A more in-depth discussion of internal energy loss
reduction is included in the "Transmission Parasitic Losses" and "Torque  Converter Losses and
Lockup Strategy" sections below.

   Another method to decrease GHG emissions is to design the entire powertrain system - the
engine and transmission - to keep the engine operating at the highest available efficiency for as
much time as possible. Transmissions with more available gears (or, at the extreme,
continuously  variable transmissions) can maintain engine operation within a tighter window, and
thus maintain operation nearer the highest efficiency areas of the engine map. Likewise,
transmissions with a wider ratio spread can maintain engine operation nearer the  highest
efficiency areas of the engine map for a wider range of vehicle speeds,  in  particular lowering the
engine speed  at highway cruise for reduced GHG emissions.

   In addition, the highest engine efficiencies for a given power output tend to be at lower
speeds, so transmission control strategies that allow very low engine  speeds (i.e.,
"downspeeding") also reduce GHG emissions. Shifting strategies are discussed in the
"Transmission Shift Strategies" section below.

   As a  practical matter, transmissions with an increased number of gears tend also to have a
wider ratio. For example, the ZF 8HP eight-speed RWD transmission has a spread of 7.07,53 the
Aisin eight-speed FWD transmission has a spread of 7.58,54 the Mercedes 9G-TRONIC nine-
                                             5-46

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
speed transmission has a ratio spread of 9.15,55 and the ZF 9HP48 nine-speed FWD transmission
has a spread of 9.8.56

   The effects of additional gears and a wider ratio can be seen in Figure 5.23, which compares
engine operation of the same engine when coupled with a six-speed transmission and with an
eight-speed transmission. Compared to the six-speed transmission, the eight-speed transmission
allows the engine to operate over a narrower speed range and at lower speeds, both of which tend
to reduce GHG emissions.
                     Q.
                     HI

                     CQ
                            600  800  1000  1200  1400  1600  1800  2000
                                              Engine Speed (RPM)
                            600  800  1000
1200  1400  1600  1800  2000
   Engine Speed (RPM)
              (a) Six-speed
                   (b) Eight-speed
     Figure 5.22 Engine Operating Conditions for Six-Speed (Left) and Eight-Speed (Right) Automatic
                           Transmissions on the FTP-75 Drive Cycle57

   The dominant trends in transmissions have been toward a larger number of gears and a wider
ratio spread. However, it is recognized, including by the 2015 NAS Report, that above certain
values, additional gearing and ratio spread provide minimal additional fuel economy benefits.58
59 eo jhu^ increasing the number of gears (except when going to effectively infinite the case of
CVT transmissions) and ratio spread beyond that exhibited by the current market leaders is
unlikely to result in significant fuel consumption benefits, although other vehicle attributes such
as acceleration performance and shift smoothness  may benefit.

   In fact, it is well-understood that typical implementations of high-gear transmissions provide
both fuel consumption and acceleration performance benefits. Performance benefits come from
                                              5-47

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
two factors: first, the gear ratio spread of transmissions with higher number of gears will
typically "straddle" the ratio spread of the lower number of gear transmission they replace (i.e.,
first gear is a numerically higher ratio and the final gear is a numerically lower ratio). This
provides more launch torque and quicker acceleration from stop.  Second, the gear ratios of
sequential gears tend to be closer together in transmissions with a higher number of gears.  This
not only narrows the on-cycle operation range of the engine for improved fuel economy (as in
Figure 5.23), but also maintains engine performance nearer the maximum power point in high
power demand situations for better acceleration performance at higher vehicle speeds.

   To determine the relative cost-effectiveness of different technologies, it is important to
account for all technology benefits where possible. As the NAS point out, "objective
comparisons of the cost-effectiveness of different technologies for reducing FC can be made
only when vehicle performance remains equivalent."61 This is particularly relevant for advanced
transmissions, which do affect performance when coupled with the same engine as transmissions
with a lower number of gears. In evaluating information on measured or modeled fuel
consumption effects of advanced transmissions, it is important to consider both reported fuel
consumption benefits and any simultaneous acceleration performance benefits, so that
transmission effectiveness can be objectively and fairly estimated.

   Transmission design parameters that substantially affect engine operation - gearing ratios,
ratio spread, and shift control  strategy - are all used to optimize the engine operation point, and
thus the effectiveness of these transmission parameters depend in large part on the engine it is
coupled with.  Advanced engines incorporate new technologies, such as variable valve timing
and lift, direct injection, and turbocharging and downsizing, which improve overall fuel
consumption and broaden the  area of high-efficiency operation. With these more advanced
engines, the benefits of increasing the number of transmission gears (or using a continually
variable transmission) diminish as the efficiency remains relatively constant over a wider area of
engine operation.  For example, the NAS estimated that the benefit of an eight-speed
transmission over a six-speed  transmission is reduced by approximately 15 percent when added
to a modestly turbocharged, downsized engine instead of a naturally  aspirated engine.62 Thus,
the effectiveness of transmission speeds, ratio, and shifting strategy should not be considered as
an independent technology, but rather as part of a complete powertrain.

   Additionally, because the engine and transmission are paired in the powertrain, the most
effective design for the engine-transmission pair is where the entire powertrain is running at the
highest combined efficiency.  This most effective point may not be at the highest engine
efficiency, because a slightly different operation point may have higher transmission efficiency,
leading to the best combined efficiency of the entire powertrain.

5.2.3.5 Automatic Transmissions  (ATs)

   Conventional planetary automatic transmissions remain the most numerous type of
transmission in the light duty fleet. These transmissions will typically contain at least three or
four planetary gear sets, which are connected to provide the various gear ratios. Gear ratios are
selected by activating solenoids which engage or release multiple clutches and brakes. A
cutaway of a modern RWD transmission (in this case the ZF 8FIP70) is shown in Figure 5.23.
                                              5-48

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                        Figure 5.23 ZF 8HP70 Automatic Transmission63

   Automatic transmissions are packaged with torque converters which provide a fluid coupling
between the engine and the driveline, and provide a significant increase in launch torque. When
transmitting torque through this fluid coupling, energy is lost due to the churning fluid. These
losses can be eliminated by engaging ("locking up") the torque converter clutch to directly
connect the engine and transmission. A discussion of torque converter lockup is continued in the
"Torque Converter Losses and Lockup Strategy" section below.

   In general, ATs with a greater number of forward gears (and the complementary larger ratio
spread) offer more potential for CCh emission reduction, but at the expense of higher control
complexity.  Transmissions with a higher number of gears offer a wider speed ratio and more
opportunity to operate the engine near its most efficient point (as shown in the previous section).

   In the past few years, manufacturers have taken advantage of this fact. Four- and five-speed
automatic transmissions, which dominated the market in 2005, have substantially declined in
number, being replaced by six-speed and higher transmissions (see Figure 5.20 above). In fact,
the average number of AT gears in the fleet has rapidly increased, and in 2014 was above six for
both cars and trucks (see Figure 5.24 below).
                                             5-49

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                             Car
Truck
             6 —
        O    5-
        |
             4 —
             3-
                    Manual
               1980 1985 1690 1995 2000 2005 2010 2015 1990 1985 1990 1995 2000 2005 2010 2015

                                         Model Year

        Figure 5.24 Average Number of Transmission Gears for New Vehicles (excluding CVTs)64

   As six-speed ATs have supplanted the four-and five-speeds, seven- and eight-speed
transmissions have also appeared on the market. In the FRM, eight speed ATs were not expected
to be available in any significant number until approximately 2020. However, even as of 2014
seven- and eight-speed transmissions occupy a significant and increasing portion of the market.

   Seven-speed transmissions currently available include the RWD 7G-Tronic from Mercedes
and the JATCO JR710E available in Nissan products. RWD eight-speed transmissions available
include offerings from General Motors and Hyundai, as well as transmission suppliers Aisin and
ZF. The ZF 8HP, introduced in 2009, has been incorporated into offerings from a range of
manufacturers, including Fiat/Chrysler, Jaguar/Land Rover, and Volkswagen. ZF has begun
production of a  second generation of 8HP transmissions (the 8HP50), which features a higher
ratio spread, lower drag torque, and improved torsional vibration absorption compared to the first
generation.65 Aisin also offers a FWD eight-speed used by multiple manufacturers. This
includes use in the compact 2016 Mini Cooper Clubman,66 a vehicle smaller than those assumed
eligible for eight-speed transmissions in the FRM.

   In the FRM, the agencies limited their consideration of the effect of additional gears to eight-
speed transmissions. However, some ATs with more than eight gears are already in production,
and more examples are in development.  At this time,  nine-speed transmissions are being
manufactured by ZF67 (which produces a FWD nine-speed incorporated into Fiat/Chrysler,
Honda, and Jaguar/Land Rover vehicles68) and Mercedes69 (which produces a RWD nine-speed).
In addition, Ford and General Motors have announced plans to jointly design and build nine-
speed FWD transmissions and ten-speed RWD transmissions (2017 F150 and 2017 Camaro
ZL1), and Honda is developing a ten-speed FWD transmission.70

   Manufacturers have claimed substantial fuel consumption benefits associated with newer
transmissions. ZF claims its first generation 8HP can reduce fuel consumption by 6 percent on
                                             5-50

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
the NEDC compared to a circa 2005 ZF 6HP, using the same engine, along with improving
vehicle acceleration performance.71 ZF also outlined a series of potential improvements to the
first generation 8FIP that could provide an additional 5 to 6 percent fuel consumption reduction
on the U.S. combined cycle.72 The second generation ZF eight-speed73 is expected to achieve up
to 3 percent efficiency gain on the NEDC due to the improvements noted above; ZF also outlined
additional potential savings associated with a third generation eight-speed transmission.74
Likewise, Mercedes clamed a 6.5 percent fuel consumption improvement on the NEDC with its
nine-speed transmission compared to the previous seven-speed.75 It should also be noted that the
percent fuel consumption reported on the NEDC drive cycle will be different from the U.S.
combined cycles.

   In FWD vehicles, ZF claims its nine-speed FWD transmission reduces fuel  consumption by
10 percent - 16 percent compared to an early- 2000s six-speed transmission.76  Aisin claims its
new FWD eight-speed transmission decreases fuel consumption 16.5 percent compared to an
early generation six-speed, and nearly 10 percent compared to the previous generation six-
speed.77  In addition, the new eight-speed improves acceleration performance.   BMW, using the
Aisin FWD transmission, reports a 14 percent fuel consumption reduction on the NEDC over the
previous six-speed transmission.78

   These efficiency improvements are due to a range of design changes in the transmissions. In
addition to improving the engine operation efficiency through changing the number of gears,
overall ratio, and shift points, these transmissions also reduce parasitic losses, change torque
converter behavior, and/or shift to neutral during idle. Mercedes claims a total of 6.5 percent
fuel economy improvement on the NEDC by using its nine-speed 9G-TRONIC in place of the
earlier generation seven-speed.79 Of this, 2 percent is due to the change in the number of gears,
ratio spread, and shift strategy, with the remainder due to transmission efficiency improvements.

   With the positive consumer acceptance, higher effectiveness, and increasing production of
transmissions with up to ten forward gears,  it may be possible that transmissions with even more
gears will be designed and built before  2025. Researchers from General Motors have authored a
study showing that there is some benefit to be gained from transmissions containing up to 10
speeds.80 However this appears to be near the limit for improved fuel consumption, and studies
have shown that there is no added potential  for reduction in CO2 emissions beyond nine or ten
gears.8182 In fact, ZF CEO Stefan Sommer has stated that ZF would not design transmissions
with more than nine gears: "We came to a limit where we couldn't gain any higher ratios. So the
increase in fuel efficiency is very limited and almost eaten up by adding some weight and
friction and even size of the transmission."83 Although manufacturers may continue to add gears
in response to consumer preference for other performance attributes, it is unlikely that further
increases will provide CO2 emissions benefits beyond that of optimized eight, nine or ten-speeds.

5.2.3.6 Manual Transmissions (MTs)

   In a manual transmission, gear pairs along an output shaft and parallel layshaft are always
engaged. Gears are  selected via a shift lever, operated by the driver. The lever operates
synchronizers, which speed match the output shaft and the selected gear before engaging the gear
with the shaft.  During shifting operations (and during idle) a clutch between the engine and
transmission is disengaged to decouple engine output from the transmission.
                                             5-51

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Manual transmissions are in general lighter, cheaper to manufacture, and have lower parasitic
losses than automatic transmissions.  The 2015 NAS report found the overall energy loss in a
manual transmission to be only about 4 percent, as compared to a 13 percent loss in automatic
transmissions.84

   As with ATs, the average number of gears in MTs has increased (Figure 5.24), albeit at a
reduced rate compared to ATs. As in ATs, the higher number of gears and associated increase in
ratio spread increases potential fuel savings.

   However, manual transmissions have only a small market share, estimated at only 3.7 percent
in 2014.85 Automatic transmissions (ATs, CVTs, and DCTs) are more popular at least in part
because customers prefer not to manually select gears.

5.2.3.7 Dual Clutch Transmissions (DCTs)

   Dual clutch transmissions are similar in their basic construction to manual transmissions, but
use two coaxial input shafts with two clutches to shift between the two shafts.  By
simultaneously opening one clutch and closing the other, the DCT "hands off power from one
shaft to the other,  and thus to sequential gears.  Unlike  the MT, the DCT selects the appropriate
gear automatically (as in an AT). DCTs offer an efficiency advantage over a typical automatic
because their parasitic losses are significantly lower. In addition, DCTs in general do not require
a torque converter, as gradually engaging the clutch (much like with a manual transmission)
provides the application of launch torque.
                     r
                                                                   transmission-
                                                                   output
                     I
                     I
                     I	L
                        dual-clutch
transmission
                        Figure 5.25 Generic Dual Clutch Transmission86

   Multiple DCTs have been introduced into the marketplace, primarily in six- and seven-speed
versions. Volkswagen has used multiple generations of DCTs in their products. Ford has used
six-speed DCTs jointly developed with Getrag. Fiat has another version of a six-speed DCT,
while both Honda and Hyundai have developed seven-speed versions. Honda introduced an
eight-speed DCT with a torque converter on the 2015 Acura TLX.87
                                             5-52

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   However, DCTs have encountered issues with customer acceptance, and, as the NAS stated in
its 2015 report, "are not likely to reach the high penetration rates predicted by EPA/NHTSA ...
primarily due to customer acceptance issues."88 As noted by the NAS in their 2015 report, "This
difference in drivability and consumer acceptance [between wet and dry clutch DCTs] can be
seen in the comparison of two of Volkswagen's MY2015 vehicles, the VW Golf and the VW
Polo. The Golf, with a wet-clutch DCT, has received many positive reviews and awards, while
the Polo, with a dry-clutch DCT, has received poor reviews for transmission-related
drivability."89

   Getrag  announced the 7DCT300 which has a wet clutch with lubrication on demand (we refer
to these as damp clutch DCTs), equaling the efficiency of a dry DCT.  The "damp" clutch is also
smaller and has a higher tolerance for engine irregularities.90 Wet/damp clutch DCTs tend to
have better consumer acceptance than dry clutch DCTs.  The 7DCT300 is available in Europe on
the 2015 Renault Espace.

   As in ATs, it is expected that additional gears above the current maximum will not
significantly decrease fuel consumption and resulting GHG emissions.  A 2012 study by DCT
manufacturer Getrag indicated that additional gears above seven and additional ratio spread
above 8.5 provided minimal additional fuel economy benefits.91

5.2.3.8 Continuously Variable Transmissions (CVTs)

   Conventional continuously variable transmissions consist of two cone-shaped pulleys,
connected with a belt or chain.  Moving the pulley  halves allows the belt to ride inward or
outward radially on each pulley, effectively changing the speed ratio between the pulleys. This
ratio change is smooth and continuous, unlike the step changes of other transmission varieties.
CVTs were not chosen in the  fleet modeling for the 2017-2025MY FRM analysis because of the
predicted a low effectiveness  associated with CVTs (due to the high internal losses and narrow
ratio spans of CVTs in the fleet at that time).  However, improvements in CVTs in the current
fleet have increased their effectiveness, leading to rapid adoption rates in the fleet. In their 2015
report, the NAS recommended CVTs be added to the list of considered technologies, and the
agencies are re-evaluating the cost and effectiveness numbers for this technology.
                                             5-53

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                   (a)
(b)
                     Figure 5.26 (a) Toyota CVT92 (b) Generic CVT sketch9
   One advantage of CVT's is that CVT's continue to transmit torque during ratio changes.
During a ratio change or shift the energy from the engine is wasted on ATs and some DCTs.
ATs and some DCT have a hesitation during shifts caused by the torque disruption during gear
changes. This shift feeling is well known to consumers and in some cases comforting to drivers
(they miss it when driving a vehicle with a CVT). As mentioned in the AT section ATs
efficiency peaks with 9 to 10 gears, while going to a CVT (with an effectively "infinite" number
of gear steps) adds a new level of efficiency to the overall system.  This is in part due to the fact
that CVTs do not need to stop transmitting torque to change ratios.

   Another advantage of a CVT is that, within its ratio range, it can maintain engine operation
close to the maximum efficiency for the required power. However, CVTs were not considered in
the FRM because at the  time CVTs had a ratio range of near 4.0, limiting the range where the
engine operation could be optimized.  In addition, the CVTs were less than 80 percent efficient
94, and thus required more total output energy from the engine. These limitations overwhelmed
the CVT's inherent advantage compared to conventional ATs.

   However, in the recent past, manufacturers and suppliers have intensified development of
CVTs, reducing the parasitic losses and increasing the ratio spread. The current generation of
CVT are now nearly 85  percent efficient, with ongoing work by suppliers to push that number to
90 percent.95 Ratio spreads for new CVTs from Honda, Toyota, and JATCO now range between
6.0 and 7.0.96 97 98 JATCO has introduced a very small CVT what has a two speed output with
                                             5-54

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
take a CVT with a small ratio spread and doubles it for an overall ratio spread of 7.3" in the base
version and 8.7 in the "wide range" version.100 As in ATs and DCTs, it is expected that
additional increase in ratio range above the current maximum will not significantly decrease fuel
consumption and resulting GHG emissions.101

   Reducing losses in CVTs has been a particular focus of manufacturers. The JATCO CVT8
featured a 40 percent reduction in mechanical losses compared to their earlier generation CVT.102
The losses were reduced by decreasing the size of the oil pump, implementing a new, higher
efficiency belt, and reducing the fluid churning losses. Honda's new compact car CVT increased
efficiency 1.0 percent to 1.5 percent at higher vehicle speeds compared to their previous
generation CVT.103 The increased efficiency was primarily due to a reduction in oil pump losses
and bearing friction.  Honda's new midsize CVT increased efficiency up to 5 percent compared to
the earlier generation CVT, primarily by reducing the required hydraulic pressure (by up to 38
percent).104 Toyota's new Kl 14 CVT reduced torque losses by 22 percent, compared to the
earlier generation of CVTs, primarily by reducing the losses associated with the oil pump, and
reducing the size of the bearings.105

   The decreased transmission losses (5-10 percent) and increased ratio spread (from 4 to
between 6 and 8.7) of CVTs has made them more effective in CCh reduction than estimated in
the FRM, and thus CVTs are anticipated to be used in an increasing share of the fleet (see Figure
5.20). The supplier JATCO supplies CVTs to Nissan, Chrysler, GM, Mitsubishi, and Suzuki 106
In addition, other manufacturers' - Audi, Honda, Hyundai, Subaru, and Toyota — all make their
own CVTs.

   The JATCO CVTS demonstrated a 10 percent improvement in fuel economy for both the
highway and city cycles compared to earlier generation CVTs.107 Honda's new compact car
CVT increased fuel economy approximately 7 percent compared to the earlier generation CVT
over both the U.S. test cycle and the Japanese JC08 test cycle.108 Honda's new midsize CVT
increased fuel economy 10 percent over the earlier generation SAT on the U.S. cycle, and 5
percent compared to the earlier generation CVT on the Japanese JC08 test cycle.109 Toyota's
new Kl 14 CVT increased fuel economy by 17 percent on the Japanese JC08 test cycle compared
to the earlier generation CVT. no

   Initial introductions of CVTs suffered from consumer acceptance issues, where customers
complained of the "rubber band" feel of the transmission, due to the indirect connection between
the driver's throttle input and the vehicle's acceleration response.  To combat this perception,
vehicle manufacturers have added a shift feel calibration to the CVT control strategy, which
mimics the feel of a conventional AT.111  This calibration, although having a slight effect on fuel
economy, has improved consumer acceptance.112

   In this document, only conventional belt or chain CVTs are considered. At least two other
technologies - toroidal CVTs and Dana's VariGlide® technology113 - are under development and
may be available in the 2020-2025 timeframe. The Dana VariGlide is considered a CVP
(Continuously Variable Planetary) with the major design difference being it using balls to
transmit torque and vary the ratio. Dana has stated that it is currently in development with an
OEM. Targeted production could be as early as 2020. These technologies hold promise for
increased efficiency compared to current design belt or chain CVTs.
                                             5-55

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.2.3.9 Transmission Parasitic Losses

   Reducing parasitic loses in the transmission improves drivetrain efficiency and lowers the
required energy output from the engine. In general, parasitic losses can come from (a) the oil
supply, (b) electricity requirements, (c) drag torque, (d) gearing efficiency, and (e) creep (idle)
torque.114

5.2.3.9.1      Losses inATs

   A study by ZF suggests that the largest sources of losses over the combined city/highway
cycle in conventional automatic transmissions are the oil supply and the drag torque.115  This is
followed by the creep torque (on the city cycle), with the electrical requirements and gearing
efficiency being relatively minor.

   For conventional ATs, power required to supply oil to the transmission is one of the largest
sources of parasitic loss. An oil pump is required for lubrication and for hydraulic pressure for
clamping the clutches. A baseline transmission would typically use a gerotor-type pump driven
off the torque converter. Replacing or resizing the oil pump can result in a substantial decrease in
torque losses. For example, Aisin claims a 33 percent reduction in torque loss in its new
generation transmission from optimizing the oil pump,116 and Mercedes claims a 2.7 percent
increase in fuel  economy on the NEDC by changing the pumping system.117 Pump-related losses
can be reduced by substituting a more efficient vane pump for the gerotor. Losses can be further
reduced with a variable-displacement vane pump, and by reducing the pressure of the system.
Losses can be further decreased by using an on-demand electric pump: Mercedes claims an
additional 0.8 percent increase in fuel economy on the NEDC by implementing a lubrication on
demand system.118  Another way to reduce losses from the pump is by reducing leakage in the
system.  Reducing leakage reduces parasitic losses by reducing the amount of fluid that needs to
be pumped through the system to maintain the needed  pressure.

   A second large source of parasitic loss in ATs is the drag torque in the transmission from the
clutches, brakes, bearings, and seals. These components have the potential to be redesigned for
lower frictional losses. New clutch designs offer potential reductions in clutch drag, promising
up to a 90 percent reduction in drag.119 Replacing bearings can reduce the associated friction by
50 to 75 percent. New low-friction seals for can  reduce friction by 50 percent to provide an
overall reduction in bearing friction loss of approximately 10 percent.120

   Optimizing shift elements improved fuel economy on the Mercedes 9G-TRONIC by  1 percent
over the NEDC.121

   Drag torque can be further reduced by decreasing the viscosity of the automatic transmission
fluid used to lubricate the transmission.  A study  of transmission losses indicate that about a 2
percent fuel consumption reduction was obtained on the FTP 75 cycle by switching to the lowest
viscosity oil.122  However, reduction of transmission fluid viscosity may have an adverse effect
on long-term reliability.

   Transmission efficiency may also be improved through superfinishing the gear teeth to
improve meshing efficiency.
5.2.3.9.2     Losses in DCTs
                                             5-56

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Advanced DCTs typically have lower losses than ATs, largely due to having an on-demand
pump, splash lubrication, and fewer open clutches.  The primary losses in DCTs are load-
independent drag and splash losses. Unlike ATs, DCTs typically depend on splash lubrication for
their internal components rather than forced lubrication.  This eliminates the losses associated
with oil supply pumps, but adds churning losses due to rotating components moving through the
oil. Churning losses can be minimized by keeping oil levels low and warming up the lubrication
oil.

   A primary consideration in DCT losses is the use of wet or dry clutches.123 Dry clutches do
not require oil cooling flow, and therefore do not contribute to oil churning losses that are
incurred with wet clutch systems; this has traditionally meant that dry clutch reduced GHG
emissions by an additional  0.5 to  1 percent over wet clutch DCTs.  However, dry clutches have a
limited maximum torque capacity, and have suffered from customer acceptance issues. In
response, so-called "damp" clutches have been introduced, where on-demand cooling flow has
substantially reduced the parasitic losses associated with wet clutches.

   DCTs also may benefit from the same improvements in bearing and seal drag and gear
finishing that are outlined in the AT section above.

5.2.3.9.3     Losses in CVTs

   CVTs tend to have higher losses than either ATs or DCTs, in large part due to the high oil
pressures required to keep the belt and pulleys securely clamped.  These losses increase
significantly at high input torques, as even higher pressures are required to maintain the
clamping force.124

   A study by JATCO suggests that losses in the CVT are dominated by oil pump torque and
losses in the belt-pulley system, with fluid churning losses as the next largest player.125 By
reducing leakage in the  oil  system and reducing line pressure when possible, JATCO's CVTS
was able to run with a reduced  size oil pump and considerable  reduction in oil pump torque loss.
JATCO also redesigned the belt for lower loss, and reduced the oil level and viscosity to reduce
churning losses.  The overall result was a 40 percent reduction in mechanical losses compared to
the earlier generation CVT.

   Honda developed a new CVT using a comparable strategy.126 They decreased the required
pulley thrust by refining the control strategy and by using a fluid with increased coefficient of
friction, which combined for a transmission efficiency increase of 2.8 percent.  They also altered
the belt trajectory around the pulley for an added 0.4 percent efficiency increase.

Another opportunity for reduced losses in CVT's is Dana's VariGlide System.  Dana's VariGlide
system can provide more favorable system losses than traditional belt or chain technologies.  The
VariGlide system eliminates the requirement for a high pressure pump, using instead a fully
passive mechanical clamping mechanism.  The unique coaxial configuration, similar to a
planetary gearset coupled with  high power density, allows for simple integration into traditional
transmission architectures and makes it uniquely suited for RWD applications.
5.2.3.9.4     Neutral Idle Decoupling
                                             5-57

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   An additional technology that has been implemented in some transmissions, which was not
considered in the FRM, is the application of a "neutral idle."  In this strategy, a neutral clutch is
opened when the vehicle is at a stop, which effectively reduces the creep torque required from
the engine.127'128 BMW demonstrated a reduction in fuel consumption of 2 - 3 percent on the
NEDC for an optimized neutral idle decoupling system on an eight-speed transmission.129
Similarly, ZF calculated that implementing a neutral idle decoupling system on its eight-speed
transmission would reduce fuel consumption by 0.5 percent to 1.4 percent on the U.S. combined
cycle, depending on the K-factor of the torque converter.130  It should be noted, of course, that
the neutral idle decoupling simply reduces idling losses, and implementing stop-start system
would eliminate the effectiveness of this technology.

5.2.3.10       Transmission Shift Strategies

   The transmission shift schedule can strongly influence the fuel consumption over a drive
cycle. A more aggressive shift schedule will downshift the transmission earlier and upshift later
(i.e., at lower engine speeds). This moves engine operation, for a particular required power, to
lower speeds  and higher torques where engine efficiency tends to be higher. Along with this,
reducing time between shifts (i.e., allowing more shifts), reducing the minimum gear where fuel
cutoff is used, and altering torque converter slip (covered in the next section) will also decrease
fuel consumption.  Applying an aggressive shift strategy can reduce fuel consumption by about 5
percent in a generic six-speed transmission or 1-3 percent in a generic nine-speed
transmission.131  Similarly, BMW showed about a 2 percent reduction in CCh from
downspeeding the engine, comparing their current generation six-speed transmission to an earlier
       •   11'7
generation. z

   However, the application of the strategy is limited by NVH and drivability concerns, as lower
engine speeds produce more significant driveline pulses and allowing more shifts may increase a
shift busyness perception. Manufacturers reduce the NVH impact by using allowing partial
lockup, adding a torque converter dampener, and/or adding a pendulum dampener. These
changes along with decreasing the ratio between gears has made higher gear numbers and
increased shifting more acceptable.  Reducing the ratio between gears allows shifting to be less
perceptible due to the smaller change in engine speed.

5.2.3.11       Torque Converter Losses and Lockup Strategy

   Torque converters are typically associated with conventional ATs and CVTs, although they
have appeared on Honda's newest eight-speed DCT.   Torque converters provide increased torque
to the wheels at launch, and serve as  a torsional vibration  damper at low engine speeds.
However, this comes at the cost of energy loss in the torque converter fluid, and modern torque
converters typically have a lockup clutch that mechanically locks the impeller and turbine
together, bypassing the fluid coupling.
                                              5-58

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                         Figure 5.27 ZF Torque Converter Cutaway133

   Although in the past torque converters remained unlocked up to high vehicle speeds, recent
trends are to lock up at much lower speeds. Improvements in torsional vibration dampers, and
the ability to utilize micro-slip across the lockup clutch has enabled lower lockup speeds.
Mazda, for example, claims torque converter lockup as low as 5 mph for its SKYACTIV-Drive
AT.134 Although not as aggressive, BMW claims a 1 percent reduction in CCh from an early
torque converter lockup.135

5.2.4   Electrification: State of Technology

   Electrification includes a large set of technologies that share the common element of using
electrical power for certain vehicle functions that were traditionally powered mechanically by
engine power. Electrification can thus range from electrification of specific accessories (for
example, electric power steering) to electrification of the entire powertrain (as in the case of a
battery electric vehicle). Powering  accessories electrically can reduce their energy use by
allowing them to operate on demand rather than being continuously driven by the crankshaft
belt. Some electrical components may also operate more efficiently when powered electrically
than when driven at the variable speed of a crankshaft belt. Electrified vehicles that use
electrical energy from  the grid also  provide a means for low-GHG renewable energy to act as a
transportation energy source where  it is present in the utility mix.

   In the Technical Support Document (TSD)136 accompanying the 2012  FRM, electric power
steering and other improved accessories were discussed along with electrified vehicles under the
                                              5-59

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
topic of electrification.  In this Draft TAR, electric power steering and improved accessories are
now discussed separately in Section 5.2.8 in order to focus the current discussion on electrified
vehicles and 12V stop-start systems, which share many common themes.  As in the TSD, air
conditioning is not explicitly examined as an electrified accessory technology but is discussed
separately in Section 5.2.9  with respect to leakage, efficiency, and off-cycle credit provisions,
oriented primarily toward conventional vehicles. Where applicable, electrified air conditioning
is discussed in the context of electric vehicles, where it can have a strong impact on onboard
energy consumption and driving range.

   Electrified vehicles (or xEVs) are considered for this analysis to mean vehicles with a fully or
partly electrified powertrain. This includes several electrified vehicle categories, including:
battery electric vehicles (BEVs), which have an all-electric powertrain and use only batteries for
propulsion energy; plug -in hybrid electric vehicles (PHEVs), which have a primarily electric
powertrain and use a combination of batteries and an engine for propulsion energy; and hybrid
electric vehicles (HEVs), which use electrical components and a battery to manage power flows
and assist the engine for improved efficiency and/or performance. HEVs are further divided into
strong hybrids (including P2 and power-split hybrids) that provide strong electrical assist and in
many cases can support a limited amount of all-electric propulsion, and mild hybrids (such as
belt integrated starter generator (BISG) hybrids, crankshaft integrated starter generator (CISG)
hybrids, and 48V mild hybrids) that typically provide only engine on/off with minimum
electrical assist.

   BEVs and PHEVs are often referred to collectively as plug-in electric vehicles, or PEVs.
Although the FRM referred to battery electric vehicles as EVs, this Draft TAR adopts the term
BEV which is now more commonly used in the technical literature.

   Fuel cell electric vehicles (FCEVs) are another form of electrified vehicle having a fully
electric powertrain, and are distinguished by the use  of a fuel cell system rather than grid power
as the primary energy source.  Although EPA has not included FCEVs in its Draft TAR fleet
compliance modeling analysis, NHTSA did simulate the vehicles for its analysis. Technology
developments relating to FCEVs are reviewed in Section 5.2.4.5.

   As with the other technologies presented in this chapter, the agencies are reviewing, and
revising where necessary, the assumptions for effectiveness and cost of electrification
technologies for this Draft  TAR. The agencies have carried out this effort along several paths.
The agencies gathered information from many sources, including public sources such as journals,
press reports, and technical conferences, as well as manufacturer certification data and
information gathered through stakeholder meetings with  OEMs and suppliers.  EPA has also
benchmarked selected vehicles by means of dynamometer testing at the EPA National Vehicle
and Fuel Emissions Laboratory (NVFEL).   The agencies have also been leveraging instrumented
vehicle test data from the Argonne National Laboratory (ANL) Advanced Powertrain Research
Facility (APRF). Among other purposes, EPA has used this data to inform development of the
ALPHA model, and NHTSA has used this data to ensure that current powertrain technologies
and controls used in Autonomie are representing state-of-the-art as well as include additional
powertrains (i.e., the Voltec system).  The agencies have also leveraged electric machine
component performance data collected by Oak Ridge National Laboratory (ORNL) under U.S.
DOE funding, and similar component and vehicle test data provided by other laboratories such as
Idaho National Laboratory (INL).  EPA also worked closely with ANL to improve and update
                                             5-60

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
the battery costing model, known as BatPaC,137 which the agencies have used to update the
projected costs of electrified vehicle battery packs. All of these sources have contributed to our
assessment of the progress of electrification technology since the FRM.

   Overview of Section

   This Section 5.2.4 is intended to briefly review the assumptions for cost and effectiveness of
the electrification technologies described in the FRM, and to review industry developments since
the FRM that could inform the question of revising those assumptions for this Draft TAR
analysis. The information described in this section thus forms the basis for revised cost and
effectiveness assumptions described in Section 5.3.4.3, which become inputs to this Draft TAR
analysis. Source data for most charts in Sections 5.2.4 and 5.3.4.3 are available in the Docket.138

   Section 5.2.4 is organized in the following way:

       •  Subsection 5.2.4.1 provides a high-level overview of the major developments in
          electrification technologies since the FRM. This section is intended only as an
          executive summary to help place the topic of electrification into context.

       •  Subsection 5.2.4.2 provides a background in non-battery electrical components that
          are common to many of the electrification technologies, and briefly reviews the major
          directions of their development since the FRM. An understanding of these
          components is helpful to understanding developments in cost and effectiveness of
          each of the electrified vehicle categories.  Developments in the cost or performance of
          specific classes of components are discussed in the context of the electrified vehicles
          in which they have been implemented.

       •  Subsection 5.2.4.3 includes subsections detailing each of the major electrified vehicle
          categories (stop-start, HEVs, PHEVs and BEVs).  These sections serve to: (a) briefly
          review the significance of each electrified vehicle category as a means of reducing
          GHG emissions; (b) briefly review the major assumptions made about the electrified
          vehicle category in the FRM; and (c) review industry developments relating to how
          the category has evolved and been taken up in the fleet since the FRM.

       •  Subsection 5.2.4.4 focuses on developments in battery technology. Batteries are
          discussed separately and after discussion of the vehicle categories for several reasons.
          First, the battery performance requirements for each of the xEV categories is best
          understood after the categories have been fully defined and discussed.  Second, a
          greater level of technical detail is required to adequately assess some battery
          developments that have a strong influence on effectiveness or cost of xEV
          technologies. Finally and  perhaps most importantly, battery cost estimation is a
          particularly influential input to the cost assumptions for xEVs, and the battery cost
          estimates for different xEV categories rely on many detailed assumptions that are best
          understood and contrasted in the context of a battery discussion after trends in xEVs
          have been reviewed. The bulk of battery-related assumptions affecting xEVs are
          therefore covered in the battery section rather than the xEV sections.

   Finally, Subsection 5.2.4.5 focuses on developments in FCEVs.
                                              5-61

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
5.2.4.1 Overview of Electrification Technologies

   The 2012 TSD and the FRM analysis identified electrified vehicles as offering a strong
potential for reducing greenhouse-gas emissions. In the analysis conducted for the 2017-
2025MY FRM, the cost-minimizing compliance pathway showed electrified vehicles playing an
important supporting role in a fleet composed primarily of non-electrified powertrain
configurations. The pathway presented by EPA showed OEM compliance with 2025MY
standards with overall fleet penetrations of 2, 5, and 26 percent for BEVs, strong hybrids, and
mild hybrids, respectively.139

   Since the FRM, there has been significant growth in the number of FIEV, PHEV, and BEV
models available to consumers. HEVs are now part of the product line of almost every major
OEM. In 2014, U.S. HEV sales were in excess of 450,000 units but declined to about 385,000
units in 2015.140  Plug-in vehicles (BEVs and PHEVs) are also being offered in increasing
numbers. In MY2015, 28 models of plug-in vehicles were available, an increase from 23 models
in MY2014, and only a handful in 2012. In each of 2014 and 2015, U.S. plug-in vehicle sales
were in excess of 115,000 units.140

   Some aspects of BEV implementation and penetration have developed differently than
predicted in the FRM. The FRM conceived that the BEVs most likely to play a significant part
in OEM compliance would offer a real-world range of between 75 miles at the low end and up to
150 miles at the high end. Since then, the BEV market appears to have formed two segments,  a
consumer segment offering a driving range of around 100 miles at a relatively affordable price,
and a luxury segment offering a much higher range (well in excess of 200 miles) at a higher
price. Tesla Motors has had notable success at producing and marketing BEVs in the luxury
segment, causing significant numbers of BEVs to enter the fleet that may not have been
predicted by OMEGA on a pure cost-effectiveness basis.

   Going forward, both BEV segments appear to be aggressively pursuing range increases in
their second and third generation models. Both of the leading manufacturers in the consumer
segment, Nissan and GM, have recently announced firm plans to offer a 200-mile range BEV in
the 2017 time frame. Tesla is also making progress toward a long-stated intention to enter the
consumer segment with the Model 3, which is widely described as a 200-mile BEV and targeted
for introduction in 2017.

   An increasing number of OEMs are beginning to add PHEVs to their product lines, utilizing
both blended-operation architectures as well as extended-range architectures that offer varying
amounts of all-electric range. The cost-minimizing pathway for compliance with the 2025MY
standards did not project a necessity for significant fleet-level penetration of PHEVs (nominally,
zero percent), although it did project that some primarily luxury- and performance-oriented
OEMs might include PHEVs as part of their individual pathways.141 The 2015 and 2016 MYs
saw a discernible increase in PHEV20-style architectures from OEMs that tend to specialize in
luxury or high-performance vehicles, suggesting that this projection was accurate. Second-
generation PHEV models have begun to appear, typically offering an increased all-electric  range
or a more robust blended-mode operation that allows for increased all-electric capabilities in
normal driving. Manufacturers have often cited customer demand for a more all-electric driving
experience in making these changes.
                                             5-62

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Advancements in the cost and effectiveness of xEVs are closely related to advancements in
battery, electric motor, and power electronics technologies. These technologies have advanced
steadily since the FRM, with significant improvements in battery specific energy, battery cost,
and non-battery component efficiency and cost contributing to improvements in production
xEVs.  The pace of industry activity in this area suggests that further advancements are likely to
occur between now and the 2022-2025 time frame of the rule.

   At the time of the FRM, empirical data regarding the cost and efficiency of xEV components
was limited by the small number of production vehicles from which it could be gathered. Today,
the relatively large number of production models provides much greater opportunity to
empirically validate projections made in the FRM.

   Battery cost is a major consideration in the cost of xEVs. At the time of the FRM there was
great uncertainty in the manufacturing costs for these components and their potential to be
reduced.  There was also uncertainty regarding battery lifetime. Today, evidence of the need for
battery replacement is rare, with most PHEV and BEV batteries showing good durability within
the limits established by OEM warranties. Although the battery cost projections published in the
FRM were significantly lower than estimates of prevailing costs at the time, recent evidence
strongly suggests that these estimates were quite accurate, with at least  one major manufacturer
having announced battery costs from a major battery supplier that are very close to FRM
projections even for the 2017-2018 time frame.  Recent reports have suggested that lithium-ion
battery cost has historically followed a pace of improvement of about 6 to 8 percent per year.142
Advancements in cost and energy capacity of battery technology continue to be pursued actively
by OEMs and suppliers alike, suggesting that there is room for further improvement within the
2022-2025 time frame of the rule.

   Analysis of current and past production BEVs and PHEVs suggests that the FRM predicted a
larger battery capacity per unit driving range than manufacturers have found necessary to
provide. This could be due in part to differences in assumed powertrain efficiencies, usable
battery capacity, or application of road load reducing technologies.  Similar analysis also
suggests that the industry  has achieved comparable acceleration performance with significantly
lower motor power ratings than the FRM anticipated. In other words, it is clear that in many
ways the industry has found ways to do more with less, compared to many of the predictions of
the 2012 FRM.

   Because the vehicle architecture for electrified vehicles is fundamentally different from that
of conventionally-powered vehicles, the consumer experience is likely to be different as well. In
particular, the fueling requirements of BEVs and PHEVs call for changes in accustomed fueling
habits,  some of which may improve convenience (e.g. the ability to charge at home) while others
may pose a challenge (e.g. a relatively long fueling time). A BEV with limited range might not
provide an exact substitute for a conventional vehicle for many consumers today, while at the
same time electrified vehicles can provide benefits of quiet operation, reduced maintenance, and
the potential integration with future mobility systems that might include shared and autonomous
vehicles.  Chapter 6 contains a more complete discussion of the impact  of efficiency
technologies on other vehicle attributes.

   The primary factors that influence the cost and effectiveness of electrification technologies
are the cost and efficiency of their components. These include: energy  storage components such
as battery packs; propulsion components  such as electric motors; and power electronics
                                             5-63

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
components, such as inverters and controllers, that process and route electric power between the
energy storage and propulsion components. For the purpose of this analysis, these components
are divided into battery components and non-battery components.

   Battery components have a particularly strong influence on cost of xEVs. Because
developments in battery technology may apply to more than one category of xEV, they are
discussed collectively in Section 5.2.4.4.  That section details developments in battery-related
topics that directly affect the specification and costing of batteries  for all xEVs, such as usable
capacity, durability, thermal management, and pack topology, among others.

   Non-battery components have a strong influence on both cost and effectiveness of xEVs.
Because non-battery technologies are important to understanding the differences in architecture
among xEVs, they are introduced prior to discussion of the individual electrified vehicle
categories in Section 5.2.4.2.

5.2.4.2 Non-Battery Components of Electrified Vehicles

   Non-battery components largely consist of propulsion components and power electronics.
Propulsion components typically include one or more electric machines (an umbrella term that
includes what are commonly known as motors, generators, and motor/generators).  Depending
on how they are employed in the design of a vehicle, electric machines commonly act as motors
to provide propulsion, and/or act as generators to enable regenerative braking and  conversion of
mechanical energy to electrical energy for storage in the battery. Power electronics refers to the
various components necessary to route current between the battery system and the propulsion
components, including such devices as inverters and rectifiers, DC-to-DC converters, motor
controllers, and on-board battery chargers.

   The energy efficiency of non-battery components has been the focus of much industry
research and development since the 2012 FRM.  The impact of resulting improvements in
efficiency and overall system optimization therefore need to be considered in developing
estimates of xEV effectiveness.  The agencies have studied and considered such improvements in
developing new estimates of xEV effectiveness for this Draft TAR.

   Costs of non-battery components have also begun to decline. Compared to engines and other
conventional powertrain components,  many of which have been reduced to commodity products
for many years, the market in xEV non-battery components is far less developed.  As OEMs seek
xEV components for their products, they are less likely to encounter stock items that fully meet
their requirements and therefore  have often chosen to either produce them in limited numbers in-
house, or to source them from suppliers that build to specification. While this dynamic may be
expected to limit the potential for economies of scale to develop and be reflected in component
costs in the near term, it is also likely that standardization and commoditization will occur as the
industry matures.  One example of industry movement in this direction is shown by the decision
of LG to leverage  its position as  xEV battery supplier to several OEMs by expanding into xEV
non-battery components. In a joint announcement with LGChem in October 2015,143 GM
described LG's role not only as supplier of battery cells for the Chevy Bolt BEV but also as
supplier of many of its non-battery components.  LG's established  role  as battery supplier to
multiple OEMs suggests that it may be planning to supply non-battery  components across the
rest of the xEV industry as well.  As another example, in 2016 Siemens and Valeo announced the
formation of a joint venture for the production of high-voltage components across the full range
                                             5-64

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
of electrified vehicle types, citing among other advantages "substantial synergies in
manufacturing and sourcing" and a focus on global markets.144 Developments such as these can
promote the potential for economies of scale to develop, and may be a significant driver of cost
reductions if they continue in the future.

5.2.4.2.1     Propulsion Components

   The components that provide propulsion for xEVs are known variously as electric motors,
traction motors, motor/generators, e-motors, or electric machines. In this discussion, they will be
referred to either as electric motors or generators (depending on the functional context), or
collectively as electric machines.

   The two main types of electric machines currently seen in production xEVs are permanent-
magnet motors (also known as synchronous motors) and induction motors (also known as
asynchronous motors). Although the permanent-magnet motors used in xEVs are sometimes
called brushless direct-current (DC) motors, these  as well as induction motors are powered by
alternating current (AC), which must be converted from DC battery current by an inverter.

   In the duty cycles typical of xEV applications, permanent-magnet motors have certain
advantages in energy efficiency due in part to the presence of integral permanent magnets to
generate part of the magnetic field necessary for operation.  However, these magnets add to
manufacturing cost, particularly when they contain rare earth elements.  In contrast, induction
motors use copper windings to generate all of the magnetic field and can be manufactured
without rare earth elements. Although the windings are significantly less costly than magnets,
generation of the field in the windings is subject to additional I2R losses that are not present in
permanent magnet motors. In some conditions, this causes induction motors to be slightly less
energy efficient than permanent-magnet motors,145'146 although the choice between the two types
of motor ultimately depends on the specific application.

   The majority of current xEV products use permanent-magnet motors. Induction motors are
found in products of Tesla Motors, as well as the Fiat 500e and Mercedes-Benz B-Class Electric
Drive.  The BMW Mini-e and the Toyota RAV4 EV, both now discontinued, also used induction
motors; in the case of the RAV4, the motor was supplied by Tesla.

   Another type of motor, the switched reluctance or axial flux motor, has recently been
suggested for use in xEVs.147'148  Although current examples of this technology are challenged
by difficulties with controllability, vibration, and noise, in the future these motors may
potentially offer a lower cost solution than either permanent-magnet or induction motors.

   Since the FRM,  some manufacturers have demonstrated successful cost reductions in
propulsion components.  For example, the  use of rare-earth metals in permanent-magnet motors
is commonly cited as a concern due to their high cost and potential supply uncertainty.  The 2016
second-generation Chevy Volt has reduced the use of rare-earths in its drive unit by more than 80
percent by using lower-cost ferrite magnets in place of rare-earths in one of its motors149 and
significantly reducing the rare-earth content of the other.150 Another approach is seen in the
BMW 13, which uses a hybridized motor design that combines aspects of the permanent-magnet
motor and the reluctance motor, allowing rare earth content to be reduced by about half
compared to a permanent-magnet motor of similar torque capability.146
                                             5-65

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Component integration has also contributed to lower costs.  GM has cited integration of
power electronics with the transmission and drive unit of the 2016 Volt as a significant enabler
of cost reductions in that vehicle by eliminating long stretches of heavy cable and improving
packaging efficiency.151'152 Major changes to the configuration of the electric propulsion system
reduced the total torque and power requirements, allowing the use of smaller bearings and rotors,
and an increase in maximum motor speed to  11000 rpm from the 9500 rpm of the previous
system. This led to a 20 percent reduction in motor volume and a 40 percent reduction in mass
compared to the previous generation, as well as improved efficiencies. Similar improvements
have propagated to the Cadillac CT6153 and the Chevy Malibu Hybrid154 through the sharing of
related components. The 2016 Toyota Prius  also utilizes improvements to the transaxle and
motor that result in significant weight reduction and efficiency.  A more compact motor design
and an improved reduction gear allows for an improved power-to-weight ratio and provides for a
20 percent reduction in frictional losses.155

   Industry activity is also focused toward improving the efficiency of propulsion motors.
Although electric motors are already highly efficient (well in excess of 90 percent in many
normal usage conditions), even small improvements in efficiency can pay significant dividends
by reducing the battery capacity necessary for a given driving range. For example, GM has said
that the increased range of the second generation Chevy Volt was achieved in part by
improvements in motor efficiency.151  Even the first generation of the Chevy Spark EV was
described as having the highest drive unit efficiency in the industry, with an average battery-to-
wheels efficiency of 85  percent in the city cycle and 92 percent in the highway cycle.156 These
efficiencies are  higher than EPA assumed in  the 2012 FRM xEV battery sizing analysis.

5.2.4.2.2     Power Electronics

   Power electronics refers to the various components that control or route power between the
battery system and the propulsion components, and includes components such as: motor
controllers, that issue complex commands to  precisely control torque and speed of the propulsion
components; inverters and rectifiers, that manage DC and AC power flows between the battery
and the propulsion components; onboard battery chargers,  for charging the BEV or PHEV
battery from AC line power; and DC-to-DC converters that are sometimes needed to allow DC
components of different voltages to work together.

   Inverters are power conditioning devices that manage electrical power flows between the
battery and propulsion motors. While all batteries are direct current (DC) devices, modern
traction motors  operate on alternating current (AC) and therefore require an inverter capable of
converting DC to AC of widely variable frequencies at variable power levels.  As implemented
in an electrified vehicle, the component commonly known as an inverter may also act as a
rectifier, that is, convert AC to DC to send energy to the battery.

   Modern inverters are semiconductor based, utilizing metal-oxide-semiconductor  field-effect
transistors (MOSFET) or insulated-gate bipolar transistors (IGBT). These designs are highly
efficient, often operating well above 90 percent efficiency. Inverter designs vary in output
waveform (square wave, sine wave, modified sine wave, or pulse-width modulated), which
accounts in part for differences in their efficiency and the potential for heat generation. Inverter
manufacturing cost is strongly associated with wafer size in manufacturing of substrate materials
such as silicon carbide.  While most wafer sizes are currently around 4 inches in diameter, larger
wafers of 6 to 12 inches would reduce scrap rates and reduce cost substantially.157
                                             5-66

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Despite these low losses, the high power levels of electrified vehicles generate significant heat
and require inverters to have aggressive liquid cooling, often residing on the coolant loop in a
position prior to the propulsion motor to ensure sufficient cooling.  Cooling elements such as
fans, heat exchange surfaces and fins or heat sinks can add to volumetric requirements and are a
common target of size and cost reduction.  The similarity of materials and cooling needs offer an
opportunity to further reduce cost by integrating the inverter with other power electronics
components such as DC converters.158

   The 2016 Chevy Volt provides one example of how improvements to the inverter and its
packaging can lead to significant improvements in packaging and related costs. Major changes
to the electric propulsion system served to reduce the current requirements of the inverter,
reducing its volume by about 20 percent (from 13.1L to 10.4L) and its mass from 14.6 kg to 8.3
kg. This allowed the inverter module to be integrated into a small space at the top of the
transmission.  This integration into the transmission saved on assembly costs, served to protect
the components and their sensitive interfaces in a sealed environment, and eliminated the need
for heavy 3-phase cables. It also saved valuable underhood space for other components
commonly associated with electrification.  The reduction in inverter current was also said to
reduce inverter switching loss by about half in conjunction with accompanying improvements to
cooling.  GM attributed a 6 percent improvement in electric drive system efficiency over the FTP
cycle, a 30 percent increase  in vehicle range and an 11 percent improvement in label fuel
economy to these inverter improvements.151'152 Similar improvements have carried over to other
models that share related components, such as the Cadillac CT6 and the Chevy Malibu
Hybrid.153'154  Toyota also has introduced changes that improve  inverter efficiency.155 The 2016
Toyota Prius includes a new power control unit to which it attributes a 20 percent reduction in
power losses. The power control unit also benefits from integration, residing in a position above
the transaxle.  Advances in the use of a silicon carbide substrate in the power control unit are
also expected to significantly reduce power switching losses and allow a 40 percent reduction in
the size of the coil and capacitor of the power control unit in production Toyota vehicles by
around 2020.159

   Many systems require DC-to-DC converters to allow DC components of different voltages to
work together. They do not convert between AC and DC, but instead step up (or down) the DC
voltage between two or more components  or subsystems, either unidirectional or bi-directionally.
One common application of a DC-to-DC converter is to allow low-voltage accessories to be
powered by energy from the high-voltage battery by reducing the voltage from 300+ V to 14V.
These are also known as buck converters, and may operate at about 1.5 kW160 to 3 kW.167
Although many current production BEVs and PHEVs retain a low-voltage battery to power
accessories, a buck converter is needed to keep the low-voltage battery charged in the absence of
an engine-driven alternator,  and can supplement power to the accessories. Another purpose of a
DC-to-DC converter is to allow certain powertrain components to operate at their optimum
voltage rather than being tied to the voltage of the high voltage battery. For example, a fuel cell
stack or super capacitor may operate more efficiently at a higher or lower voltage than the high-
voltage battery, or along a variable range of voltages.161 A variety of topologies are under
development to suit these varied applications.160'161

   Controllers are electronic devices that implement control algorithms that control power flows
through the electrified powertrain. Motor controllers are responsible for issuing the  complex
commands that precisely control torque and speed of the propulsion motor. A primary task of
                                             5-67

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
this controller is to determine the exact frequency of alternating current necessary for the motor
to deliver the demanded speed and torque, and to control the inverter to provide it. A
supervisory controller is another form of controller that implements higher-level vehicle control
algorithms, including issuing high-level torque and speed commands to the motor controller.
Supervisory controllers are not unique to electrified powertrains but may be functionally
integrated with other components that are. Compared to other power electronics components,
controllers are not typically large consumers of energy, but can benefit from cost reductions
applicable to other components.

   Onboard chargers are charging devices permanently installed in a PHEV or BEV to allow
charging from grid electrical power.  Level 1 charging refers to  charging powered by a standard
household 110-120V AC power outlet. Level 2 charging refers to charging at 220-240V AC
power. The available power depends on the amperage of the household circuit, and can range
from about 1 to 2 kW for Level 1 to about 5 to 7 kW for Level 2. Onboard chargers travel with
the vehicle, and are distinct from stationary charging equipment (Electric Vehicle Supply
Equipment, or EVSE) commonly installed at public or private charging stations. More
information on PHEV and BEV charging infrastructure and EVSE  can be found in Chapter 9,
Infrastructure Assessment.

   The widespread home availability of 110-120V AC power does  not necessarily mean that
Level 1 charging is preferable either for convenience or efficiency. Charging time at the Level 1
rate is much slower than at Level 2, and can become impractically slow for longer-range BEVs
that may take longer than overnight to bring to full charge at Level 1 even after only partial
depletion. However, Level 1 residential charging is widely relied upon by the current users of
BEVs and PHEVs and provides a lower cost option for ownership that may continue to be
sufficient for households with lower daily driving needs.

   Charging efficiency can also vary significantly. In general, the efficiency with which a battery
accepts DC charge current is highest at low charge rates.162 However, the degree to which the
manufacturer has optimized the charging circuitry for a specific preferred charge rate can also
have a strong influence, because the efficiency of AC to DC conversion is also an important
factor. According to tests performed by Idaho National Laboratory on a 2015 Nissan Leaf, the
efficiency of Level 1 charging ranged from only 61.8 percent to a maximum of 78.4 percent,
while that of Level 2 charging ranged from 81.5 percent to 90.5 percent.163 This suggests that
the design of the charging circuitry can have a greater effect on  charging efficiency than charge
rate alone, and that manufacturers may optimize the charging system to accommodate the mode
of charging it expects customers to most commonly utilize.

   DC fast charging is increasing in availability and popularity,  and can support charging at
much higher rates than Level 2 (up to 150 kW in some cases, subject to the capability  of the
vehicle being charged). Charging at these higher rates may result in a lower net efficiency
relative to Level 2, and may require more robust battery cooling to  dissipate the heat generated
during a charge.

   Although charging efficiency is primarily relevant to upstream emissions and is not a factor in
onboard energy consumption, there is significant potential for efficiency improvement in these
components that may be indicative of similar potential in other power electronics components.
For example, between Genl and Gen2 of the Chevy Volt, the energy efficiency, size and weight
of its onboard charger was improved significantly.164'152 Level 1 charging efficiency improved
                                             5-68

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
from 86.8 percent in Genl to 94.5 percent in Gen2, an improvement of 8.9 percent.  Efficiency at
Level 2 increased similarly from 89.6 percent to 95.5 percent, an improvement of 6.5 percent.
These improvements allowed the overall system efficiency (from the wall plug to the battery) of
Level 2 charging to improve to 88.4 percent, and that of Level 1 to 86.7 percent (improvements
of 8.6 percent and 9.3 percent, respectively). Power density of the unit improved from 326 W/kg
to 605 W/kg (85 percent), while volumetric power density improved from 492 W/liter to 889
W/liter (81 percent), which led to significant packaging advantages.  The fact that these
improvements to charger efficiency were achieved despite their lack of a strong impact on highly
visible attributes such as driving range or power suggests that similar improvements to other
components that do affect range or power are even more likely to be pursued successfully.

   Battery management systems (BMS)165'283 are an important factor in maintaining and utilizing
the available capacity of the traction battery. The primary role of a BMS is to maintain safety
and reliability by preventing usage conditions that would damage or excessively degrade the
battery. The BMS may therefore limit voltages and currents on the pack, module, or individual
cell level, and monitor pack or cell temperature as well as other parameters.

   Another important role of the BMS is to balance the charge levels of the individual battery
cells so that each cell is maintained at a similar voltage and state of charge.  This can play an
important part in determining the usable portion of total battery capacity and in maintaining
battery life. In a battery containing hundreds of cells, small variations in resistance will exist
among individual  cells, and differences in cell temperature will result not only from these
differences but also from differences in cell location within the pack and proximity to  cooling
media. During a normal charge or discharge of the pack, these differences will affect cell
efficiency and cause some cells to approach their voltage or charge limits sooner than  others.
Without balancing, the  entire pack will effectively reach its charge or discharge limit when the
weakest cell reaches its limit. In this case, the charge contained in the remaining cells goes
unutilized. Effective cell balancing can increase utilization significantly.

   BMS systems may employ passive or active balancing. Passive balancing acts to identify the
cells that are approaching their limits and selectively modifies their charge or discharge rates,
usually by dissipating their energy resistively, to allow the remaining cells to continue operating.
Active balancing shuttles energy among cells rather than dissipating the energy.  Active
balancing is potentially more energy efficient than passive balancing but is typically more costly
to implement. The cost and effectiveness of active balancing  is an active area of industry
research toward reducing the necessary battery capacity and power for a given application.

5.2.4.2.3     Industry Targets for Non-Battery Components

   Establishing targets can be an effective way of focusing industry effort toward a common
goal. For example, the battery cost and performance targets established by the United States
Advanced Battery Consortium (USABC) are familiar to most in the battery industry and have
become important reference points by which developments in battery technology are often
measured. While  industry targets  such as these can vary in their purpose and achievability, they
can provide valuable guidance on what some in the industry consider to be potential directions
for future technology.

   Targets for cost and  performance of non-battery components have been established by U.S.
DRIVE,166 a government-industry partnership managed by the U.S. Council for Automotive
                                              5-69

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Research (USCAR), which also manages USABC. Members include the U.S. Department of
Energy, industry members of USCAR, and several other organizations including major energy
companies and public energy utilities. The U.S. DRIVE targets apply to electric motors,
inverters, chargers, and other power electronics components for the 2015 and 2020 lab yearu
time frames.167  These targets,  some of which are shown in Table 5.1, include performance
targets such as specific power, specific energy, and energy and power density (volumetric), as
well as cost targets.

   The U.S. DRIVE targets were established specifically with respect to HEVs, which were seen
as presenting the greatest challenge in meeting the  targets due to their being on the low end of
the power range compared to PEVs.  The targets therefore apply best to an HEV-sized 55 kW
system. U.S. DRIVE expects the targets to be less difficult to meet for higher-power PEV
systems, in part because their more powerful powertrains may incur less overhead cost (for
connectors and the like) that are not necessarily directly proportional to power.168  This suggests
that the U.S. DRIVE targets would be relatively conservative when applied to PEVs.

   Although the U.S. DRIVE figures are only targets, the industry has shown remarkable
progress in approaching these goals.  It is notable that U.S. DRIVE targets for specific power are
quite close to what was already available in some production HEVs at the time they were set.
Since some of the goals were being met in higher-priced products, bringing these levels of
performance to the average PEV may largely be a matter of cost reduction rather than
technological breakthrough.
             Table 5.1 U.S. DRIVE Targets for Electric Content Cost and Specific Power
Component
Electric motor
Power electronics
Motor and electronics combined
3 kW DC/DC converter
U.S Drive Target (Lab Year)
2015
1.3 kW/kg
$7/kW
12 kW/kg
$5/kW
1.2 kW/kg
$12/kW
1.0 kW/kg
$60/kW
2020
1.6 kW/kg
$4.7/kW
14.1 kW/kg
$3.30/kW
1.4 kW/kg
$8/kW
1.2 kW/kg
$50/kW
   The 2020 lab year target for specific power of combined motor and power electronics has
some support in current literature. Assuming a five year lag between lab demonstration and
production, the 2020 lab year corresponds to 2025. A presentation by Bosch169 at The Battery
Show 2015 states that the electric motor and power electronics for a  100 kW, 20 kWh BEV
system in the 2025 time frame is expected to comprise about 37 percent of electric content
weight, with battery weight comprising the remaining 63 percent. Assuming the 20 kWh battery
pack has a specific energy of about 140 Wh/kg (as indicated by ANL BatPaC for an NMC622
u It should be noted that a minimum of five years typically passes between successful demonstration of a technology
  in a lab and its introduction into the market.
                                             5-70

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
pack at 115 kW net battery power), and a corresponding weight of 143 kg, the non-battery
content would be estimated at about 53 kg.  The 100 kW system would then represent a non-
battery specific power of 100 kW/53 kg, or 1.88 kW/kg. While the U.S. DRIVE target of 1.4
kW/kg is not directly comparable because it is based on a 55 kW traction motor, the result for the
100 kW example is directionally correct in the sense that U.S. DRIVE considers the targets
easier to achieve for more powerful systems.168  Most BEV and PHEV motors modeled in this
analysis are larger than 55 kW, suggesting that the U.S. DRIVE figure for a 55 kW system may
represent a fairly conservative figure for these applications.

   Although the U.S. DRIVE figures are targets and therefore not necessarily indicative of
industry status, EPA has confidence that the targets for specific power represent attainable
production goals during the time frame of the rule.  This is based in part on the observation that
the 2020 specific power target for electric motor and power electronics combined is very close to
levels that were already being attained by some production vehicles at the time they were set.170
Further, the motor of the recently announced Chevy Bolt BEV already appears to exceed the
U.S. DRIVE target at 1.97 kW/kg (based on a mass of 76 kg and peak power of 150 kW).171
This example is consistent with confidential business information conveyed to EPA through
private stakeholder meetings with OEMs that suggests that cost and performance targets for
some types of components are already being met or exceeded in production components today,
or are expected to be met within the time frame of the rule.

5.2.4.3 Developments in Electrified Vehicles

   In this Draft TAR, each of the  electrified vehicle categories represents a distinct GHG-
reducing electrification technology that manufacturers may choose to include as part of a
compliance pathway. These technologies range from 12-volt stop-start systems without
accompanying hybridization, to mild and strong hybrids (HEVs), to plug-in vehicles (PHEVs
and BEVs) and fuel-cell electric vehicles (FCEVs). The propulsion and power electronics
technologies discussed in the previous section are integral to understanding the architecture and
capabilities of each of these electrification technologies. Developments in each of these
electrification technologies are described in this section.

5.2.4.3.1     Non-hybrid Stop-Start

   In this analysis,  non-hybrid stop-start refers to a technology that reduces idling by temporarily
stopping the engine when the vehicle stops and restarting it when needed.  This eliminates much
of the fuel consumption associated with idling. In urban driving conditions that include a large
amount of idling at intersections and in congested traffic, stop-start can provide significant GHG
benefit.

   Non-hybrid stop-start is also commonly known as idle-stop or micro hybrid. In the 2012
FRM, it was referred to as conventional stop-start. In this Draft TAR analysis, as in the FRM,
non-hybrid stop-start is limited to engine stopping and restarting in a 12V context, with no
accompanying hybridization. For this reason, the term micro-hybrid will not be used to refer to
non-hybrid stop-start systems. The non-hybrid stop-start classification should not be confused
with mild and strong hybrids that include a stop-start function.  Systems that include brake
energy regeneration or other hybrid features would be  classified as hybrids. However, as in the
Ricardo analysis of the 2012 FRM, non-hybrid stop-start may include  a strategy known as
                                             5-71

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
"alternator regen" that charges the 12V battery more aggressively by increasing the alternator
field upon vehicle deceleration.

   Non-hybrid stop-start is therefore the simplest form of electrification discussed in this section.
It is typically implemented by: (a) upgrading to a higher-performance starter capable of higher
power and increased cycle life, (b) upgrading to a higher-performance 12V battery to improve
cycle life and reduce voltage drop on restart; (c) adding an appropriate control system to manage
stopping and starting as transparently as possible; and in many cases, (d) modifying certain
accessories to allow for adequate service while the engine is off.

   In the 2012 FRM, the effectiveness  estimates for stop-start were derived from the Ricardo
modeling study. The agencies estimated the 2-cycle effectiveness of stop-start technology to be
in the range of 1.8 to 2.4 percent, depending on vehicle class.  The 2012 FRM considered stop-
start to be a new technology and assigned it a steep learning curve for the years 2012-2015 and a
flat learning curve for the years 2016-2025. On the basis of projected costs and effectiveness,
the agencies projected that stop-start would achieve a fleet-level penetration of 15 percent172 in
the cost-minimizing pathway for compliance with the 2025MY standards.

   Since the 2012 FRM, rapid growth in the application of 12V stop-start systems is evidence of
the technology's potential to provide cost-effective emissions reductions.  The 2015 EPA Trends
Report projects that non-hybrid stop-start will be present on almost 7 percent of new non-hybrid
car and truck production in MY2015, with total penetration of stop-start at nearly 9 percent when
mild and strong hybrids are included.173 Penetration has grown steadily each year, reaching 0.6
percent in 2012, 2.3 percent in 2013, and 5.1 percent in 2014, with 6.6 percent projected for
2015.174 BMW and Mercedes-Benz are the most notable adopters, each including stop-start in
about 70 percent of their projected 2015 production.175

   As a GHG-reducing technology, the effectiveness of stop-start depends on the amount of idle
time included in the assumed test cycle. The standard EPA test cycles contain short periods of
idle, but less than some believe is present in real world driving.  In order to provide a more
accurate credit basis for the real-world benefit of stop-start, the 2012 FRM provided for stop-
start technology to be eligible for off-cycle credits under the Off-Cycle Program.  The Off-Cycle
Program is discussed further in Section 5.2.9.

   In contrast to the FRM projections of 1.8 to 2.4 percent effectiveness under EPA test cycles,
other sources have suggested an average of 3.5 percent.176'177'178 As one example, the 2015 Ford
Fusion 1.5L TGDI is available with and without a  12V stop-start option, providing an
opportunity to assess the effectiveness  of stop-start as implemented in this vehicle. The
difference in estimated fuel economy between the two versions suggests an effectiveness of
about 3.5 percent on a fuel economy basis.  The automotive supplier Schaeffler Group has
presented an engine stop-start technology179 it describes as capable of providing a 2-cycle
combined fuel economy improvement  of about 6 percent over the city cycle and 2 percent over
the highway cycle, or about 3.42 percent combined. The 2015 Mazda 3 is available with and
without the Mazda i-ELOOP regenerative braking and stop-start  system. A comparison of
certification test data for this vehicle with and without the system suggests that its two-cycle
GHG effectiveness is about 3.35 percent.180

   Some test cycles used in other parts of the world include a greater proportion of idle time  and
therefore assign a greater benefit to stop-start. This would naturally make stop-start more
                                              5-72

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
attractive to manufacturers in regions that certify under these cycles, and may be a factor in the
greater penetration of stop-start that has been observed worldwide.  Stop-start176 has been
popular in Europe due to high fuel prices and the stringent EU CCh emission target established in
2009. In 2014, about 60 to 70 percent of vehicles sold in the European market offered stop-start.

   Because stop-start technology alters the customary operation of the engine, it has potential to
alter the traditional feel of driving. Frequent restarts of the engine, although rapid and seamless
in most implementations, can increase the sense of noise, vibration, and harshness (NVH).
Drivers unaccustomed to stop-start may feel uncomfortable having the engine switch off in stop
and go traffic, particularly if accessories such as heat or air conditioning are also affected. Some
of the seamlessness and potential benefit  of stop-start can be  eroded by individual driving habits.
For example, if a driver repeatedly pulls up toward the leading car as traffic compacts while
waiting at an intersection, the engine may restart each time, reducing fuel savings and adding to
NVH.

   Manufacturers often cite consumer acceptance factors in the adoption of stop-start in the U.S
market. Early introductions of the technology involved lower volume vehicles and adaptations
of systems originally designed for the European market.  Manufacturers have considered
customer feedback from these early applications in the implementation of recent stop-start
systems, which are now smoother and more unobtrusive to the driver. For example, some
suppliers have proposed continuously engagement of the starter motor to improve the restart
process. Others have implemented systems that maintain a specific piston position while stopped
in order to achieve a fast and smooth restart by firing a single cylinder.  As  a result, improved
systems promise greater effectiveness through more frequent and longer periods of idle stop time
while operating in a more transparent manner.

   Vehicles with sufficiently smooth and seamless stop-start  technology have been well-received
by consumers,181 especially when paired with some explanation of the system's benefits and
operating characteristics at the time of delivery. With these more recent implementations, it is
more common now for stop-start systems to be applied as standard equipment on high-volume
vehicles like the Chevrolet Malibu, Chrysler 200, Jeep Cherokee, and Ram  1500 truck. Ford
expects to offer it on 70 percent of its North America vehicle lineup by 2017,182 including the
2015 F-150 truck.

   The introduction of stop-start has stimulated development of 12V battery systems capable  of
providing the enhanced performance and cycle life that it requires.  Much of this activity has
involved variations of lead-acid chemistries, such as absorbed-glass-mat (AGM) designs and
lead-carbon formulations. For example, at the 2015 Advanced Automotive Battery Conference
(AABC), a Planar Layered Matrix (PLM) 12V enhanced lead-acid battery was exhibited by
Energy  Power Systems (EPS). EPS claimed this technology increases battery power and
regenerative charging capability by a factor of four while increasing the battery life by a factor of
five, at a similar cost to a conventional AGM lead-acid battery.

   Other developments have shifted toward lithium-ion chemistries specially adapted for stop-
start applications. As one example, Maxwell Technologies has developed a 12V lithium-ion
battery combined with a 395V ultra-capacitor pack designed  for 12V stop-start systems.183  The
dual pack was said to provide quicker engine start, lower voltage drop, capacity and life
improvement while providing capability to operate at -30 degrees Celsius.  Since the battery and
ultra-capacitor operate at different voltages, these systems require additional electronics for DC
                                              5-73

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
to DC conversion.  These systems are also likely to cost more than lead-acid based systems. The
cost of the Maxwell dual pack stop-start system is estimated at about $230/pack, which is higher
than that of an advanced lead-acid battery. In general, use of the lithium-ion chemistry for 12V
stop-start applications continues to face challenges with regard to cost as well as cold-start
operation.

   The Mazda i-ELOOP system184 represents an incremental step beyond basic stop-start, using
ultracapacitors to store regenerative brake energy during deceleration and coasting. While the
system cannot use the reclaimed energy for propulsion, it supplements the energy used by
accessories and climate control, potentially saving energy by allowing the engine to stay off for
slightly longer periods.

   Based on a review of these and similar industry developments, as well as data collected from
other sources, the agencies have updated  effectiveness estimates for stop-start technology.
Updated cost and effectiveness estimates are discussed further in Sections 5.3 (GHG
Assessment) and 5.4 (CAFE Assessment).

5.2.4.3.2      MildHybrids

   In this analysis,  mild hybrid refers to a technology that supplements the internal combustion
engine by providing limited hybridization, typically including a limited amount of electrical
launch assistance, some regeneration, and stop-start capability.  Together, these features reduce
energy consumption by optimizing loading of the engine, enabling some engine downsizing,
allowing the engine to turn off at times, and recovering a portion of the energy that would
otherwise be wasted by friction braking.  Mild hybrids commonly are implemented in part by
replacing the standard alternator with an enhanced power, higher voltage, higher efficiency belt-
driven starter-alternator which can provide some propulsion assist and also recover braking
energy while the vehicle slows  down (regenerative braking).  Although the belt-driven basis of
these systems can limit their power capability to approximately 10 kW to 15 kW,185 mild hybrids
can provide greater benefit than stop-start systems while keeping cost significantly lower than
that of a strong hybrid.

   Mild hybrids operate at a higher voltage than 12V stop-start systems. Even the relatively mild
demands of stop-start186 technology pushes a 12V electrical system to its limits. Achieving the
10 to 15 kW demanded of a mild hybrid application at 12V could lead to discharge currents of
1000 Amps or more, and would require very thick, heavy, and expensive electrical conductors.
In order to achieve effective launch assist and regeneration, mild hybrids therefore operate at
higher voltages of 48V to 120V or higher, with  an increased battery capacity as well.  The higher
system voltage allows the use of a smaller, more powerful electric motor and reduces the weight
of the motor,  inverter, and battery wiring harnesses.

   In the 2012 FRM analysis, mild hybrid technology was referred to as "higher-voltage stop-
start/belt integrated starter generator (BISG)" and was limited to BISG architecture, as
exemplified by the Chevrolet Malibu eAssist system. The primary source of effectiveness data
used by EPA was derived from the Lumped Parameter Model based on modeling of the Malibu
Eco BAS (BISG) system with a 15 kW motor and 0.5 kWh battery. EPA cost estimates were
based on an analysis of this system with a 0.25 kWh battery.  NHTSA used estimates of BISG
mild hybrid effectiveness developed by ANL using Autonomie. EPA assumed an absolute CCh
effectiveness ranging from 6.8 to 8.0 percent depending on vehicle class (2012 RIA, p. 1-18).
                                              5-74

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
The absolute effectiveness for the CAFE analysis ranged from 8.5 to 11.6 percent depending on
vehicle class. These effectiveness values include only the effectiveness related to the hybridized
drivetrain (battery and electric motor) and supported accessories.

   On this basis, the agencies projected that mild hybrids would achieve a fleet-level penetration
of 26 percent187 in the cost-minimizing pathway for compliance with the 2025MY standards.

   The EPA Trends Report does not distinguish between mild and strong hybrids in its
accounting of hybrid vehicle penetration.  Therefore it is difficult to separate the relative
penetration of mild hybrids from that of strong hybrids since the 2012 FRM. Although most
analysts had forecast the market share of hybrid vehicles to slowly but steadily rise, hybrid
market share (including mild and strong hybrids) has leveled off at about 3 to 3.5 percent188 of
the total light vehicle market since 2009. According to a report by the International Council on
Clean Transportation (ICCT),189 GM mild hybrid systems accounted for about 2 percent of the
2014 U.S. market, a decline from about 5 percent in 2013.  Other sources have remained
optimistic that penetration levels will eventually grow substantially.  For example, the
automotive supplier Continental has projected market penetration rates of three million BEVs, 12
million strong hybrids and 13 million 48V mild hybrids by 2025.19°

   Like stop-start technology,  mild hybrid technology alters the customary operation of the
engine and so can alter the traditional feel of driving. In many situations the engine may turn off
less frequently and be off for longer periods, although the cycling may appear more random
because it is not necessarily connected to stop and go operation. Some of the effectiveness of
mild hybrids may be diminished by individual driving habits, leading to possible dissatisfaction
with fuel economy. For example, the fuel economy benefit of mild hybrids may fall off more
quickly with aggressive driving due to the lower potential for engine-off operation under these
conditions.

   The 2015 National Academy of Sciences (NAS) report estimated a 10 percent effectiveness
for mild hybrid technology191 based upon the 11 percent fuel consumption reduction observed in
the 2013  GM Malibu Eco. The NAS estimate appears reasonable when considering
improvements in the GM Ecotec engine and six-speed automatic transmission,  and when
considering differences between the vehicle's 0-60 mph acceleration times (which are reported to
be about 7.8 seconds for the base 2013 Malibu LT192 and 8.2 seconds for the 2013 Malibu
Eco193).

   The GM Malibu 15 kW 115V eAssist BISG mild hybrid improved fuel economy about 11
percent over the conventional Malibu Eco 2.5L PFI engine with a six speed transmission.  This
effectiveness figure includes the benefits of other non-hybrid technologies (such as low rolling
resistance tires, underbody aerodynamic panels and radiator grille active shutters) that are
present on the e-Assist mild hybrid package.

   The 2013 GM Malibu Eco's eAssist system uses a 15 kW BISG induction motor with 11 kW
launch assist during heavy acceleration and 15  kW of recuperative braking power.194 The
effectiveness of a 12 to 15 kW electric machine with a liquid-cooled  integrated inverter in a 48V
mild hybrid is comparable to that of a 15 kW motor in 100V+ mild hybrid when taking into
consideration the 30 pound weight reduction from the battery pack and the three, long and heavy
3-phase AC cables used in the 100+V BISG system. For an equivalent mass, 48V mild hybrid
technology effectiveness195 will be slightly less than that of 100V+ mild hybrids.
                                             5-75

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Since the 2012 FRM, the GM eAssist platform has migrated to other vehicles in the GM
lineup.  In February 2016, General Motors announced a limited pilot program offering a version
of its eAssist mild hybrid system on approximately 200 GMC Sierra 1500196 and 500 Chevrolet
Silverado197 2WD pickups in California. This option is offered at a retail price of $500,
significantly lower than the approximately $1000 cost attributed to the 2013 Malibu Eco hybrid
system by an FEV teardown analysis.198 GM credits this system with up to a 13 percent
improvement in city fuel economy.  This development is significant in part because it is the first
example of a BISG system applied to production pickup trucks by a major manufacturer.  GM
stated that it would "monitor the market closely [...] and adjust as appropriate moving forward."
GM is also offering the eAssist BISG mild hybrid as an option to Chevrolet Equinox and GMC
Terrain  midsize SUVs, and Buick Verano, Buick Regal, and Buick Lacrosse.  At least one
analyst expects annual sales of these vehicles to grow to about 100,000 by 2020,188 suggesting
that BISG may become a significant contributor to the compliance path of manufacturers that
rely on this technology.

   The Honda  Civic IMA (Integrated Motor Assist) or PI mild hybrid integrates a 1.5L inline
four cylinder Atkinson cycle engine199 with a CVT transmission and a 17 kW CISG motor to
achieve a 29.7  percent total GHG effectiveness (calculated from two-cycle certification data
comparing the  2015 1.5L Honda Civic IMA to the 2015 1.8L Honda Civic sedan).  The
effectiveness attributable to the mild hybrid technology alone  can be estimated by subtracting the
effectiveness of the other technologies present on the vehicle.  This includes about  1.9 percent
for low rolling resistance tires (LRRT1), 0.7 percent for low drag brakes (LDB), 1.3 percent for
electrical power steering (EPS), 0.7  percent for LUB, 3 percent for use of Atkinson cycle ICP
and DCP,  3.5 percent for use of a CVT, 3 percent for HEG, 0.8 percent aerodynamics and 1.5
percent  for weight difference, resulting in about  13.3 percent GHG effectiveness for this system.
This comparison does not consider the small  0-60 acceleration performance loss (from 9 seconds
to 9.8 seconds) between the  standard 1.8L sedan and the IMA hybrid.

   Combined two-cycle certification test data comparing the 2015 Mercedes-Benz E400
20kW120V P2 mild hybrid and the comparable E350 conventional vehicle indicated about 13
percent  GHG effectiveness.

   In addition to its own benefits, mild hybridization may help enable the use of other
technologies that can further improve efficiency.  For  example, fuel consumption reduction may
approach 20 percent when an electric supercharger is used in 48V mild hybrids combined with
regenerative braking energy recovery, engine downsizing and downspeeding.200 Audi is
expected to market a system utilizing this technology in 2017. As another example, a 48V, 7 kW
electric  supercharger201 has been shown to deliver an extra 40 to 70 kW at the crankshaft by
boosting the engine combustion process. Hence, the electric supercharger may be an effective
accompaniment to engine downsizing and downspeeding.

   The only high-voltage BISG mild hybrid systems currently present in the U.S. market are the
115V Buick Lacrosse eAssist and the 90V 2017 Chevrolet Silverado truck197 mild hybrid  system.
Hyundai is, however, using BISG technology for torque smoothing in its high voltage BISG
Hybrid Starter  Generator (HSG) drivetrain. About 15 percent of the weight reduction in the 2017
Chevrolet Silverado large truck mild hybrid system  was achieved by reducing the battery cell
count from 32  cells to 24 cells, and eliminating three 3-phase AC cables that had previously
connected the battery pack to the motor. EPA estimates the cost of the 90V, 15 kW Silverado
                                             5-76

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
system would be approximately 85 percent of the 115V, 15 kW 2013 Malibu Eco cost projected
by the FEV teardown study ($1045), or about $890.

   A 48V mild hybrid truck was announced in the recent FCA business plan202 for the 2018
Dodge Ram 1500 large truck using next-generation powertrains.203 Schaeffler204 and Hyundai205
also recently demonstrated advanced engineering prototypes of small and mid-size SUV 48V
mild hybrids.

   Compared to 12V systems, high voltage BISG imposes higher costs for the battery pack,
shock protection safety, and active cooling, but with a higher return in effectiveness. For
example, A123 Systems has projected a fuel economy effectiveness of 12 percent for a 48V mild
hybrid system utilizing its 48V battery technology.206  At this level of effectiveness, this system
was described as being more cost effective (at $55 per percent fuel economy gain) than  a full
hybrid solution (at $83).

   To date, most mild hybrids such as the aforementioned Malibu eAssist have been designed to
operate at a voltage of 100V or higher.  However, since the 2012 FRM, evidence has
accumulated to suggest that many functions of a BISG mild hybrid can be provided at a lower
voltage, such as 48V, at significantly reduced costs. Although the effectiveness of 48V mild
hybrids195 will be slightly less than that of higher-voltage mild hybrids (for example, a 48V
system may have a regenerative energy capturing efficiency of about 50 percent207 compared to
perhaps 85 percent for a typical strong hybrid), it can still provide up to  10 to 15 kW of launch
assist and battery charging power. 48V mild hybrid prototype demonstration vehicles from Audi,
Hyundai, Mitsubishi, and Johnson Controls have been described as delivering about 10 to 15
percent CCh reduction and fuel economy improvement.208  Continental, a major Tier 1 supplier
of electrified automotive systems, has presented a prototype small car with a 10 kW BISG 48V
mild hybrid system,  said to provide a 7 percent CCh reduction.209 In the FRM, the agencies
calculated a 7.4 percent GHG effectiveness for small cars equipped with a 10 kW BISG mild
hybrid system, which is comparable to the Continental results.

   Industry appears to be coalescing on a 48V standard for such mid-voltage hybrid applications,
with manufacturers such as Audi, BMW, Daimler, Porsche and VW having initiated a 48V
standard known as LV148.210

   48V mild hybrid technology can also be understood as an alternative to stop-start that is not as
costly as adopting a higher voltage mild hybrid technology. Compared to 12V stop-start, 48V
mild hybrids provide several benefits for a relatively small cost increase,211 such as faster engine
starting, more engine-off time, significant regenerative braking capacity, and better electrical
support for accessories while the engine is off.

   For these reasons, the agencies now expect 48V mild hybrid technology to become more
common than anticipated at the time of the 2012 FRM. The agencies are therefore adding the
48V mild hybrid architecture to this Draft TAR analysis.

   48V mild hybrid technology has received an increasing amount of attention since the 2012
FRM, with  a number of OEMs and suppliers introducing several developmental 48V mild hybrid
systems capable of significant CCh and fuel consumption reductions. At the 2015 SAE Hybrid
and Electric Vehicle Technology Symposium, Controlled Power Technology (CPT) exhibited a
switched-reluctance  motor-generator technology and an electric supercharger for 48V vehicle
                                             5-77

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
electrification. Bosch has presented a 48V mild hybrid system scheduled to be ready for
production by 2017212 that it describes as capable of a 15 percent reduction in fuel consumption.
At the 2015 Consumer Electronics Show (CES), Continental exhibited a 48V mild hybrid system
which consists of a 48V Belt Integrated Starter Generator (BISG) replacing 12V alternator,
DC/DC converter and a 48V lithium-ion battery pack. The BISG motor is an induction motor,
and liquid cooled by engine coolant.  The motor can be decoupled for downhill coasting by
disconnecting the transmission from the engine. Continental expects this 48V mild hybrid
system to begin production in 2016.213 In concert with these introductions, suppliers are also
predicting significant market penetration for 48V systems within the time frame of the rule.
Bosch projected some 4 million 48V mild hybrid vehicles worldwide in 2020, while Eaton
expected up to 3 million 48V mild hybrids globally by 2020.189

   48V mild hybrid technology is estimated to be significantly less expensive than strong hybrid
technology at about 25 percent of the cost. Several advantages of 48V systems contribute to this
lower cost. The voltage is lower than the 60V safety threshold that would otherwise require more
robust electrical shock protection.  The small power levels associated with these components
promotes integration of the inverter with the motor and the  elimination of long stretches of cable,
further isolating the AC portion of the circuit. The relatively small 48V battery pack is
significantly less costly than for a strong hybrid due to its smaller capacity, and may be
composed of fewer cells due to its lower voltage. The battery may not require liquid cooling,
instead being passively cooled with appropriate placement and packaging.  The relatively low
power requirements of a 48V system also promotes use of relatively inexpensive motor
technology (such as induction or switched reluctance) without as strong a concern over NVH or
efficiency.

   Recent developments in the 48V platform have suggested that it is also capable of pushing the
limits of what would be considered a mild hybrid. New P2, P2/P4 and PO/P4 48V system
architectures have been presented by various suppliers such as Bosch, Shaeffler,  Continental, and
Control Power Technologies, ranging from 20 kW to 45 kW of assist capability.190 The
effectiveness for these new, more powerful systems, particularly those on the higher end of the
power range (30-45kW) may approach that of P2 strong hybrids but at a much lower cost. For
example, Bosch has presented a 2nd generation, 48V P2-architecture mild hybrid currently in
development.212 In this 48V P2 system, a more powerful motor-generator is integrated into the
transmission (to create a transmission-integrated starter-generator or TISG architecture). As with
a P2 strong hybrid, the motor can be decoupled from the engine to propel the vehicle in an
electric-drive mode in stop-and-go traffic and for short distances.

   Transcending the BISG format provides a way around common mild hybrid limitations, such
as the 15 kW peak motor power limit, belt efficiency losses, and tandem operation of the engine
with the motor. Stronger formats such as Crank-Integrated Starter Generator (CISG) PI
architecture, as well as Transmission Integrated Starter Generator (TISG) P2 architecture,
overcome the peak motor power limitation in BISG PO mild hybrids and further increase the
potential effectiveness of mild hybrid technology. The Honda IMA CISG PI mild hybrid system
cannot run the electric motor alone without simultaneously  operating the internal combustion
engine,214 while the TISG P2 mild hybrid format allows the engine shut down while the electric
motor works independently  for braking energy recuperation and vehicle propulsion. The
effectiveness of TISG P2 mild hybrids therefore may have higher effectiveness potential than
that of CISG PI mild hybrids.
                                             5-78

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The effectiveness of TISG P2 mild hybrids appears to be higher than that of CISG PI mild
hybrids. GETRAG projected about 15 percent effectiveness for a 48V 21 kW TISG P2 mild
hybrid at the 14th VDI Congress.185 This system employs a 7 speed dual clutch hybrid
transmission, which integrates one common oil circuit for cooling and lubrication, and a
combined e-machine and inverter applicable not only to the 48V 21 kW mild hybrid but also to
other variants such as a 220V+, 50 kW strong hybrid and a 360V+, 110 kW plug-in hybrid
application.  This hybrid transmission also supports other efficiency-enhancing features such as
pure electric driving, extended sailing, more efficient launch assist and brake energy
recuperation, battery charging when the vehicle is standing, and generator-mode/load shift;
features very similar to those provided by strong hybrids.

   Based on a review of these and similar industry developments, as well as data collected from
other sources, the agencies have updated effectiveness estimates for mild hybrid technology.
Updated cost and effectiveness estimates are discussed further in  Sections 5.3 (GHG
Assessment) and 5.4 (CAFE Assessment).

5.2.4.3.3     Strong Hybrids

   In this analysis, strong hybrid refers to hybrid technologies that have higher power capability
and larger battery capacity than mild hybrids, thus providing for more effective management of
power from the internal combustion engine, greater levels of regenerative braking, and more
powerful electric propulsion capable of accelerating the vehicle with less (if any) assistance from
the engine. Strong hybrids provide greater effectiveness than mild hybrids by better optimizing
loading of the engine,  allowing additional engine downsizing, allowing the engine to turn off for
longer periods, and recovering a greater portion of braking energy. These enhanced functions
tend to require higher voltages (as high as 300V to 400V) and more powerful batteries with
greater energy capacity, typically on the order of 1 to 2 kWh.  These attributes add to complexity
due in part to safety requirements associated with  higher voltages and greater battery capacity.
Although strong hybrids are costlier than mild hybrids, they can access a greater degree of fuel
economy and CCh reduction than mild hybrids, and include some of the highest fuel economy
vehicles currently in production.

   Strong hybrids include several distinct architectures.  On a sales-weighted basis, the power-
split hybrid electric vehicle (PSHEV) represents the most common architecture, largely by virtue
of its use for many years in the Toyota Prius hybrid.  This system replaces the traditional
transmission with a single planetary gearset and two motor/generators. The smaller
motor/generator uses the engine to either charge the battery or supply additional power to the
drive motor. The second, more powerful motor/generator is permanently connected to the
vehicle's final drive and always turns with the wheels, as well as providing regenerative braking
capability. The planetary gearset splits engine power between the first motor/generator and the
output shaft to either charge the battery or supply power to the wheels.

   The two-mode hybrid electric vehicle (2MHEV) is a hybrid electric drive system that uses an
adaptation of a conventional stepped-ratio automatic transmission by replacing some of the
transmission clutches with two electric motors that control the ratio of engine speed to vehicle
speed, while clutches allow the motors to be bypassed. This improves both the transmission
torque capacity for heavy-duty applications and reduces fuel consumption and CCh emissions at
highway speeds relative to other types of hybrid electric drive systems.
                                              5-79

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The P2 hybrid is a hybrid technology that uses a transmission integrated electric motor placed
between the engine and a gearbox or transmission, with a wet or dry separation clutch which is
used to decouple the motor/transmission from the engine.  A P2 hybrid would typically be
equipped with a larger electric machine than a mild hybrid system but smaller than a power-split
or 2-mode hybrid architecture. Disengaging the clutch allows all-electric operation and more
efficient brake-energy recovery.  Engaging the clutch allows efficient coupling  of the engine and
electric motor. Based on simulation, when combined with a DCT transmission, the P2 hybrid
architecture provides similar or improved fuel efficiency to other strong hybrid  systems with
reduced cost.

   In the 2012 FRM, P2 hybrid was the only hybrid architecture that was applied in the agencies'
analysis.  Although PSHEV and 2MHEV technology were discussed because they were present
in the market at the time of the FRM, they were not included in the analysis because the industry
was expected to trend toward more cost-effective hybrid configurations such as P2.

   The primary reference EPA used for strong hybrid effectiveness in the 2012  FRM was the
Ricardo modeling study which modeled a P2 with a future DCT. On this basis  EPA estimated an
absolute CCh effectiveness for P2 strong hybrids ranging from 13.4 to 15.7 percent depending on
vehicle class (see 2012 RIA, p. 1-18).  These figures included only the effectiveness related to
the hybridized drivetrain (battery and electric motor) and supported accessories, and did not
include the effect of any accompanying advanced engine technologies. The quoted figures were
based on electric motor sizes assumed in the Ricardo vehicle simulation results  and would vary
with other motor sizes.

   On this basis, the agencies projected that strong hybrids would achieve a fleet-level
penetration of 5 percent215 in the cost-minimizing pathway for compliance with the 2025MY
standards.

   The EPA Trends Report does not distinguish between mild and strong hybrids, nor specific
architectures of strong hybrids, in its accounting of hybrid vehicle penetration.  Therefore it is
difficult to use this source to assess the relative penetration of P2 and other strong hybrid
architectures since the 2012 FRM. However, it is expected that strong hybrids are making up the
majority of the market.

   A recent report by the International Council on Clean Transportation (ICCT)189 reviews
market penetrations for various hybrid architectures. According to this report, the market share
of the P2 hybrid architecture among all hybrids has been relatively small, having grown from
about 9 percent in 2013 to about 12  percent in 2014. Toyota has continued to lead the U.S.
hybrid market with 66 percent of U.S. hybrid sales in 2014. These sales largely account for the
dominance of power-split hybrids in the market. In  the same year, Ford claimed a 14 percent
share of the U.S. hybrid market, also with power-split hybrids. P2 hybrids are primarily
represented in the U.S. market by Hyundai/Kia and Honda, with 8 percent of total 2014 hybrid
sales. The Honda integrated-motor-assist (IMA) architecture represented only 3 percent of the
2014 hybrid market, and is expected to be replaced  by a P2 system in the near future.

   Compared to the more mature, 4th generation power split hybrid architectures of Toyota and
Ford, EPA believes the P2 hybrid architecture is still in a relatively early stage of development
and has yet to be fully optimized.  Manufacturers are continuing to make strides toward
improving this architecture in recently introduced models by refining power electronics and
                                             5-80

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
component efficiency and integrating parts. For example, Hyundai has improved the 2nd
generation Sonata hybrid by fully integrating a 38 kW traction motor and all of the other hybrid
powertrain components within the transmission. The reduced weight has led to improved fuel
economy with reduced costs, as evidenced by the observation that there is no major difference in
effectiveness between this P2 vehicle and the 2015 Toyota Camry power-split hybrid. Going
forward, similar opportunities for major cost reduction and fuel economy improvement are likely
to arise in competing P2 hybrid systems.

  Differences in configuration account for some of the cost and effectiveness differences
between P2 and power-split architectures.  The input power-split architecture requires two
motors, which consist of a small generator and a bigger traction motor which drives through a
simple power-split planetary gear set.  The P2 architecture uses a single, smaller traction motor,
but drives through a more complex conventional transmission gearing. The Honda two-motor
architecture does not use a power-split planetary gear set, and therefore requires a bigger motor
to directly transmit power to the drive axle compared to the typical input power-split hybrid
system.  For example, the Honda Accord 2-motor hybrid uses a 124 kW traction motor216 while
the  Toyota Camry power-split hybrid uses a 105 kW traction motor.217 Highly efficient motor-
integrated DCT transmissions have recently entered production or are under development and are
being adopted in the latest P2 parallel hybrid designs. P2 parallel hybrid architecture also
provides higher towing capacity while the power-split hybrid architecture is limited to less than
3500 pounds trailer towing capacity.

  Even the relatively well-developed power-split architecture continues to show room for
efficiency improvements.  Toyota redesigned the 2016 Prius218 transaxle and motor in its 4th
generation Hybrid Synergy Drive (HSD) to reduce combined weight by 6 percent and volume by
12 percent.  The planetary gear arrangement in the reduction gear has been replaced with parallel
gears, reducing mechanical losses by approximately 20 percent.  The 53 kW main traction motor
is mounted  on a parallel shaft, enabling the transmission case volume to be reduced substantially
while also reducing frictional losses by about 20 percent. The power control unit, which
combines the controller, inverter and DC/DC converter, was attached to the top of the transaxle
and its size  reduced by about 33  percent by eliminating several high-voltage cables. The lithium-
ion battery pack, which is available on the Eco trim level, is 6 percent smaller and 31 percent
lighter than the nickel-metal hydride (Ni-MH) version, while providing the same power output
and degree of hybridization.

  Further evidence that the effectiveness of input power-split hybrids  and P2 parallel hybrids
are  getting closer is shown by the 2017 Hyundai IONIQ P2 hybrid, announced in 2016. The
combined fuel economy of this vehicle, with the GEN2 Hyundai P2 parallel hybrid drive, is
expected to be about 53 mpg, which is comparable to the 52 mpg fuel economy of the 2016
GEN4 Toyota Prius hybrid.  This vehicle  also employs advanced technologies such as a gasoline
direct injection (GDI) inline 4 cylinder Atkinson cycle engine, cooled EGR, CVVT, dual circuit
cooling system, 6 speed dual clutch transmission (DCT), exhaust heat recovery system, and an
intake oil control valve, which act together to increase engine thermal efficiency to as high as 40
percent.

  As reported by ICCT189 (and reproduced here in Figure 5.28), the estimated costs for hybrid
systems have tended to decline steadily in the years after their introduction.  If these trends
                                             5-81

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
continue, significant reductions in hybrid system cost may be expected during the time frame of
the rule.
3D.UUU
$4.500
$4,000
$3.5OO
$3.OOO
$2.5OO
$2,OOO
$1.5OO

$1.0OO
$5OO
$0
2(
\ Prius
\ generation
\ adjustments
^"^••^^powerspllt - FEV
P2 - FEVCS.
^S. P2 - TSD
P2-FEV **C~~-2.5%/year
Hl-power battery £ * * <-. ^ ^5%/year
-2.5%/year ~-S»
Mild - TSD
Mild - Valeof '"••-.,
\ 	 / -5%/year
)OO 20O5 2O10 2015 2O2O 2(
            Figure 5.28 Hybrid System Direct Manufacturing Cost Projection (ICCT, 2015)

   The overall cost of power-split, P2 and two-motor hybrid systems appear to be comparable.
For example, as estimated by an FEV teardown in 2010,219 the power-split hybrid cost of
$2,565220 is only slightly higher than the $2,392 cost estimate for a P2 hybrid system.  EPA is
therefore combining all strong hybrid architectures under the strong hybrid category for the
purposes of this Draft TAR analysis. NHTSA has included both power split and pre-
transmission FtEVs in its analysis. While Atkinson engines were exclusively used for the power
split FIEV,  multiple engine and transmission technologies were included for the pre-transmission
analysis.

   Based on a review of these and similar industry  developments, as well as data collected from
other sources, the agencies have updated cost and effectiveness estimates for strong hybrid
technology. Updated cost and effectiveness estimates are discussed further in  Sections 5.3
(GHG Assessment) and 5.4 (CAFE Assessment).
5.2.4.3.4
Plug-in Hybrids
   A plug-in hybrid electric vehicle (PHEV) is much like a hybrid electric vehicle, but with at
least three significant functional differences.  The first is the addition of a means to charge the
battery pack from an outside source of electricity (usually the electric grid).  Second, a PFLEV
has a much larger battery capacity, and often a greater usable fraction as well.  Finally, it has a
control system that allows the battery to be significantly depleted during normal operation.

   Deriving some of their propulsion energy from the electric grid provides several advantages
for PHEVs. PHEVs offer a significant opportunity to replace petroleum used for transportation
energy with domestically-produced electricity.  The reduction in petroleum usage does, of
                                              5-82

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
course, depend on the amount of electric drive the vehicle is capable of providing under its duty
cycle. PHEVs also provide electric utilities the possibility to increase electric generation during
off-peak periods overnight when there is excess generation capacity and electricity prices are
lower.  Utilities like to increase this base load because it increases overall system efficiency and
lowers average costs. PHEVs can lower localized emissions of criteria pollutants and air toxics
especially in urban areas by operating on electric power. The emissions from the power
generation occur outside the urban area at the power generation plant which provides health
benefits for residents of the more densely populated urban areas by moving emissions of ozone
precursors out of the urban air shed. Unlike most other alternative fuel technologies, PHEVs can
initially use an existing infrastructure for refueling (charging and liquid refueling) so investments
in infrastructure may be reduced.

  Depending on the operating strategy chosen by the manufacturer, a PHEV either provides for
a significant all-electric range (AER) during which the engine does not operate, or provides for
blended operation in which the engine provides some of the propulsion energy while the battery
contributes the remainder.  In this discussion, the former is referred to as a PHEV with AER, and
the latter is referred to  as a blended PHEV.

  In the 2012 FRM analysis, PHEVs were modeled in two configurations, designated PHEV20
and PHEV40 (having 20 miles and 40 miles, respectively, of all-electric range).  Range was
modeled as an approximate real-world range comparable to an EPA label range (specifically, it
was modeled as 70 percent of a projected two-cycle range). Both PHEV configurations were
assigned component sizing consistent with their operation as PHEVs with AER. This tended to
assign a more powerful electric powertrain than would have been required by a blended PHEV,
which is assisted by the engine.

  In the 2012 FRM analysis, EPA assigned PHEVs an effectiveness derived from the SAE
J1711 recommended procedure for accounting for utility factor (the balance between miles
traveled on electricity in all-electric mode and other miles powered by fuel).  On this basis
PHEV20 was assigned an absolute CCh effectiveness of 40 percent, and PHEV40 was assigned
63 percent (see 2012 RIA, p.  1-18). NHTSA modeled a PHEV30 and PHEV 50 with utility
factors of 0.5226 and 0.6887 respectively.

  In the FRM analysis, the cost-minimizing pathway for compliance with the 2025MY
standards did not project a necessity for significant fleet-level penetration of PHEVs (nominally,
zero percent). The analysis did project that some primarily luxury- and performance-oriented
OEMs might include PHEVs  as part of their individual pathway to achieve compliance with
2025MY standards.221

  At the time of the FRM, only a few PHEVs were commercially available in the U.S. market.
The most prominent examples were the Chevy Volt and the Fisker Karma, both of which
debuted as MY2011 vehicles, and the 2012 Toyota Prius Plug-In Hybrid. Production of the
Karma was discontinued in late 2012 as Fisker encountered financial difficulties. Fisker now
belongs to the Chinese company Wanxiang Group and has not resumed significant production to
date.

  Even these early PHEVs demonstrated important differences in their operating strategy that
remain visible in today's market. The Volt and Fisker both offered a significant AER by
including a distinct charge-depleting mode in its operating strategy.  In contrast, the Prius
                                             5-83

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
utilized a more blended mode of operation in which the engine could regularly operate during the
charge depletion stage depending on driving conditions, for example, if the vehicle exceeded a
certain speed or power demand. Both strategies continue to appear in the market today, with
some vehicles emphasizing AER and others emphasizing overall fuel economy in blended
operation. Some PHEVs that employ blended operation are able to achieve an all-electric range
during EPA city and highway test cycles, but may operate in blended mode (using a combination
of gasoline and electricity) when driven more aggressively. Operation in blended mode may be
converted to an equivalent AER by applying a utility factor that considers the contribution of
stored electricity to the total distance traveled in this mode. Therefore both types  of PHEVs are
capable of displacing conventionally-fueled mileage with electrically fueled mileage.

   The 2011 Chevy Volt had an EPA-rated AER of 38 miles, while that of the Fisker was 32
miles.  The Prius was rated at 6 miles AER (11 miles including blended mode).

   Since the FRM, several new models of PFffiV have entered production, with several
additional models announced for future production or otherwise known in the form of concept
cars. Table 5.2 shows a summary of PHEV models that have been in production during the
period since the FRM and their EPA-estimated range (which may include operation in blended
mode).
                                             5-84

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                      Table 5.2 Trends in EPA-Estimated Range of PHEVs

PHEV model
Chevy Volt
Fisker Karma
Toyota Prius Plug-In Hybrid
Ford Fusion Energi
Ford C-Max Energi
Honda Accord PHV
McLaren PI
BMW i3 Rex
BMWiS
Cadillac ELR
Cadillac ELR Sport
Porsche Panamera S E-Hybrid
Porsche 918 Spyder
Mercedes-Benz S550e
BMW X5 xDrive40e
Porsche Cayenne S e-Hybrid
Hyundai Sonata PHEV
Mercedes-Benz C350e
Audi A3 e-tron
Audi A3 e-tron ultra
BMW 330e
Mercedes-Benz GLE 550e
4MATIC
Volvo XC90 T8 Hybrid
Mean AER (not sales weighted)
EPA range (mi)
2012
35
33
11
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-

-
-
-
26.3
2013
38
-
11
20
20
-
-
-
-
-
-
-
-
-
-
-
-
-
-

-
-
-
22.3
2014
38
-
11
20
20
13
19
72
15
37
-
16
-
-
-
-
-
-
-

-
-
-
26.1
2015
38
-
11
20
20
-
19
72
15
37
-
16
12
14
NA
14
-
-
-

-
-
-
24.0
2016
53
-
NL
20
20
-
-
72
15
40
36
16
-
14
14
14
27
18.6*
16
17
14
18*
14
24.4
       Notes:
       NL = vehicle not listed in Fuel Economy Guide
       NA = rating not available in Fuel Economy Guide
       * = approximated from press or manufacturer estimates
   Since the FRM, the continued development and production of PHEVs as evidenced in Table
5.2 has likely been driven in part by manufacturers having chosen to consider PHEVs as part of
their pathway for compliance with the 2017-2025 standards, but even more so by California's
zero emission vehicle (ZEV) regulation. In 2012, CARB adopted increased requirements for
ZEVs and PHEVs through MY2025.  A 2015 National Academy of Science report on PEV
deployment222 cites the California ZEV regulation as being particularly influential in increasing
PEV production and adoption.

   In addition, PHEVs from all manufacturers continue to be eligible for a federal tax credit of
up to $7,500, effectively reducing their net cost to consumers.223'224 This credit applies to the
first 200,000 PEVs (PHEVs and BEVs combined) that are produced by a given manufacturer and
rapidly phases out thereafter. While most manufacturers are unlikely to approach this limit for at
least several years, some of the leading PEV manufacturers such as General Motors, Nissan, and
Tesla are beginning to approach the limit. For example, if the Gen2 Chevy Volt sells well, and
the recently introduced Chevy Bolt EV does also, it is possible that General Motors could reach
the limit by the end of 2017. Strong future sales of the Tesla Model X and Model 3, or the
                                             5-85

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
anticipated 200-mile version of the Nissan Leaf, could cause Tesla and Nissan to approach the
limit by the end of 2018.225  However, in addition to Federal incentives, many states including
California and most of the states that have adopted California's ZEV regulation offer incentives
at the state and local levels.

   It is important to note that most PHEVs are built on global platforms, meaning that economies
of scale for the U.S. market may be driven in part by incentives in other countries. Incentives for
PHEVs in the European Union and China are particularly notable because many manufacturers
that serve the U.S. also serve these markets.

   Trends in PHEV Electric Range

   The electric range of a PHEV (either AER or equivalent AER) is largely a function of the
provided battery capacity. Figure 5.29 shows the relationship between the battery capacity of the
PHEVs in Table 5.2 and their EPA-estimated electric driving range (or the best estimate
available at writing).
                     1C
20      30      40      50      60
     EPA estimated electric range (mi)
                                                                      70
                                                                              EC
          Figure 5.29 Battery Gross Capacity and Estimated AER or Equivalent for PHEVsv

   As the Table and Figure shows, PHEV electric range varies considerably among models.
Among the 2012-2016 PHEVs depicted, two distinct clusters appear, one consisting of longer-
range PHEVs with AER in the vicinity of 35 to 40 miles, and another consisting of shorter-range
vehicles offering between 10 and 20 miles of range (either AER or its equivalent in blended
operation).

   The longer-range cluster consists of various versions of the Chevy Volt and Cadillac ELR
(which shares the Voltec powertrain), the Fisker Karma (at 33 miles), and the BMW i3 (at 72
miles). These are PHEVs with AER that can provide a true all-electric drive mode under a wide
v Range figures gathered from 2012-2016 EPA Fuel Economy Guides.
                                             5-86

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
range of operation. Longer-range PHEVs require a larger battery capacity which tends to
increase their purchase price relative to shorter range PHEVs.

   The shorter-range cluster includes several blended-operation PHEVs. With the exception of
the Toyota Prius PHV (11 miles) and the Ford Energi models (20 miles), these emerged
primarily in the 2015 and 2016 MYs from OEMs that tend to specialize in luxury or high-
performance vehicles.  This suggests that these OEMs are considering PHEVs as a compliance
strategy, as projected in the FRM. For example, when BMW announced the U.S. versions of the
330e and the X5 xDrive40e PHEVs in November 2015, BMW Product Manager Jose Guerrero
was quoted as saying that the timing of introductions such as these "wasn't a competitive impulse
by any manufacturer ... it was an internal impulse that we know that in the future our cars need
to be more efficient, and this is a way ... into that efficiency."226 The Mitsubishi Outlander
PHEV, expected to enter the U.S. market in 2016 as a 2017MY vehicle,227 is also expected to
have an EPA AER in the neighborhood of 20 miles.  These and similar announcements suggest
that a distinct segment of PHEV20-type vehicles is likely to continue in the future as
manufacturers continue to select this lower cost pathway.

   Where new generations of the same model have been announced, the range has in some cases
been increased. For example, the AER of the 2016 Chevy Volt increased from 38 miles to 53
miles.  Going forward,  several OEMs have indicated that second generation PHEV products will
have more AER and more electric power capability, by targeting US06 capability, with minimal
if any reliance on the engine and 30 miles or more of AER.  For example, the FCA Pacifica plug-
in minivan was announced in January 2016 as targeting a 30 mile all-electric range, with
capability to operate all-electric over most operating  conditions.228 Honda is reported to be
considering a 40 mile AER for an upcoming PHEV that would replace the now-discontinued
Honda Accord PHV, which had an AER of only 13 miles.229 Similarly, other manufacturers
including Toyota, GM, and Ford have suggested that their 2017 to 2018 PHEV products will be
targeting at least 30 miles of electric range.

   In such announcements, manufacturers have frequently cited customer desire for an all-
electric driving experience.  As one example, GM appears to credit consumer demand for more
range as part of the impetus for increasing the range of the 2016 Volt. According to Chief
Engineer Andrew Farah, "We listened to our customers ... they were very clear when they told
us that they wanted more range."230 General Motors may have even coined the term "range
anxiety" in order to promote the extended range of the Volt PHEV versus battery-only BEVs.
These manufacturers appear to be responding by increasing the potential for all-electric operation
by increasing electric powertrain power ratings and battery capacity.

   The California Zero-Emission Vehicle (ZEV) program also may be influencing PHEV range.
To qualify as transitional-zero emission vehicles (TZEVs) under the program, PHEVs must
provide at least 10 miles of AER operation on the UDDS drive schedule (as well as meet certain
criteria pollutant standards).231  Since many PHEV manufacturers market in ZEV states as well as
other states, the ZEV program provides a strong incentive for producing PHEVs with AERs
above this threshold.

   Other incentive programs may be encouraging longer PHEV electric range. One example is
the China New Energy Vehicles Program.232 Renewal of this program in 2013 increased the
eligibility requirements for PHEVs to a minimum 50 km (30 mile) AER (under the NEDC cycle)
in order to qualify for purchase subsidies.233  There is some evidence that this may be
                                            5-87

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
encouraging manufacturers of global-market PHEVs to increase AER to at least this level.234
For example, the Cadillac CT6 PHEV was announced in April 2015 at the Shanghai Auto Show,
where it was described as qualifying for the New Energy Vehicles incentives with a range in
excess of 60 km (37 miles). 235 The U.S. version will have the same 18.4 kWh battery pack as
the China version, suggesting that its AER will be similar. As of July 2016, at least one local
U.S. incentive in the state of Washington will also adopt a 30 mile PHEV range requirement to
qualify for a sales tax exemption up to $3,100.236

   Manufacturers have continued to pursue and implement improvements in the efficiency and
cost of battery and non-battery components for PHEVs. One example is the 2016 Chevy Volt, in
which the weight of the battery pack was reduced by 14 kg despite an increase in its capacity
from 17.1 kWh to  18.4 kWh.  The weight of the traction motor was also reduced by 45 kg, and
additional weight and cost were saved by integrating the inverter with the motor and eliminating
long runs of high voltage electrical cable.151'152

   Improvements in component efficiency and road load have both improved performance of
production PHEVs. For example, GM has indicated that the 2016 Chevy Volt improved its
average electric powertrain efficiency over the EPA city and highway cycles by 3  percentage
points (or 4 percent absolute)  compared to the first generation Volt, improving from 86 percent
to 89 percent for the city, and from 84 percent to 87 percent for the highway. Drive unit losses
(including losses of the electric motor, inverter, and transmission) were reduced by 39 percent in
the city cycle and by 35 percent in the highway cycle.237 The Gen2 Volt also provides a good
example of the use of standard road load improvements to increase range in a PHEV.164 Here,
significant changes to the electric propulsion system were accompanied by improvements in
brake drag, reductions in accessory load, and significant improvement of vehicle mass
efficiency.

   In the 2012 FRM, the  agencies envisioned PHEV20 and PHEV40 as representative of PHEVs
that were likely to play a significant role in achieving fleet compliance during the time frame of
the rule. As Table 5.2 and Figure 5.29 show, PHEV20 continues to be represented by several
20-mile and shorter range PHEVs that either continue to be available or have been introduced
since the FRM.  PHEV40 has now been surpassed in real-world range by the 2016 Chevy Volt at
53 miles, and by the BMW i3, which with its range extender option becomes classified as a
PHEV with 72 miles AER.

   EPA and CARB therefore considered replacing PHEV40 with a longer range, such as
PHEV50, in this Draft TAR analysis. Several uncertainties made it unclear as to whether it
would be preferable or useful  to do so.  First, although at least two PHEVs have exceeded
PHEV40, it is also true that other production PHEVs such as the Cadillac ELR and CT6 continue
to fall on the lower side of this line.  Second, if PHEV ranges do in fact increase toward PHEV50
in the future, it is likely to be enabled at least in part by developments other than increased
battery size alone, such as a larger usable capacity, improved powertrain efficiency, improved
battery specific energy, and reduced  road loads. Revising the PHEV40 range would therefore
require that the agencies not simply increase the battery size alone, but also must acquire a full
understanding of the factors that are enabling this increased range in practice, and represent them
accordingly in the battery sizing model.  Because not all manufacturers are likely to be following
the same path, modeling these factors faithfully required careful consideration. For this Draft
                                            5-8

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
TAR analysis, EPA has chosen to retain PHEV40 with a 40-mile label range. NHTSA models
PHEV50 with a 50 mile 2-cycle range.

   In later sections, the agencies will reexamine the 2012 FRM assumptions for other parameters
that affect PHEV battery sizing for this Draft TAR analysis. These include assumptions for
usable battery capacity, electric powertrain efficiencies, battery specific energy, and specific
power of electric machines and power electronics.

   Trends  in PHEV Motor Sizing

   In addition to driving range, the motor power of PHEVs is another important input to the
agencies' projection of battery and system costs for PHEVs. In the 2012 FRM, PHEVs of a given
vehicle class (small car, large car, etc.) were assigned an electric motor power rating (in kW) that
would preserve the same engine-power-to-weight ratio that was observed in conventional
vehicles of that class. This method assumed that the all-electric acceleration of PHEVs relates to
the power  rating of the electric motor in the same way that the engine-powered acceleration of
conventional vehicles relates to the power rating of the engine. However, electric motors differ
markedly from combustion engines, particularly in their delivery of low-speed torque. Electric
motors deliver maximum torque at the lowest end of their speed range, while combustion
engines must develop significant speed to deliver a comparable torque.  This strong low-end
torque allows electric-drive vehicles to deliver high acceleration at low speeds.  This might allow
a PHEV or BEV to deliver acceleration performance similar to that of a conventional vehicle but
with a significantly lower nominal motor power rating than a comparably performing
combustion engine.  At the time of the 2012 FRM, it was unclear to what extent this
phenomenon would influence electric motor sizing in production vehicles, leading to the
decision to assign PHEV motor power based on the nominal power-to-weight ratios of
conventional vehicles.

   The issue of proper motor sizing for PEVs is being revisited for this Draft TAR analysis.
Accurately assigning PEV motor power is important on several fronts. First, the motor power
rating has  a direct effect on the battery power rating, which determines its power-to-energy (P/E)
ratio and its cost. Second, EPA is revising the battery sizing methodology that was used in the
2012 FRM by accounting for the weight of the propulsion motor and power electronics
separately from  the weight of the battery.  This makes an accurate determination of motor power
ratings more critical than before.

   An accurate accounting of motor cost also requires an accurate accounting of motor power.
As in the 2012 FRM, EPA estimates PHEV motor cost as a function of peak power output.  For
more discussion of the decision to scale motor cost to power output, see Section 5.3.4.3.6, Cost
of Non-Battery Components for xEVs.

   Since the 2012 FRM, the number of production PHEVs has increased and provides a much
larger sample size from which some observations may be drawn. Figure 5.30 plots the drive
motor power output ratings and curb weights of moderate- and high-performance PHEVs
produced from MYs 2012 to 2016.
                                            5-89

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
llshed all-«lectric motoring power (kw)
ui o tn o K o t
3OOOOOOC








9]




Fisker Karma
Large Car
and Truck ,0
Porsche^-
8 Spyder

-_(p


-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Figure 5.31 compares the battery capacities of 2012-2016MY PHEVs (from Figure 5.29) to
the battery capacities that were estimated for the 2012 FRM analysis.
                                                                   o  2012 FRM

                                                                   •  MY 2012-2016 PHEVs

                                                                   	Trend ine
                     20     30     40     50     60

                       EPA estimated electric range (mi)
                                                       70
BO
       Figure 5.31 Comparison of 2012-2016MY PHEV Battery Capacities to 2012 FRM Estimates

   For each PHEV range (20 and 40 miles), several values (depicted by the blue circles) are seen
in Figure 5.31 above, corresponding to the FRM estimates for each of the vehicle classes (Small
Car, Standard Car, Large Car, etc.) and several glider weight reductions ranging from 0 percent
to 20 percent.

   It can be seen from the plot that the FRM estimates for PHEV20 battery capacity line up quite
well with the population of production vehicles of a similar range. The FRM estimates for
PFLEV40 also appear to line up fairly well, but show a wider spread and tend to predict a larger
battery capacity per unit range than the trend line would suggest.

   There are several possible reasons the 2012 FRM sizing methodology may have estimated
larger battery capacities for a given range than seen in production. First,  differences in vehicle
weight are not represented in the plot comparison and  could be responsible for some of the
difference in predicted battery capacity for a given range.  Second, it is possible that the
relatively high cost of battery  capacity being experienced in the 2012-2016 time frame
(compared to the agencies' predicted costs for the 2020 time frame) may have caused
manufacturers of some PEVs to apply higher levels of aerodynamic drag reduction,  rolling
resistance reduction, and mass reduction than assumed in the agencies' analysis, in order to save
on battery cost.  The 2012 FRM analysis assumed only a 10 percent reduction in each of
aerodynamic drag and rolling  resistance for battery sizing purposes, with varying levels of mass
reduction. Finally, it is possible that the 2012 FRM assumptions for electric drivetrain
efficiency, usable battery capacity, or other parameters under predicted what the industry has
actually achieved.

   For these reasons EPA is examining the  assumptions used in its battery sizing methodology
and making adjustments where appropriate.  Specific adjustments to the PFLEV battery sizing
                                              5-91

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
methodology used by EPA will be developed and discussed in Section 5.3. NHTSA will be
directly using Autonomie results for battery power and energy, based on multiple sizing
algorithms that were developed and validated in Autonomie to size a wide range of vehicle
powertrains to meet performance requirements.

5.2.4.3.5     Battery Electric Vehicles

   Battery electric vehicles (BEVs) are vehicles with all-electric drive powered by batteries
charged from an outside source of electricity (usually the electric grid).  The analysis conducted
for the 2012 FRM modeled three BEV configurations, designated EV75, EV100 and EV150
(having 75, 100, and 150 miles range, respectively)*. The cost-minimizing compliance pathway
projected a 2 percent fleet-level penetration of BEVs.238

   At the time of the FRM, only a few BEV models had become commercially available in the
U.S. market.  The most prominent examples were the 2011-12 Nissan Leaf and the Tesla
Roadster, which were available nationwide. A few other BEVs were  available in 2012 to very
limited markets or through demonstration programs, such as the BMW Mini E and Toyota
RAV4 EV. Production of the Tesla Roadster was discontinued in early 2012 but was soon
replaced by the  Tesla Model S. Other BEVs available near the time of the FRM were the
Mitsubishi i-MiEV, BYD e6, Coda Sedan, and Ford Focus Electric.

   These early BEVs were designed for different market segments, and showed significantly
different philosophies on the matters  of performance and driving range. Most, such as the Leaf
and Mini E, were designed as moderate-performance vehicles with a driving range of 100 miles
or less, seen as best suited to driving  in urban areas. In contrast, the Tesla Roadster was designed
for a high-performance market segment at a much higher price, allowing it to offer a much
longer range (245 miles by EPA estimate).  Subsequent Tesla vehicles have continued to pursue
similarly aggressive range and performance targets at relatively high purchase prices, while
several other manufacturers continue to define a distinct segment targeting shorter ranges and
moderate performance at lower purchase prices. These two segments will likely continue to
exist within the  time frame of the rule.239'240

   Since the 2012 FRM, several new models of BEV have entered production, with several
additional models announced for future production or otherwise known in the form of concept
cars.  Table 5.3  shows a summary of BEV models that have reached production since the FRM,
and their EPA estimated range.
w As with PHEVs, the indicated range was meant to represent an approximate real-world range comparable to an
  EPA label range (specifically, 70 percent of a projected two-cycle range).
                                             5-92

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                        Table 5.3 Driving Range of 2012-2016MY BEVs

BEV model
Azure Dynamics Transit
Connect
Coda
BYDeS
Toyota RAV4 EV
Mitsubishi i-MiEV
Ford Focus Electric
Tesla Model S (85 kWh)
Nissan Leaf (24 kWh)
Tesla Model S (40 kWh)
Tesla Model S (60 kWh)
Scion iQ EV
Honda Fit EV
Smart fortwo
Fiat 500e
Kia Soul EV
BMW i3 BEV
Chevy Spark EV
Volkswagen e-golf
Mercedes-Benz B-Class ED
Tesla Model S (70 kWh)
Tesla Model S 70D
Tesla Model S 85D
Tesla Model S P85D
Tesla Model S (90 kWh)
Tesla Model S 90D
Tesla Model S P90D
Tesla Model X90D
Tesla Model X P90D
Nissan Leaf (30 kWh)
EPA range (mi)
2012
56
88
122
103
62
76
265
73
-
-
-
-
-
-
-
-
-
-
-

-
-

-
-
-
-
-
-
2013
-
88
127
103
62
76
265
75
139
208
38
82
68
87
-
-
-
-
-

-
-

-
-
-
-
-
-
2014
-
-
127
103
62
76
265
84
-
208
-
82
68
87
-
81
82
NA
87

-
-

-
-
-
-
-
-
2015
-
-
127
-
NL
76
265
84
-
208
-
-
68
87
93
81
82
83
87

240
270
253
265*
270*
253*
NA
-
-
2016
-
-
-
-
62
76
265
84
-
-
-
-
68
84
93
81
82
83
87
234
240
270
253
265*
294
270
257
250
107
       Notes:
       NL = vehicle not listed in Fuel Economy Guide
       NA = rating not available in Fuel Economy Guide
       * Manufacturer applied 85 kWh EPA range figure for EPA labeling purposes
   Since the FRM, the continued development and production of BEVs as evidenced in the
above table has likely been encouraged in part by several regulatory factors.  The 2017-2025
GHG/CAFE regulation assigns a high GHG effectiveness to BEVs, further enhanced by
assigning 0 g/mi for upstream emissions and a multiplier for the earlier years of the rule.  Some
manufacturers have therefore chosen to consider BEVs as part of their pathway for compliance
with the 2017-2025 standards.  Production of BEVs also generates GHG credits that may be used
for regulatory compliance or sold to other manufacturers. Production of BEVs can also assist
manufacturers in meeting fleet average criteria pollutant regulations such as EPA's Tier 2 and
Tier 3 standards or CARB's LEV II and LEV III standards. And, just as with PHEVs,
California's ZEV regulation continues to drive BEV production to generate ZEV credits as
manufacturers prepare for ever increasing requirements through 2025 model year.
                                             5-93

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   In addition, BEVs from all manufacturers continue to be eligible for a federal tax credit of up
to $7,500, effectively reducing their net cost to consumers.223'224  Because this credit applies to
the first 200,000 eligible vehicles (BEVs and PHEVs combined) produced by a given
manufacturer, it is likely to continue to influence the BEV market for some time. At current
rates of production, it is possible that some manufacturers may begin approaching the 200,000
limit by 2018, with others following soon after.225  In addition to the Federal tax credit, many
states, including California and many of the states that have adopted California's ZEV regulation
offer incentives for ZEVs at the state and local levels.

   BEVs continue to be offered at a significant price premium to  conventional vehicles, largely
due to the cost of the battery, as well as non-battery components that have yet to reach high
production volumes. Some BEVs, particularly those targeted primarily for sale in the ZEV
states, are available for purchase only in those states. Despite the higher purchase price and
limited availability, BEV production levels have grown significantly since the FRM.

   Through November 2015, Nissan had sold about 88,000 Leaf EVs, and GM had sold about
90,500 Volt PHEVs and Spark EVs combined. Analysts have widely speculated that a slight
decline in PEV sales in MY2015 (relative MY2014) is due at least in part to anticipation of new
models with longer range and enhanced features. For example, expectations of a refreshed
version of both the 2016 Volt and 2016 Leaf existed long before either became available. The
2016 Leaf offers a larger 30kWh pack, increasing range significantly, while the 2016 Volt also
offers a longer range, better fuel economy and other enhancements such as improved seating.

   The demand for high-end BEVs, such as those produced by Tesla Motors, has accounted for a
significant portion of this production despite their high purchase price. These vehicles compete
in a market segment with other high-priced vehicles and are seeing success in that segment.  This
suggests that a demand for BEVs exists relatively independently of the regulatory factors that are
largely oriented toward the broader automotive market.  If the performance attributes that are
attracting this segment of buyers can be sufficiently retained at a lower price point, this could
further drive demand for BEVs in the future.

   Trends in BEV Driving Range

   Growth in the BEV market since the 2012 FRM has greatly expanded not only the available
choice of vehicle models and trims, but also the available driving ranges. BEV driving range is
largely a function of battery capacity. Figure 5.32 shows the relationship between the battery
capacity of the BEVs in Table 5.3 and their EPA estimated driving range.
                                             5-94

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                          50
100        150        200
  EPA estimated range (mi)
250
300
              Figure 5.32 Battery Gross Capacity and EPA Estimated Range for BEVsx

   It has become apparent since the FRM that manufacturers have been pursuing increased
driving range.  Several examples serve to illustrate this trend. The Nissan Leaf was introduced in
2011 with an EPA-rated range of 73 miles.  The 2013 model increased this to 75 miles, while
2014 and later models earned a higher rating of 84 miles by eliminating  a partial charge option,
allowing the range to be evaluated at 100 percent charge.  This trend indicates that Nissan
perceives increased range as a desirable goal. As another example, in January 2016, it was
reported that the range of the BMW i3 might increase by about 50 percent due to improved
battery chemistry  and electronics, to approach perhaps 120 miles.241 In May 2016, BMW
announced that the range would be approximately 114 miles, due in part to a 50 percent increase
in cell capacity.242 In January 2016, Volkswagen also indicated that a new version of the e-Golf
could expect a possible 30 percent increase in range over the current model (or about 108 miles)
due to an increase in cell capacity from 28 A-hr to 37 A-hr.243 The 2017 Ford Focus BEV is also
expected to increase its range to 100 miles compared to its original range of 76 miles.244  The
2017 Hyundai loniq BEV also targets a range of more than 100 miles.245

   A trend toward increased range also seems to be playing out across manufacturers, as new
products are introduced to compete in the market.  For example, the Kia Soul EV was introduced
in 2014 with a range of 93 miles, surpassing the Leaf.  Not long after in  2015, Nissan announced
the 2016 Nissan Leaf, offering an EPA range of 107 miles with a new 30 kWh battery pack.

   Future vehicles expected to enter the market in the relatively near term (2016-2017) have
increasingly targeted even longer ranges. Both the Chevy Bolt (expected to debut as  a MY2017
vehicle) and a future version of the Nissan Leaf have been described by  their manufacturers as
targeting a 200 mile range.  As of April 2016, the Tesla Model 3 is being described as offering a
215 mile range and entering production in late 2017.246

   Even Tesla Motors, which already  offers a range in excess of 200 miles in its current vehicles,
has shown an interest in increased range as evidenced by regular increases in battery capacity.
 : Range figures gathered from 2012-2016 EPA Fuel Economy Guides.
                                              5-95

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
After announcing in 2012 that the Tesla Model S would be available in three battery sizes (40
kWh, 60 kWh, and 85 kWh), the 40 kWh version was canceled in 2013, prior to its production.
In April 2015, the battery capacity of the 60 kWh version was increased to 70 kWh, which along
with powertrain improvements increased its range from 208 miles to 240 miles. In September of
the same year the 85 kWh version was increased in capacity to 90 kWh by use of an improved
chemistryY.  Tesla also announced in 2015 an available battery upgrade for the discontinued
Roadster that would increase its range by about 40 percent.

  Manufacturers have frequently cited customer demand in the quest for increased range.  When
the 40 kWh Model S was canceled, Tesla attributed the decision to low demand, further saying,
"Customers are voting with their wallet that they want a car that gives them the freedom to travel
long distances when needed."247 Although this statement clearly promotes Tesla's market
strategy of offering a longer driving range than other BEV-manufacturing OEMs, similar
sentiment has been expressed by other OEMs in marketing their electrified vehicles or
announcing future plans.  Customer demand for an affordable BEV with  a longer driving range
than currently available is implicit in the 200-mile range target of both the future Nissan Leaf
and the 2017 Chevy Bolt.

  Obviously, one way for an OEM to increase range is to increase the battery capacity.  Simply
increasing the battery capacity in the absence of other improvements may be prohibitive because
it increases the cost of the vehicle accordingly. On the other hand, improved battery
manufacturing or battery chemistry (in terms of cost or energy density) might enable a larger
capacity while offsetting some  of the cost penalty. For example, both Tesla and Nissan have
utilized  improved chemistry to increase  capacity within the existing footprint of their respective
packs; while GM and Nissan have hinted strongly at improved chemistry being the enabler of the
affordable 200-mile range target for the  Bolt and future Leaf.  These and other examples are
discussed in more detail in Section 5.2.4.4.1.

  Increasing the usable capacity (i.e. widening the usable state-of-charge window) of the battery
may also be possible; for example, by use of an improved chemistry, or by acting on experience
that indicates that the existing buffer capacity may be reduced. Improvements in battery
management systems (BMS) may also lead to greater utilization of the available battery capacity.
Examples of OEM activity in this area are reviewed in more detail in Section 5.2.4.4.3

  Range can also be increased by reducing vehicle energy consumption. This can be  done by
improving the energy efficiency or weight of non-battery powertrain components (electric
machines and power electronics) or even the battery itself. For example, the dual motor versions
of the Tesla Model S achieve a slightly higher range than the single motor versions due to an
improved powertrain efficiency resulting from the ability to selectively operate one or both
motors as conditions warrant. Range may also benefit from standard road load improvements
such as light-weighting, improved aerodynamics, and lower rolling resistance.

  In addition to increased range, a larger battery may carry other ancillary benefits for
manufacturers and consumers.  Because a large battery stores more energy per charge cycle than
a small battery, it is likely to experience fewer charge-discharge cycles in the course of providing
Y The manufacturer chose to apply the 85 kWh EPA range figure to the 90 kWh version for EPA labeling purposes.
  Marketing materials attribute an additional 6% range to the 90 kWh version.
                                              5-96

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
a given number of vehicle miles. For example, a battery that provides for a range of 200 miles
can provide a lifetime mileage of 150,000 miles with about 750 charge-discharge cycles, while a
100-mile battery may require 1,500 cycles. The smaller number of expected cycles may promote
a longer battery lifetime or relax manufacturer provisions for battery durability, such as
increasing the permissible charge rate or the usable capacity. A larger battery might also
experience a much shallower average state-of-charge (SOC) swing in the course of meeting its
mileage target, with similar implications for durability. Another advantage of a large battery is a
reduction in average discharge rate (C-rate), which can allow consideration of chemistries and
configurations that would not be suited to smaller batteries. For example, Tesla may have
selected a chemistry that supports notably low C-rates in recognition that the large size of the
battery acts to minimize per-cell power requirements.248 Of course, a drawback of a larger
battery over a smaller battery is its greater weight, which tends to reduce the overall energy
efficiency of the vehicle.

   In the same way that cabin air conditioning can have a significant impact on fuel economy of
conventional vehicles,249 both heating and air conditioning can have a strong impact on BEV
energy efficiency and range.  While the impact of passenger comfort on range can be great for
both BEVs and PHEVs, BEVs are at a particular disadvantage because all energy for heating and
cooling must  come from the battery. In contrast, PHEVs may choose to operate the engine if
needed (for example, the Chevy Volt operates the engine to help with cabin heating in cold
weather).  Cabin heating and cooling for BEVs is therefore an active area of research toward
increasing BEV range.250'251

   Some BEVs, such as the Nissan Leaf, have employed heat pump-based HVAC in place of
resistive heating.  When the temperature differential between the outside air and the desired
cabin temperature is not too large, this method can be much more efficient than resistive heating
at controlling cabin temperature. Another approach to passenger comfort that has been used for
BEVs and PHEVs involves heated and cooled surfaces, for example, the steering wheel and
seats, instead of or in addition to heating the cabin air, which one study has shown can reduce
cooling and heating energy in a PHEV by about 35 percent.252 Pre-conditioning the passenger
cabin while plugged in to a charging station is yet another approach, which  can reduce the use of
onboard energy for heating and cooling (although it does consume energy at the station).

   Modeled BEV Ranges in the 2012 FRM and this Draft TAR

   As previously noted, the FRM analysis modeled three BEV range configurations (EV75,
EV100 and EV150). At the time of the 2012 FRM, the agencies envisioned EV150 as the
maximum BEV range that was likely to play a significant role in achieving fleet compliance
during the time frame of the rule. Since that time, EV150 has been surpassed by several longer-
range vehicles that are under production or recently announced.  Tesla vehicles with a range well
in excess of 200 miles were already in production at the time of the FRM, and have since
continued to grow in range and market share. Although these vehicles currently constitute a
luxury segment that may not be fully representative of a mass-market vehicle, their success at
achieving significant market penetration shows that at least one OEM has found it preferable to
comply with the 2017-2025MY standards and generate additional GHG credits by producing an
EV200 instead of an EV150.  Announcements from Nissan and GM that target a 200-mile range
in BEVs to be produced as early as 2016 also suggest that EV200 may be a more  accurate
representation of the higher-range BEV segment than EV150. Therefore, based on the current
                                             5-97

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
direction of the industry, the agencies have replaced EV150 with EV200 in this Draft TAR
analysis.

   It is uncertain whether adoption of EV200 in place of EV150 is likely to have a significant
impact on the projected cost-minimizing pathway for fleet compliance with the 2017-2025
standards. There is limited potential for either EV150 or EV200  to be selected by OMEGA and
the Volpe model as part of a cost-effective compliance path, because the relatively high cost of
the larger battery is likely to overshadow any gain in effectiveness. That is, since EV75, EV100,
and EV150/200 all are assigned a GHG effectiveness of 100 percent (with upstream emissions
assessed at 0 grams per mile), the incremental cost of EV150/200 vs. EV75 or EV100 strongly
discourages its selection on a pure cost-effectiveness basis. On the other hand, adopting EV200
has the advantage of more accurately reflecting the evolving electrified vehicle fleet.

   In adopting EV200, the agencies gave careful consideration to the resulting implications for
the battery sizing and costing methodology. The increase in range from 150 to 200 miles had to
be modeled in a way that accounts for how manufacturers would be expected to achieve the
incremental range. In addition to increasing gross battery capacity, manufacturers would likely
also rely on other changes to better utilize the available capacity, perhaps by increasing the
usable capacity, improving powertrain efficiency, improving battery specific energy, and
reducing road loads. Many of the refinements to the battery sizing methodology that are
discussed in Section 5.3 resulted from this effort to faithfully represent the paths available to
manufacturers to achieve EV200.

   In Section  5.3 EPA is reexamining the 2012 FRM assumptions for all xEV parameters that
affect battery  sizing for this Draft TAR analysis. These include assumptions for usable capacity,
electric powertrain efficiencies, specific energy of the battery, and specific power of electric
machines and power electronics.  Also, because the cost effectiveness of standard road load
improvements is greater for BEVs than for conventional vehicles and even other xEVs (due to
the potential to save on battery cost), EPA is also reexamining the assumptions for road loads as
they affect battery sizing for BEVs.  In  addition, NHTSA will be reassessing the battery and
electric machine performance parameters based on available literature and vehicle test data from
the ANL APRF.

   Trends in BEV Motor Sizing

   In addition to driving range, the motor power of BEVs is another important input to the
agencies' projection of battery and system costs for BEVs. As discussed previously with respect
to PHEVs, the 2012 FRM analysis assigned BEVs of a given vehicle class a motor power rating
that would preserve the same engine-power-to-weight ratio observed in conventional vehicles of
that class. This method assumed that the electrically-powered acceleration of BEVs relates to
the power rating of the electric motor in the same way that the engine-powered acceleration of
conventional vehicles relates to the power rating of the combustion engine. However (as
discussed in the PHEV  section previously),  electric motors differ markedly from combustion
engines in their delivery of low-speed torque, delivering maximum torque at the lowest speeds,
while combustion engines must develop significant speed to deliver a comparable torque. This
might allow a BEV to deliver acceleration performance similar to that of a conventional vehicle
while using a significantly lower nominal motor power rating than a comparably performing
combustion engine. At the time of the 2012 FRM, it was unclear to what extent this
                                             5-98

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
phenomenon would influence BEV propulsion motor sizing, leading to the decision to assign
BEV motor power based on the nominal power-to-weight ratios of conventional vehicles.

   As previously discussed in relation to PHEVs, the issue of proper electric motor sizing for
BEVs is being revisited for this Draft TAR analysis. Accurately assigning the motor power of a
BEV is important for several reasons. First, the motor power rating has a direct effect on the
battery power rating, which determines its power-to-energy (P/E) ratio and its cost. Second, the
agencies have revised the battery sizing methodology to account for the weight of the electric
motor and power electronics separately from the energy content of the battery.  This makes an
accurate determination of motor power ratings more critical than before. Finally, an accounting
of motor cost requires an accounting of motor power. As in the 2012 FRM analysis, EPA
estimates electric motor and power electronics costs as a function of peak output power, in
accordance with several examples of similar industry practice. For more discussion of the
decision to scale motor cost to power output,  see Section 5.3.4.3.6, Cost of Non-Battery
Components for xEVs.

   Since the FRM, the number of production BEVs has increased and provides a much larger
sample size from which to draw observations. Figure 5.33 plots the motor power ratings and
curb weights of BEVs produced from MYs 2012 to 2016.
- 3OO
I
rzso
o
°- 200
1
| 150
•o
ii
:§ loo
-Q
3
°- 50
n



C





Bl
levy Spark E\
^
0
»



riWiS
0
<\
A<
&>*• :
*
Tesla 60
1
IP *
o°° •
„>
f
»

%*

Merced

•

Tesla 85

es B250e



                                                                       o 2012 FRM

                                                                       • MY2012-2016 BEVs
                500       1000      1500      2000
                               Curb weight (kg)
2500
3CCC
     Figure 5.33 Comparison of Motor Power of 2012-2016MY Production BEVs and FRM Estimates

   In the Figure, the solid orange dots represent the motor power ratings and curb weights of
production BEVs (excluding the highest-performing Tesla vehicles in excess of 350 kW)
produced for MYs 2012-2016. The blue circles represent the motor power ratings and weights
assigned to BEVs of various ranges and classes in the 2012 FRM. The chart suggests that the
FRM assigned significantly higher BEV motor power ratings than the majority of BEV
manufacturers have actually provided. Among moderate-performance vehicles, the BMW i3 and
the Chevy Spark EV have motor power ratings that are closest to the levels assumed in the FRM.
Even the higher-market Mercedes B250e is at a lower power-to-weight ratio than the FRM
would have assumed.
                                             5-99

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Based on this analysis and a new power-to-weight study described in Section 5.3, EPA has
revised the BEV motor power ratings assumed for this Draft TAR analysis.  The analysis will
therefore adopt power ratings closer to those suggested by the power-to-weight ratios that BEV
manufacturers appear to be following, while maintaining an estimated acceleration performance
equivalent to conventional vehicles. As with PHEVs (discussed in the previous section),
assigning a more accurate power rating will allow greater fidelity in the projected cost of both
the battery and non-battery components of BEVs. Specific proposed adjustments to BEV motor
power sizing are developed and discussed in Section 5.3. NHTSA will be directly using
Autonomie results to assign the power of electric motors for BEVs. Multiple sizing algorithms
have been developed and validated in Autonomie to size a wide range of vehicle powertrains to
meet specific vehicle performance.

   Trends in BEV Battery Sizing

   The 2012 FRM analysis employed a battery sizing methodology to assign battery power
ratings and energy capacities for modeled BEVs. Now that a number of BEVs are on the market
and have been rated for range by EPA, it is informative to compare the FRM projections of BEV
battery capacity and range to the BEVs that have entered the market for MYs 2012-2016. Figure
5.34 shows the  range and battery capacity plot of Figure 5.32, annotated with the assumed
battery capacities and ranges used in the FRM.
                                                                   •  MY 2012-2016 BEVs

                                                                   O  21012 FRM

                                                                  	Trendlne
                          100      150      200
                          EPA estimated range (mi)
                                                    250
                                                            300
       Figure 5.34 Comparison of 2012-2016MY BEV Battery Gross Capacities to FRM Estimates

   The FRM modeled batteries for EV75, EV100, and EV150 at several assumed glider weight
reductions ranging from 0 percent to 20 percent. For each BEV range (75, 100, and 150 miles),
several values are seen in Figure 5.34 above, corresponding to each of the vehicle classes (Small
Car, Standard Car, Large Car, etc.) and glider weight reductions of 0 percent and 20 percent.

   Following trends seen in the MY2008 fuel economy data, the 2012 FRM modeled large BEVs
such as Large Car and Small/Large MPV with substantially larger power-to-weight ratios and
                                            5-100

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
significantly higher fuel consumption compared to smaller vehicles such as Small Car and
Standard Car. This led to significantly different battery capacity projections for a given range,
obscuring the comparison to observed MY2012-2016 BEVs. In order to assess how well the
2012 FRM technique predicted battery sizes for each class, it is therefore necessary to consider
the larger and smaller vehicle classes separately. Because the vehicle classes in the Fuel
Economy guide, from which the range data is taken, are different from the six vehicle classes
used in the FRM, only an approximate comparison can be made by dividing the fleet into a group
of smaller-to-moderately sized vehicles and a group of larger vehicles.

   Figure 5.35 shows data for small and moderately sized passenger cars, which in the FRM
would be classed as Small Car and Compact Car, and in the Fuel Economy guide are classed
variously as Minicompact, Subcompact, Compact, and Midsize (importantly, the Nissan Leaf is
classed as Midsize and so is included in this group).
I
S
t
II
      ice

      3C

      EC

      7C

      ec

      5C
      2Z

      1C
                                                                     • MY 2012-2016 BEVs

                                                                     ;; 2012 FRM
                  50        100       150       200
                              EPA label range (mi)
                                                  250
300
Figure 5.35 Comparison of 2012 FRM-Projected Battery Capacity to MY2012-2016 BEVs (Smaller Vehicles)

   This plot shows that for smaller BEVs, the FRM projections of battery capacity appear to fit
reasonably well with MY2012-2016 BEVs. Even so, there is a tendency for the 2012 FRM
projections to have somewhat overestimated the battery capacity that manufacturers have
provided for these vehicles.

   Figure 5.36 shows data for larger cars and SUVs, which in the FRM were classed as Large
Car, Small MPV and Large MPV, and in the Fuel Economy guide are classed variously as Large
Car, Small SUV 2WD, and Standard SUV. Variations of the Tesla Model S are classified as
Large Car and so represent the bulk of production examples shown in the plot.
                                             5-101

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                                                      • MY 2012-2016 BEVs

                                                                      C2012FRM
                           100       150      200
                              EPA label range (ml)
250
300
 Figure 5.36 Comparison of 2012 FRM-Projected Battery Capacity to MY2012-2016 BEVs (Larger Vehicles)

   This plot shows that for larger BEVs, the tendency for the 2012 FRM methodology to
overestimate battery capacity is much stronger for all except the shortest ranges (which  are
largely not present in the market). A trend line of the FRM projections is not only at a higher
level but also appears to diverge substantially as the range increases. Although the FRM did not
project ranges beyond 150 miles, it appears that the 2012 FRM battery sizing methodology
would fail dramatically at estimating battery capacity for 200-plus mile BEVs, such as for
example the Tesla models depicted at the far right of the plot, and even the EV200 configuration
the agencies have adopted to replace EV150.

   As discussed with reference to PHEVs in the previous section, there are several possible
reasons the 2012 FRM battery sizing methodology may have estimated larger battery capacities
for a given range than seen in production. First, differences in vehicle weight are not represented
in the plot comparison,  and could be responsible for some of the difference.  Second, it is
possible that the relatively high cost of battery capacity being experienced in the 2012-2016 time
frame (compared to the agencies' predicted  costs for the 2020-2025 time frame) may have caused
manufacturers of some BEVs to apply higher levels of aerodynamic drag reduction, rolling
resistance reduction, and mass reduction than assumed in the agencies' analysis, in order to save
on battery cost.  The 2012 FRM analysis  assumed only a 10 percent reduction in each of
aerodynamic drag and rolling resistance for battery sizing purposes, with varying levels of mass
reduction. Finally, it is possible that the 2012 FRM assumptions for electric drivetrain
efficiency, usable battery capacity, or other parameters under predicted what the industry has
managed to achieve.

   For these reasons, EPA has examined the assumptions used in the BEV battery sizing
methodology and made adjustments where  appropriate.  Specific proposed adjustments  to the
BEV battery sizing methodology are discussed in Section 5.3. As noted earlier, NHTSA will use
Autonomie results for BEV batteries, based on multiple sizing algorithms that have been
                                             5-102

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
developed and validated in Autonomie to size vehicle powertrains to meet specific vehicle
performance targets.

5.2.4.4 Developments in Electrified Vehicle Battery Technology

   For many types of electrified vehicles, particularly PHEVs and BEVs, battery cost is the
largest single component of vehicle cost. Battery pack cost is determined in part by the
configuration of the pack, which should be tailored to the specific performance goals of the
vehicle.

   Pack configuration may be decomposed into a large number of primary design parameters
which the vehicle designer can specify to determine the performance of the pack and ultimately
its cost. In configuring a pack, the primary performance targets are energy capacity in kilowatt-
hours (kWh) and power capability in kilowatts (kW). These performance targets are determined
by design choices such as: battery chemistry (Li-ion is composed of a number of closely related
but differently performing chemistries); pack voltage, usable portion of total capacity, cell
capacity (Ampere-hours per individual cell), cell topology (the  electrical and physical
arrangement of cells and modules in the pack), and cooling method (passive or active, and air or
liquid), among others. Further, for a pack defined by a given set of these design parameters, the
assumed annual manufacturing volume will also influence the projected cost.

   It is customary to refer to battery cost in terms of cost per kWh.  However, in order to make
valid comparisons on this basis it is important to understand that cost per kWh is strongly
influenced by the power-to-energy (P/E) ratio of the battery.  Intuitively, a BEV battery
optimized for energy storage capacity (low P/E) will have a low cost per kWh because the
materials and construction are oriented toward providing maximum energy capacity.
Conversely, an HEV battery optimized for power (high P/E) will have a higher cost per kWh,
because the materials and construction are oriented toward providing power, while the metric of
cost per kWh continues to focus on energy. For these reasons, cost per kWh figures derived
from energy-optimized BEV or PHEV battery packs should not be used to estimate the cost of a
power-optimized HEV pack, or vice versa. Comparisons of cost per kWh are only valid when
the applications have a similar P/E ratio.

   It is also important to be aware of whether a quoted cost per kWh is on a cell basis or a pack
basis. Figures found in press literature may be of either type. Costs quoted on a cell basis will
be much lower than for a full pack that includes battery management, disconnects, and thermal
management. In the 2012 FRM and this Draft TAR, all cost per kWh figures are presented on a
pack basis.

   Finally,  the energy capacity of a battery pack (kWh) may be characterized either by gross
capacity or usable (net) capacity.  Gross capacity, also known as nominal or nameplate capacity,
is the total amount of energy that  can be reversibly stored in a complete charge and discharge
cycle of the battery, without regard to long term durability. It is a relatively fixed quantity that is
a function of the amount of electrode active materials contained in the battery.  Usable capacity
is the portion of gross capacity that the manufacturer believes can be regularly used in an
application while maintaining a desired level of durability.  Although usable capacity is the
metric that  relates best to performance attributes such as driving range, usable capacity varies
widely among different vehicle types and individual models of each type.  For consistency it has
                                             5-103

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
become customary to refer to the size of xEV battery packs by their gross capacity, and to refer
to battery cost per gross kWh. The 2012 FRM and this Draft TAR follow this standard.
5.2.4.4.1
Battery Chemistry
   In the 2012 FRM, the agencies based their battery cost analyses on outputs of the then-current
version of the ANL modeling tool BatPaC137, which models several well established lithium-ion
chemistries. As shown in Table 5.4, the choice of chemistries available in BatPaC included:

                Table 5.4 Lithium-ion Battery Chemistries Available in ANL BatPaC
Chemistry
LMO-G
LMO-LTO
NMC333-G
NMC441-G
NCA-G
LFP-G
Cathode
Lithium-Manganese Oxide
Lithium-Manganese Oxide
Nickel-Manganese-Cobalt (3-3-3)
Nickel-Manganese Cobalt (4-4-1)
Nickel Cobalt Aluminate
Lithium-Iron Phosphate (Olivine)
Anode
Graphite
Lithium Titanate Oxide
Graphite
Graphite
Graphite
Graphite
   Certain chemistries are better suited for certain types of packs than for others.  For example,
the specific versions of NMC chemistry that are modeled by BatPaC are well suited for packs
having a large energy capacity such as BEV packs, but due to limits on area specific impedance
(AST), they are not as well suited for small, power-dense packs for HEVs.  Considerations such
as these ultimately led to the chemistry choices employed by the agencies in the FRM.  BEV and
PHEV40 batteries were configured with NMC441-G, while PHEV20 and HEV packs were
configured with LMO-G.

   Since the 2012 FRM, the lithium-ion family of chemistries has continued to dominate xEV
battery technologies seen in current and announced production vehicles. As expected,
NMC/NCM cathode formulations are increasingly being seen in BEVs announced since the
FRM, including in mixed formulations with LMO. For example, the Kia Soul BEV uses an NCM
cathode.253 In the 2015 NAS report (p. 4-26), the committee mentions the use of NMC cathodes
for the 2020-2025 time frame, lending further support to the agencies' choice. PFLEVs and FLEVs
are being seen not only with LMO-dominant cathode formulations, such as in the original  Chevy
Volt, but also with NMC and blended NMC cathode formulations, as in the 2016 Chevy Volt,254
the Ford C-Max Hybrid FLEV and C-Max Energi PFLEV.255  These are presumably optimized for
the relatively high P/E ratio of these applications. Lithium-iron phosphate cathodes are also
being promoted for HEV use.256 While it is not possible for BatPaC to model every (often
proprietary) variation in cathode formulation, the available choices are likely sufficient to
represent the cost spectrum applicable to this family of chemistries.

   Since the FRM, xEV batteries have trended away from pure LMO cathodes toward blends of
NMC with LMO. 257 In particular, most HEV batteries currently in production appear to utilize
either NMC or LMO blended with NMC. For example, the 2016 Chevy Malibu Hybrid battery
is said to use an NMC cathode258 while the Volt uses NMC blended with LMO.254 This contrasts
with the agencies' assumption of LMO chemistry for HEV and PHEV20, which was the result of
the limited number of high-power chemistries modeled by earlier versions of BatPaC.

   Version 3 of BatPaC, released for beta in November 2015, includes additional cathode
chemistry  options, including the more common NMC622 in place of NMC441, and a user-
                                            5-104

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
selectable blend of NMC and LMO. The blended cathode option will allow the agencies to
consider a blended NMC-LMO cathode for HEV and PHEV20 batteries, to better represent their
usage in existing platforms.

   At the time of the 2012 FRM, practically every production xEV was using a Li-ion chemistry,
with the nickel-metal-hydride (NiMH) battery of Toyota HEV products being the primary
exception.  After using NiMH in the Prius since its introduction in 1997,  there are signs that even
the Prius may be moving toward Li-ion. By 2012 Toyota had already adopted a lithium-ion
chemistry for the Prius PHEV, a platform which requires a larger battery capacity than the
standard hybrid. In  October 2015, Toyota announced that the 2016 Prius hybrid would also
begin offering a Li-ion battery as an option.155'259

   Since the 2012 FRM, industry research has continued into more energy- and power-dense
variations of the lithium-ion platform, including  improved cathode material blends, lithium-rich,
manganese-rich, nickel-rich, and higher voltage (e.g. 5V) spinel cathodes, and the use of silicon
in the anode. Other research is concerned with even more advanced platforms, including
lithium-sulfur, and several metal-air chemistries  (lithium-air, aluminum-air and zinc-air) among
others.  These advanced chemistries are not yet available in cells suitable for xEV use, but
potential examples are beginning to emerge.

   Lithium-sulfur (Li-S) cells are beginning to be seen in some highly specialized applications.
A Li-S cell manufactured by Sion Power is used in the Airbus-sponsored Zephyr high-altitude
unmanned aerial vehicle (UAV) to  store solar energy for nighttime flight. The low-temperature
performance of Li-S cells may have in part led to the choice of this chemistry for this
application.260 Oxis Energy is expected to release a commercial Li-S battery cell in 2016, with
an eye toward xEV applications.261'262

   Silicon is also beginning to appear in the anode of commercial Li-ion  cells. While it takes 6
carbon atoms in a carbon anode to accept 1 lithium ion, a silicon atom can accept several.
However, uptake of lithium ions by silicon is accompanied by extreme volumetric expansion,
leading  to complications such as disintegration of the anode matrix and loss of electrical
conductivity. For this reason many  are currently  focusing on very small additions of silicon to an
otherwise carbon-based anode to achieve incremental improvements in specific energy.  In 2015
Tesla Motors Inc. announced a 90-kWh Model S pack that was said to achieve a greater specific
energy by including a small amount of silicon in the anode.263

   Solid-state lithium-ion cell technology is another active area of research.  Most solid-state
construction concepts retain the traditional anode and cathode couples but replace the liquid
electrolyte with a solid (usually polymer) electrolyte. Others seek to enable use of lithium metal
as the anode by leveraging the solid nature of the electrolyte to prevent dendrite formation.  Solid
state construction leads to the possibility of more efficient production techniques, such as
building complete battery cells by printing or deposition, potentially in complex shapes that
conform to available packaging space, or in flat shapes that could be integrated structurally with
the vehicle.  Minimizing the resistance of the solid electrolyte is a primary research target for
enabling this technology. As an indicator of interest in this technology, the British appliance
manufacturer Dyson purchased the  solid-state lithium-ion battery firm SaktiS for $90 million in
October 2015.264 In March 2016, it was widely reported that Dyson may be planning to produce
an electric vehicle, as  suggested by evidence that the company is receiving U.K. government
funding for this purpose.265  Similarly, Bosch, a major automotive supplier, acquired solid-state
                                             5-105

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
lithium-ion developer Seeo in 2015, citing potential applicability of the technology for increasing
the range of electric vehicles.266

   While promising, these and similar early examples of Li-S electrode couples, silicon anodes,
and solid-state construction will need time to show that engineering targets for cycle life,
dimensional stability, and durability in demanding xEV applications have been reliably met.
Until then, reliable estimates of their cost or commercial availability will not be available.
Metal-air chemistries will require even more development before they will be mature enough to
characterize their potential use in automotive applications or their production costs.  The 2015
NAS report (Finding 4.5, p. 4-44) further supports the conclusion that "beyond Li-ion"
chemistries such as these are unlikely to be commercially available during the time frame of the
rule. At this time the agencies consider it unlikely that fully proven forms of such chemistries
will become commercially employed  in xEV applications on a broad scale during the time frame
of the rule. The developmental state of these chemistries and the unavailability of well-
developed cost models prevent their inclusion in the agencies' analysis.

5.2.4.4.2      Pack Topology, Cell Capacity and Cells per Module

   Pack topology refers in general to the  way cells and modules are electrically connected to
form a pack.  Modules are collections of cells that act as building blocks for a pack.  Cell
capacity is the charge capacity of an individual cell, and is closely related to pack topology.

   To fully understand developments  in these areas  and the agencies' choices for these
parameters in the modeling of battery packs for costing purposes, an example of how these
parameters interact will now be presented as background.

   One approach to configuring a battery pack would start with a target pack voltage for the
application.  Target voltage typically refers to the nominal voltage expected at about 50 percent
SOC. For PEVs, the targeted voltage is typically between 300 V and 400 V. The most
commonly used Li-ion chemistries provide a nominal voltage between 3 V and 4 V  per cell.
Assuming a 3.8 V cell and a target of 365 V,  a BEV pack might be constructed of 96 cells
connected as  series elements (3.8 V * 96 =  365 V).  The target energy capacity  of the pack
(kWh) would then be achieved by specifying the capacity of each cell. The larger the  target pack
capacity, the  larger the required capacity of the cell.  In this example, to target a 24 kWh pack
capacity, each series element would need to have a capacity of about 66 A-hr:

   24,000 W-hr / 3.8 V / 96 cells = 66 A-hr

   Manufacturers have several options for providing this cell capacity. The simplest would be to
manufacture cells of 66 A-hr capacity. This results  in one cell at each series position,
minimizing the number of cells and interconnections, potentially minimizing the cost of the
pack. In practice, manufacturers may instead be compelled to use smaller cells, perhaps to better
address thermal management considerations, or to match an existing cell  size offered by a cell
supplier. The 66 A-hr required at each series position might then be provided by two 33 A-hr
cells, or three 22 A-hr cells, connected in parallel. The exact cell capacity could vary  slightly to
match available products if some variation in pack capacity or voltage are permissible.
Increasing the pack capacity, for instance doubling it to 48 kWh, could in theory be  achieved
either by doubling the number of series elements (from 96 to 192) or by doubling the A-hr
capacity of each series element (to  132 A-hr). The first option is problematic because it would
                                             5-106

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
double the voltage to 730 V, which presents a safety issue and may be outside the typical
operating voltage range of available power electronics. The larger cell capacity of the second
option may be difficult to achieve in a single cell while maintaining effective thermal and current
distribution characteristics within the cell. For these reasons, larger packs are often found to
include parallel strings of two or more smaller cells at each series position. Tesla products are an
extreme example, composed of thousands of very small cells, which results in as many as 36
cells in each series position.

   Another important aspect of pack topology is the format of the individual cell. As at the time
of the FRM, most industry cell development and current automotive cell applications continue to
be centered on prismatic (rectangular) cell formats composed of stacked or flat-wound electrode
strips housed in metal cans or polymer pouches. ANL BatPaC models a prismatic format housed
in a stiff polymer pouch.  Tesla remains almost  unique in its use of small, cylindrical 18650-
format cells.  Because Tesla has built significant market share, this  difference has potential
significance to the projection of future pack costs.  Also, there is some evidence that other
manufacturers are beginning to consider cylindrical cells. In 2015 Volkswagen announced the
R8 e-tron which has a pack composed of cylindrical cells; potentially, other products such as the
Q6 e-tron and the Porsche Mission E might also share this format if this is an indication of VW's
future battery construction approach.  Additionally, in November 2015 Samsung SDI announced
that it would supply cylindrical cells to a China customer for use in electric SUV battery
packs.267 According to one analysis, about 38 percent of currently available BEV models have
packs composed of cylindrical cells, with the rest roughly evenly divided between prismatic
pouch and prismatic metal can268 (although it is unclear whether the relatively large number of
Tesla sub-models are counted as separate models). About 40 percent of HEV models use packs
composed of cylindrical cells, according to the same source.

   Despite the differences between  prismatic and cylindrical cell formats, there may be limited
potential for large differences in pack costs to result.  First, material costs per unit energy storage
are likely to be similar on a cell basis. Cylindrical cells and prismatic cells differ primarily in the
manner in which layers of active materials are packaged together, one being a spiral winding of a
single electrode strip and the other a stack of multiple smaller strips.  Although the assembly
process is different, both methods utilize active  material with similar efficiency.  This is
significant because material costs are the most dominant component of total cell cost.137'269'270'142
Second, while cylindrical cells may benefit from a somewhat simpler cell manufacturing process
and the highly commoditized status of the 18650 format, the large number of 18650-format cells
that must be connected to build a pack may work against these advantages. While larger
cylindrical cells might be used, their heat dissipation properties may limit their practical size.
While 18650-format cells have good thermal qualities, larger cells begin to face challenges in
rejecting heat from the core of the cylinder where the maximum temperature tends to develop.271
Despite Tesla's success with the cylindrical format, it remains unclear whether either format
possesses a greater potential to eventually minimize pack cost. Therefore the agencies believe
that the cost estimates of the BatPaC model should be reasonably accurate for both cell
formats.272

   In the 2012 FRM analysis, xEV packs were preferably configured with a single series string
of cells.  The largest BEV packs were the exception, being configured with a  parallel string of
two cells in each series position, in  order to limit voltage to the desired range and limit the
required A-hr capacity of the cells.  Since the 2012 FRM, xEV battery packs (with the exception
                                             5-107

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
of Tesla, as previously mentioned) have largely continued to follow the practice of having one,
two or three cells in parallel at each series position.

  In the 2012 FRM analysis, maximum cell capacity was limited to 80 Ampere-hours (A-hr) or
less. While the cells of most packs configured for the analysis were well under this limit, some
larger BEV packs approached the limit. Therefore the cell capacity limit is primarily relevant to
the configuration of large BEV packs.

  An 80 A-hr cell  capacity was generally larger than the cell capacities observed in large BEV
packs at the time of the 2012 FRM.  The agencies expected that as the industry matured,
manufacturers would achieve economies by gradually optimizing cell capacities to the
requirements of the application, including an increase in cell capacity for large packs in order to
minimize the number of cells while limiting the total voltage. Since that time, there is some
evidence that manufacturers have begun moving toward larger cell capacities as expected.

  In October 2014 GM announced that the Chevy Volt generation 2 battery pack would have
fewer cells (192 vs 288) that are each about 50 percent greater in capacity.  In the original pack,
each series element was composed of three cells in parallel, while the new configuration has only
two.273 The 30 kWh trim of the 2016 Nissan Leaf, announced in September 2015, is said to
achieve its increased capacity within the same size and footprint of the lower-trim 24 kWh pack
by utilizing a more energy dense chemistry variation. The number of cells remained unchanged
at 192, implying an increase in the A-hr capacity of each cell.274  Similarly, the 2017 BMW i3
achieves a 50  percent increase in capacity over the earlier model, within the same pack volume,
by using a 94 A-hr cell in place of a 60 A-hr cell.242

  The latter example further suggests that cell suppliers are pushing the envelope of cell
capacity for vehicular applications beyond the 80 A-hr limit used in the agencies' 2012 analysis.
The 60 A-hr cell format that Samsung SDI had been supplying to BMW for the pre-MY2017
BMW i3 pack was  already  one of the larger light-duty BEV cell formats in use when it was
replaced by the 94 A-hr format. At AABC 2015, Samsung SDI presented further plans for
manufacturing prismatic  cells of 90 to 120 A-hr by 2020.275  The presenter also mentioned a goal
of eventually producing 180 A-hr cells for BEV use, using a new chemistry with high NCM
content plus silicon. This suggests that at least some suppliers are already anticipating a market
in vehicular applications for these very large format cells.

  Module configuration is another topology issue.  In general, the more cells that are included in
each module, the fewer modules and the lower the cost of their connections.  Since the number
of modules must be a whole number, the number of cells per module can depend on the total
number of cells necessary to reach a voltage or capacity target, and so need not be the same for
every size of pack.

  In the FRM, battery modules for all xEVs were configured with 32 cells per module.  The
Chevy Volt provided one example of a manufacturer that was already using at least 32 cells per
module, in a liquid-cooled application similar to that assumed in the agencies' analysis of BEVs
and PHEVs. Although most BEVs at the time had fewer than 32 cells per module, this figure
was selected to represent expected improvements in cell reliability and packaging methods as
manufacturers gained experience over time.
                                             5-108

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Since the FRM, some further evidence has emerged to support the agencies' expectation that
the industry will tend toward increasing the number of cells per module.

   It is now understood that the original Chevy Volt battery was configured with 7 modules of
36 cells each and 2 modules of 18 cells each. A similar configuration is retained in the recently
announced 2016 Volt. Similarly the Kia Soul EV battery consists of 192 cells in 8 modules,276'277
varying from 20 to 28 cells per module.

   As another example, in September 2015, Nissan announced a new battery pack option to be
available for the 2016 Leaf.  The two higher-trim versions of the Leaf, the SV and SL, will
include a 30 kWh pack in which the number of cells per module is increased from 4 to 8.278
While the number of cells per module is still relatively small, Nissan's continued use of passive
air cooling as a thermal management strategy may place a smaller limit on the number of cells
per module than for the more common liquid-cooled packs that are modeled in the agencies'
analysis.

   Subsequently, in November 2015 at the Tokyo Auto Show, Nissan revealed its IDS concept
vehicle, powered by a newly developed 60 kWh pack.279  In interviews with the press, a number
of details were shared regarding the design of this pack.  The pack was described as having 288
cells utilizing an NMC cathode chemistry. Assuming a nominal cell voltage of 3.75V typical of
these chemistries, each cell would be sized at about 55.5 Ampere-hours, significantly larger than
in the Leaf pack.  The IDS pack also appears to install in a footprint similar to that of the 30 kWh
version of the Leaf battery.  Nissan has not yet specified the number of cells per module in the 60
kWh pack, but evidence suggests it is significantly larger than in the Leaf packs.  One interesting
aspect of the design approach for this pack is its support for a variable module stack height,
suggesting a variable number of cells per module may be specified depending on the target
capacity of the pack.  In one press report,280 an official was described as saying that Nissan had
taken a conservative approach  to the number of cells per module in earlier packs, and due to the
lack of failures or other issues  with those packs, were now able to consider an approach that
supports a much larger number of cells per module in the new pack.

   In January 2016, GM announced details of the Chevy Bolt battery pack.281 As with the 60
kWh Nissan IDS pack, this 60  kWh pack is composed of 288 cells in 96 cell groups of 3 cells
each. The cells are distributed  among 10 modules, or about 28 to 30 cells per module.  Three
individual cells are connected in parallel at each series position. Assuming a nominal cell
voltage of 3.75V, this suggests an individual cell capacity of 55.5 Ampere-hours (identical to
that of the Nissan IDS pack).

   As noted above, the ideal number of cells per module may vary depending on the capacity of
the pack and the size of the cells.  In the 2012 FRM, the agencies assumed all modules would
have 32 cells. In this Draft TAR analysis,  it may be more appropriate to optimize the pack
topology by varying the number of cells per module in order to better match performance targets
and minimize cost.
                                             5-109

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
5.2.4.4.3      Usable Energy Capacity

   As previously noted in the introduction to this section, batteries may be described with respect
to their gross energy capacity or their usable energy capacity. Usable capacity refers to the
portion of gross capacity that the manufacturer believes can be regularly used in an application
while maintaining a desired level of durability. It is thus an important parameter for battery
sizing because it determines the gross capacity necessary to provide a target usable capacity for
an application.

   The concept of usable capacity is often accompanied by several closely related terms. In this
discussion, the following terms are used and defined as follows. State-of-charge, or SOC, refers
to the percentage of total energy (kWh) or charge (Ampere-hour) capacity that remains in a
battery at a given time, ranging from 0 to 100 percent on a gross capacity basis. SOC design
window,282 or simply SOC window, refers to the usable portion of total capacity intended by
design, expressed in terms of SOC; for example, an SOC design window might be described as
the range between 25 percent and 75 percent SOC, or alternatively as an SOC window of 50
percent. SOC swing may be used interchangeably with SOC window but is used here to refer
more specifically to observed in-use behavior rather than a design context.  Usable capacity is
thus determined by SOC design window (in a design context) or implied by an observed SOC
swing (in-use). Usable capacity may refer either to a usable energy (in kWh) or the usable
portion of gross capacity (in percent).

   For lithium-ion chemistries,  SOC is not always measurable with precision and is commonly
estimated by means of algorithms that include measurements of current, voltage and battery pack
temperature, both instantaneous and over time. The construct of SOC window therefore inherits
some of these traits. While it is most convenient to think of the boundaries of an SOC window in
terms of SOC percentages, it may also be defined by an allowable range of battery voltages, or a
combination of the two.

   The SOC design window that a manufacturer assigns to a battery is typically selected to
balance battery durability with energy availability. Owing to the complexity  of battery behavior
and vehicle control algorithms,  it is possible that some controllers may not refer to a single
rigidly defined SOC window, but instead, may define multiple or variable SOC windows that
apply to different usage conditions or are determined by the controller's observation of patterns
of usage or battery health monitoring over a short or long term. For example (and particularly
for HEVs), because extreme but intermittent usage conditions may have a different degree of
impact on battery life than normal usage, it is possible that some manufacturers may program
their controllers to define multiple target windows, to allow a wider swing to  accommodate
temporary, extreme conditions while following a narrower swing for normal conditions.  As
another example, some manufacturers may widen the allowable SOC swing as the battery ages
(perhaps by allowing a wider range of allowable voltages,  or modifying the allowable SOC
window) in order to maintain driving range or usable capacity. Although the concept of a single
SOC design window may therefore be overly simplistic for some vehicles,  it remains useful for
battery sizing purposes.

   Setting an appropriate  SOC window can be influenced by the effectiveness of the battery
management system (BMS).  Improved BMS systems are one potential path toward enabling a
wider SOC window or a reduced battery capacity for a given range.283
                                             5-110

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   The SOC design window is a primary factor in the sizing of a battery for a particular use.
That is, the desired electric driving range for a PEV, or the amount of energy buffering capability
desired for an HEV,  combined with the SOC window, directly suggests the necessary gross
capacity of the battery.  In the 2012 FRM, for battery sizing purposes, the agencies assumed a 40
percent usable SOC window would apply to HEVs, 70 percent for PHEVs, and 80 percent for
BEVs.

   Increases in PHEV and BEV driving range that have been observed since 2012 may have
been enabled in part by increases in SOC design window and hence usable capacity. The 2015
NAS report also stated (p. 4-5), "as extended in-use experience is obtained, the battery SOC
swing may be increased for all electrified powertrains."  For these reasons, it is appropriate to
review the usable capacity assumptions used in the 2012 FRM and to make any applicable
revisions for this Draft TAR analysis.

   Usable capacity for HEVs

   For the 2012 FRM, a 40 percent usable capacity was chosen by the agencies for strong HEVs
predicted to be available in the 2020-2025 time frame. At that time, 40 percent was greater than
the 20-30 percent observed in production HEVs of this type.  The agencies chose 40 percent on
the expectation that improvements in battery technology and manufacturer learning would enable
a wider SOC design  window by 2020.

   The 2015 NAS report (p. 4-5) was skeptical of the agencies' choice of a 40 percent usable
capacity for HEVs and suggested using a value closer to the 20 to 30 percent observed in
production HEVs. The report supported this position in part by contending that, by virtually
doubling the SOC window, the HEV batteries projected in the analysis would be "half the cost
and size" of what would be required.  However, the agencies believe that the wider SOC window
would not have this effect. At the high power-to-energy (P/E) ratio of an HEV battery, cost is not
as strong a function of capacity (kWh) as a function of power (kW). Therefore reducing battery
capacity from e.g. 0.50 kWh to 0.25 kWh, while holding the required power constant, would not
correspondingly reduce the cost by half, because the reduction in capacity would push the P/E
ratio to a higher level, counteracting much of the cost reduction.  Cost projections generated by
BatPaC confirm this trend and show that, for a given power capability, the cost of a 0.25 kWh
pack would be very similar to that of a 0.50 kWh pack. For example, BatPaC Version 3 projects
that an HEV pack sized for a power output of 15 kW would cost $634 as a 0.25 kWh pack, and
$660 as a 0.50 kWh pack, a difference of only about 4 percent.2  Therefore at these relative pack
capacities, EPA's use of a 40 percent SOC design  window for sizing purposes does not have a
large impact on projected cost.

   The agencies also believe that developments in battery technology and manufacturer learning
observed since 2012 have been consistent with the agencies' expectation that a 40 percent usable
capacity will be applicable to HEVs in the 2020-2025 time frame.  Since the 2012 FRM,
numerous HEV models and battery systems intended for such vehicles have been announced. It
is clear that although some HEV manufacturers have continued to use a rather conservative SOC
window (for example, at AABC 2015, it was reported that the 2016 Malibu Hybrid uses a  1.5
z BatPaC inputs: LMO-G chemistry, 1 module of 28 cells, EG-W (liquid) cooling, HEV-HP vehicle type, 450K
  annual production volume.
                                            5-111

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
kWh pack of which 30 percent is usable (450 Wh of 1500 Wh)258), there is also evidence that
some manufacturers have begun increasing the SOC design window in subsequent generations of
HEVs.

   Specifically, recent developments in batteries for 48V mild hybrids, which have smaller
batteries than strong HEVs but similarly demanding requirements, have supported a relatively
wide swing.  At AABC 2015, Bosch presented a 0.25 to 0.50 kWh battery system designed for
use in a 48V hybrid. This battery was described as having been designed for an SOC window
from 30 percent to 80 percent SOC (a 50 percent usable capacity) despite its small total
capacity.284 Also at AABC 2015, A123 Systems presented a battery system for a 48V hybrid
that uses a proprietary chemistry  variation on Lithium-iron phosphate which the company calls
Ultraphosphate. Like the Bosch system, this 0.37 kWh pack supports a window from 30 percent
to 80 percent SOC (50 percent usable capacity).  A123 indicated that production of this pack is
planned to begin in 2017.256

    In 2014, EPA tested a 2013 Volkswagen Jetta Hybrid supplied by Transport Canada as part
of an exploratory benchmarking exercise.  Several braking and acceleration episodes were
performed with the intention of eliciting maximum swing of the 1.1 kWh battery. Multiple
energy swings were observed in both charge and discharge ranging from 0.56 to 0.65 kWh,
equivalent to a gross SOC swing of about 51 to 59 percent.285 Although this testing documented
that the vehicle controller will permit this SOC swing to occur under these usage conditions, it
remains unclear whether this degree of swing would be observed regularly over normal usage. A
limited amount of testing over steady-state and standard test cycles elicited smaller swings of up
to approximately 30 percent.  The short duration of standard test cycles and variation in the
observed swing prevented firm conclusions from being drawn about the exact SOC design
window the controller regularly permits.

   Going forward, it is possible that improvements in cell balancing  may also act to support
downsizing of HEV battery sizes or widening of SOC windows from their current levels.  For
example, at AABC 2015, NREL presented work showing that use of active cell balancing instead
of passive  balancing can result in a 50 percent reduction in the necessary capacity of an HEV
battery while also eliminating the need for liquid cooling.286

   These findings suggest that EPA's choice of 40 percent usable capacity for HEVs remains a
conservative estimate for the 2020-2025 time frame.

   In the NHTSA analysis conducted by Argonne National Laboratory using the Autonomie
model, a 15 percent to 20 percent usable SOC window was assumed for HEVs during standard
test procedures at ambient temperature. Higher usable SOC swings were measured at Argonne's
APRF under different test conditions (i.e. A/C on).

   Usable  capacity for PHEVs

   For the  2012 FRM, a 70 percent usable capacity was chosen by the agencies to represent both
PHEV20 and PHEV40 vehicles.  The usable portion of total capacity for a PHEV tends to be
narrower than for a BEV.  One reason for this difference is that when a BEV reaches its
minimum SOC, it is taken out of operation and recharged, while a PHEV instead begins to
operate in  charge-sustaining mode (charging and discharging within a narrow SOC band) for an
indefinite time. The need to provide a proper lower-end buffer for the SOC band, and to avoid
                                            5-112

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
extensive operation at a very low SOC, encourages setting a higher minimum SOC point for a
PHEV than for a BEV. PHEV batteries also tend to have a larger P/E ratio due to their need to
provide similar power levels as a BEV battery while having a smaller capacity. A smaller SOC
window would be appropriate under such conditions to promote battery life. The 2015 NAS
report (p. 4-12) affirmed that  a 70 percent usable capacity is appropriate for a PHEV
architecture.

   At the time of the 2012 FRM, relatively few PHEVs were in production to serve as examples
of this platform. The primary production example was the Chevy Volt, which was about to be
released in its first generation (referred to here as Genl). Prior to its release, the usable capacity
of the pre-production Genl Volt battery was commonly reported as approximately 8 kWh of a
total 16 kWh, or about 50 percent.  The first production Genl Volt is now understood to have
utilized about 10.2 of 16 kWh, or about 64  percent.287 Testing of a 2012 Chevy Volt by Argonne
National Laboratory showed the vehicle to  be utilizing an SOC window between 87 percent SOC
and 18 percent SOC (69 percent usable capacity).288

   The initial generations of the Chevy Volt are often described as having adopted a conservative
battery management approach by utilizing a narrow SOC design window  and liquid cooling.
Since the 2012 FRM, GM widened the SOC window for the Volt on at least two occasions while
increasing the battery capacity on at least three. The  Genl model was upgraded in the 2013MY
from 16 kWh gross capacity to 16.5 kWh, and further increased for the 2015MY to 17.1 kWh.
During this process the usable energy increased from 10.2 kWh in the 16  kWh version to 11.2
kWh in the  17.1 kWh version. This represented a small increase in usable energy capacity, from
63.75 percent of gross capacity to 65.5 percent. The Gen2 Volt, released for the 2016MY, now
uses 14 kWh of 18.4 kWh gross, or about 76.1 percent usable capacity. This represents a 25
percent increase in usable capacity from the last Genl model.287

   The PFIEV batteries modeled in the 2012 FRM are similar to the Volt battery in that they are
liquid cooled, enabling the same level of temperature control that is often cited as being
responsible for the dependability of the Volt battery.  The production 2016 Volt battery now
exceeds the 70 percent usable capacity assumed by the agencies, potentially suggesting that the
70 percent figure is conservative.

   It should be noted that the  2016 Volt battery is sized for a 53 mile AER, and accordingly may
have a significantly lower P/E ratio than that for a PFLEV20.  This may allow it to enjoy  a wider
SOC design window than the smaller battery of a PFLEV20 or possibly even that of a PFLEV40.
Therefore the Volt example is not by itself conclusive that a wider SOC window would be
appropriate for PHEV20 or PHEV40.

   According to results  of testing at Argonne National Laboratory, the Ford Fusion Energi
utilizes about 5.9 kWh of its 7.6 kWh gross capacity, or about 78 percent.  This provides an
additional data point suggesting that a wider SOC window than 70 percent may be appropriate
even for some shorter-range PFLEVs. The Fusion Energi is rated at 20 miles of AER,  and utilizes
a blended depletion style that may utilize the engine if driven more  aggressively than in the
standard EPA test cycles.  This engine supplementation at elevated power demands is likely to
result in lower peak power demands on the battery, potentially making wider swings less
demanding  on the battery.
                                            5-113

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   These findings suggest that EPA's assumption of 70 percent usable capacity for PHEVs
remains a conservative estimate for the 2020-2025 time frame.

   Usable capacity for BEVs

   For the 2012 FRM, an 80 percent usable capacity was assigned to BEV batteries. This was
based on knowledge of manufacturer plans as well as examples seen in the press for early
production BEVs such as the Nissan Leaf and other developmental vehicles. The 2015 NAS
report (p. 4-12) affirmed that an 80 percent usable capacity is appropriate for BEVs.

   Since the 2012 FRM, a large number of BEV models have reached production, and thousands
of BEVs have accumulated a great degree of road usage. This has provided abundant
opportunity for manufacturers to begin drawing conclusions regarding the appropriateness of the
SOC design windows they chose to implement in their first generation models, and even to begin
applying the findings to subsequent model generations. It has also provided many opportunities
for research organizations to test these vehicles to ascertain aspects of their design and behavior,
including SOC swings observed in use. Table 5.5 summarizes some estimated SOC swings
observed in 2012-2016MY BEVs, which are further described below.
                 Table 5.5 Estimated SOC swings for selected MY2012-2016 BEVs
Example
ANL EV benchmarking (various)
Tesla Models 85
2015 Kia Soul EV
BMWiS
Estimated SOC swing
80 to 90 percent
85 percent
90 percent
87 percent
Source
Argonne National Laboratory
AVL
Idaho National Laboratory
Idaho National Laboratory
   Argonne National Laboratory (ANL) operates an ongoing research program to benchmark
xEVs.288 Vehicle testing from multiple instrumented battery electric vehicles has shown that the
vehicles operate usable SOC windows ranging from 80 percent to 90 percent whether air cooled
or water cooledAA.  The agencies will continue to analyze data from these tests to establish the
SOC swings being seen in current and newly released xEVs.

   At AABC 2015, AVL presented the results of a teardown of a Tesla Model S battery pack.248
AVL reported that cycling tests of the pack suggested that 73 kWh of the 85 kWh gross capacity
is accessible, suggesting that this pack may be utilizing an 85 percent usable capacity.  This
result is in line with reports from Model S owners that have suggested a usable capacity of about
75 to 76 kWh.289

   The Advanced Vehicle Testing Activity group at Idaho National Laboratory has tested the
batteries of several BEVs currently in production.290  In testing of the 2015 Kia Soul EV, the
measured battery capacity ranged from 30.4 to 30.5 kWh in each of four test vehicles.  The
service manual for the 2015 Kia Soul EV is reported to list a nominal SOC range of 5 percent to
95 percent, or 90 percent usable, for the high voltage battery system.291  A 90  percent SOC
window would amount to about 27 kWh of usable energy, the same as Kia advertises.  In a
AA Instrumented battery electric vehicles include: 2015 Chevrolet Spark EV, Kia Soul EV, 2014 Smart EV, 2013
  Nissan Leaf, 2012 Ford Focus Electric.
                                            5-114

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
departure from the practice of most other OEMs, Kia may be advertising the usable capacity
rather than the gross capacity.

   Technical specifications for the BMW i3 indicate a battery capacity of 18.8 kWh.292
Numerous press sources widely repeat this figure as a usable SOC while consistently citing a
gross SOC of 21.6 kWh or 22 kWh. The 21.6 kWh figure is highly consistent with the results of
battery testing by Idaho National Laboratory293'294'295'296 for four 2014 BMW 13 vehicles under
test, which indicated gross capacity ranging from 21.4 kWh (one vehicle) to 21.7 kWh (three
vehicles).  Like Kia, BMW appears to be advertising the usable capacity of the i3 battery rather
than the gross capacity. A gross capacity of 21.6 kWh suggests a usable capacity of 87 percent.

   In May 2014, the Chevy Spark EV underwent changes to its battery that may indicate a
widening of SOC design window.  In announcing a change  in cell supplier from A123 Systems
to LG Chem, General Motors also indicated that the new Spark battery would be reduced in
capacity from 21 kWh to 19 kWh, while keeping the same range of 82 miles and the same
mpge.297 Given that rated mpge did not change, this suggests that retention of the original  range
was more likely made possible by widening the SOC design window than by increasing
powertrain efficiency. A widened window could be enabled by either the use of a different
battery chemistry (going from A123's Lithium-Iron Phosphate to LG Chem's NMC+LMO
chemistry), and/or an increased comfort level due to ongoing experience with the platform.
Since the original  A123 cathode chemistry (Lithium-Iron-Phosphate or LFP) is comparable to
LG Chem's LMO-dominant chemistry in terms of allowable SOC  swing, it suggests that
experience may have played at least some role in this change.

   At AABC 2015, Honda reported that their decision to  extend the lease option on the Fit EV
by 2 years was based on learning that the batteries in these vehicles were experiencing lower
degradation than projected.298 This suggests that it might be possible to widen the SOC design
window in future releases while maintaining durability targets.

   The agencies' 2012 choice of 80 percent usable capacity  for BEVs appears consistent or
slightly conservative in light of the trends discussed here.

5.2.4.4.4     Thermal Management

   Battery thermal management includes battery cooling to  reject heat generated during use, and
in many cases battery heating to warm the battery in cold weather. In systems where active
thermal management is present, the battery management system (BMS) will work to keep the
battery within a preferred temperature range during use.

   Battery thermal management systems are commonly divided into passive systems  (where the
outside of the pack is exposed to ambient air) and active systems (where a cooling medium is
circulated through the pack, or thermoelectric components are integrated with the pack).  Active
cooling media may be ambient air, cabin air, air  conditioned by the vehicle A/C system, a liquid
coolant, or the A/C system refrigerant.299'300'301'302

   For the FRM, the agencies assumed PEV packs would employ active liquid cooling, which
was seen in production vehicles such as the Chevy Volt and in several limited-production PEVs
at the time of the FRM. In contrast, HEV packs  were assumed to employ passive air cooling
acting on the outside of the pack, which was the  prevalent method seen in HEVs at the time.
                                             5-115

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   One recent approach to cooling battery packs involves placement of a bottom cooling plate
beneath the packaged battery cells rather than between each cell. Coolant or refrigerant
circulates through the plate and cools the battery cells conductively. This approach is used in the
BMW i3 battery, was once used in the Chevy Spark A123-supplied battery, and is possibly being
used in the Chevy Bolt pack.303

   Direct circulation of refrigerant rather than an intermediary fluid such as a gly col-water mix
can also improve heat rejection and vehicle packaging by eliminating the secondary cooling loop
that would otherwise be needed to reject heat to the atmosphere.  The BMW i3 utilizes
refrigerant cooling.299

   Active liquid cooling continues to be the predominant thermal management method for the
battery packs of BEVs and PHEVs announced since the FRM. The notable exception is the
Nissan Leaf, which continues to use passive air cooling as it has since its first generation. At the
time of the FRM, some in the industry and press were expressing skepticism about Nissan's
choice of passive air cooling.304'305'306 Some customers had also begun reporting unexpected
battery degradation in hot climates such as Arizona, which some attributed to inadequate thermal
management. During the 2014 MY, Nissan adjusted the chemistry of the battery pack to better
withstand high temperatures.307 Although Nissan has continued to use passive air cooling in the
2016 Leaf (and also in the new 60 kWh pack under development), all other production BEV and
PHEV packs introduced since the FRM use some form of liquid or refrigerant-based cooling.
The 2015 NAS report (under "Cooling," p. 4-17) tended to affirm the agencies' assumption of
liquid cooling for BEV packs by independently noting the potential inadequacy of passive air
cooling in the Leaf pack.

   Although HEV packs were the only packs modeled with passive air cooling in the 2012 FRM
analysis, there is some evidence that even these packs may be moving toward liquid cooling.
Although air cooling continues to predominate,302 a presentation by Mahle at TMSS 2015
suggests that air cooling is increasingly being displaced by liquid cooling even in HEV packs.300
Johnson Controls has also described a 260 V, 1.7 kWh FIEV battery product with provision for
liquid cooling.308 Effective cooling and heating capability is often cited as a potential path
toward reducing the size of xEV batteries by allowing more of their capacity to be utilized while
minimizing degradation.302'294 This suggests that liquid cooling may become one of the enablers
for future FIEV batteries to provide the 40 percent usable capacity assumed in  the agencies'
analysis.

   As previously described, EPA uses ANL BatPaC to model the cost of xEV batteries,
including mild and strong FIEV batteries. BatPaC provides cost estimates for several cooling
options, including active air cooling (cabin air or cooled air) and liquid cooling (glycol/water
mix). It does not model passive air cooling without air channels between the cells, as might be
found in passively cooled HEV batteries. EPA performed several trials to investigate the impact
of the available cooling choices for HEV batteries, and found that BatPaC assigns similar or
slightly lower costs for its implementation of liquid cooling than for its implementation of active
air cooling. For these reasons EPA now uses the liquid cooling option under BatPaC to model
the cost of HEV packs, as already done for PHEV and BEV packs.

5.2.4.4.5      Pack Voltage
                                             5-116

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   In the 2012 FRM analysis, EPA limited pack voltages to certain ranges depending on whether
the pack was intended for an HEV, PHEV, or BEV. HEVs were targeted to about 120V while
PHEVs and BEVs ranged from 360V to 600V.  For this Draft TAR analysis (as described in
detail in Section 5.3), EPA lowered the voltage range for PHEVs and BEVs to between
approximately 300 and 400V to reflect trends observed since the FRM. NHTSA designed pack
voltages to meet the voltages currently in the market and to reflect the trend of lowering the pack
voltage by using high capacity batteries to reduce cost.

   To some degree, the customary voltage range for a given xEV category is an outgrowth of the
relative size of the battery. Small battery packs for HEVs can be composed of a correspondingly
small number of cells, which limits the attainable voltage even if all cells are placed in series.
These lower voltages are also consistent with the desire to maintain safety as well as with any
need to interface with the 12V electrical system that typically remains in these vehicles. Larger
packs for PHEVs and BEVs are typically composed of a much larger number of cells and so can
easily reach a much higher voltage if desired. While safety considerations continue to place a
practical upper limit on system voltage, a moderately high voltage is consistent with the greater
power flows required by these vehicles and offers the added benefit of conducting energy at a
lower amperage, which reduces the necessary weight and cost of electrical conductors and
reduces I2R losses.  Compatibility of available supplier parts may also encourage different
manufacturers to target a similar voltage envelope.  Many manufacturers of PHEVs and BEVs
appear to have targeted the range between 300V and 400V.

   The system voltages chosen by the agencies for modeling xEVs were based on those seen in
production xEVs at the time of the FRM.  Since the FRM, the agencies have not observed a
strong trend away from these general voltage ranges in newly released xEV products, with the
possible exception of the upper voltage limit for PHEVs and BEVs.

   EPA's original 600V upper limit on BEV battery voltage had been set to accommodate the
largest BEV packs that were modeled in the 2012 FRM analysis.  Most PHEV and BEV packs
modeled in the 2012 FRM were in the 300V-400V range. The only pack modeled in the 2012
FRM that approached the 600V limit was a Large Truck EV150 pack at 586V. At the time, VIA
Motors was producing a plug-in electric truck with a 650V battery pack that served as a
corroborating example. However, later versions of this and other VIA products have since
adopted a lower battery voltage of around 350V to 380V, suggesting that some advantage was
seen to adopting a lower voltage.

   Examples of PHEVs and BEVs in the 600V range continue to exist. The McLaren PI PHEV,
first introduced to the U.S. in 2014 as a very  limited production high-performance vehicle,
operates at 535V. In September 2015, Porsche announced the Mission E concept BEV that
would operate at 800V.  The higher voltage was described  as enabling much faster charging as
well as lower conductor weight.309 These examples suggest that voltage ranges higher than the
typical 300V-400V may continue to be applicable at least to high performance BEVs and
PHEVs.

5.2.4.4.6     Electrode Dimensions

   The electrodes of a lithium-ion cell are in  the form of flat foil strips coated with active
materials and stacked or rolled together. Several important parameters of cell performance are
                                           5-117

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
controlled by the dimensions of the electrode; in particular, the thickness of the active material
coatings on the electrodes and the aspect ratio (length-to-width ratio) of the electrodes.

   In general, thinner electrode coatings promote power density, while thicker coatings promote
energy density. By default, BatPaC limits coating thickness to no less than 15 microns  and no
more than 100 microns due to various practical considerations.137 The lower limit represents
interfacial impedance effects associated with very thin electrode coatings.310 The typical
precision of coating equipment, at around plus or minus 2 microns,311 would also become
challenged below this thickness.  The upper limit represents material handling and ion transport
considerations. Thicker coatings may be prone to flaking when uncut electrode sheets are rolled
or unrolled for shipment and processing.  Thicker electrodes also require ions to travel a greater
distance through the active material during charge and discharge, leading to effects such as
increased resistance, reduced power capability, and the potential for lithium plating on charging.
In the 2012 FRM, electrode coating thickness was therefore limited to  100 microns.  In practice,
this limit was only encountered by the most energy intensive packs for large BEVs.  In the latest
release of BatPaC, ANL has improved the model by which electrode thickness is determined.  In
most cases this results in somewhat thinner electrodes than would have been projected in the
version used for the 2012 FRM analysis.  This is expected to result in a slightly higher cost per
kWh for most battery packs, all other things being equal.312

   Electrode aspect ratio is important because it determines how far current must travel on
average between where ions reside in the active materials and the current collector tabs. Longer
distances are associated with greater resistance and heat generation. If the length is much greater
than the width, and the current collector tabs reside on the short dimension rather than the long
dimension, current must travel farther on average than in the inverse situation. BatPaC assumes
a default aspect ratio of 3:1, with tabs placed on the short dimension. In the FRM, EPA used an
aspect ratio of 1.5:1, loosely based on the dimensions of some commonly known cells at the
time.

   The 3:1 default aspect ratio used in BatPaC appears to be seeing increasing use in the
industry.  In announcing the 200-mile Chevy Bolt EV281 at the 2016 NAIAS, GM indicated that
its battery cells, supplied by LG Chem, have an aspect ratio of 3.35:1 (measuring 3.9 inches by
13.1 inches).  An animation accompanying the announcement shows that the cell tabs reside on
the short dimension. The Kia Soul EV battery also uses cells with a nearly identical aspect ratio
and tab placement, supplied by SK Innovation.313'253  These examples lend support to the validity
of the default 3:1 aspect ratio and tab placement assumed by BatPaC.  GM describes this aspect
ratio as "landscape format," presumably to highlight the low-profile design of the pack that
allows the entire pack to reside within the floor space of the vehicle.

   Also at the 2016 NAIAS, Samsung SDI introduced a family of cells ranging from 26 to 94
Ampere-hours,314 some of which have a similar aspect ratio to the GM Bolt cells but with tabs on
the long dimension. Samsung also displayed a line of "low height packs," suggesting that it
anticipates a trend toward low-profile applications for which these cells would be well suited.315
In December 2015, Volkswagen also announced plans to pursue flat, low-profile pack designs
for future electrified vehicles,316 which likely will also call for a similar cell aspect ratio.

5.2.4.4.7      Pack Manufacturing Volumes
                                             5-118

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   In the 2012 FRM analysis, the agencies assumed that battery pack manufacturing would reach
full economy of scale at an annual production volume of 450,000 packs in the year 2025.  This
volume was based on the annual manufacturing volumes assumed by FEV in the teardown
analyses performed for the FRM analysis.

   In BatPaC, when the user specifies a production volume of 450,000 for a given battery pack,
it means that the cost estimate for that specific pack is based on a dedicated manufacturing plant
that manufactures an annual volume of 450,000 of that identical pack.  Since all of the packs
produced by the hypothetical plant are identical, it implies that the cost estimate is most
applicable to a situation in which the packs are intended to be used by a single manufacturer in a
single model of electrified vehicle.

   The 2015 NAS report noted (p. 4-42, and Finding 7.3, p. 7-23) that the technology penetration
levels projected by the agencies for electrified vehicles are lower than the 450,000 annual
production volume that the agencies assumed in projecting battery pack costs for the 2022-2025
time frame. Further, it noted that whatever annual production did occur would likely be divided
among multiple manufacturers and multiple models, preventing the full economy of scale of
450,000 units from being achieved by any  single manufacturer. The report recommended that
the agencies use a smaller manufacturing volume for electrified vehicle battery packs to better
reflect projected technology penetration, rather than the 450,000 annual production assumed in
the 2012 FRM.

   Despite the agencies' use of an annual production of 450,000 units, it is unclear whether this
results in more optimistic  estimates of battery cost than the industry may realize. The following
discussion describes several points relevant to this consideration: (a) the potential for a "flex
plant" manufacturing approach to realize economy of scale at much lower pack volumes; (b) the
potential for economies of scale to  fully develop at production volumes at low as 60,000; (c)
examples of actual costs that are already lower than the agencies' FRM estimates at a much lower
production volume than 450,000; (d) the agencies' placement of estimated costs in the year 2025
instead of 2020; and (e) the potential for consolidation in the battery industry to increase pack
manufacturing volumes.

   There is evidence that optimizing the approach to battery manufacturing by adopting a "flex
plant" approach may allow economies of scale to be realized at pack production volumes much
lower than 450,000.  According to a recent ANL study,317 a battery manufacturing plant that is
designed to simultaneously manufacture packs for multiple vehicle types (HEVs, PHEVs and
BEVs) by standardizing on a single electrode width can significantly reduce the pack
manufacturing volumes required to achieve maximum economy of scale. The ANL study calls
this approach a "flex plant." Some manufacturers already appear to be adopting a similar
approach for production of prismatic cells. For example, at AABC 2015, Samsung SDI described
a strategy to build an "ecosystem" of xEV battery products by maintaining a "standard cell
format between generations," that is, by maintaining the same cell dimensions and container size
and achieving different target capacities by varying the chemistry.275 At the same conference,
Bosch similarly described a goal to produce packs of varying capacity by use of a standard 36
Ampere-hour cell.284 XALT Energy also described its practice of achieving variable cell
capacity (Ampere-hour) sizes by adjusting the electrode count within a cell while maintaining
one of two fixed cell footprint areas.318 Cell  standardization also may promote the economics of
battery second life applications319 and so could provide an added motivation for manufacturers to
                                             5-119

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
reduce the number of cell formats. The agencies anticipate that the most successful suppliers
may continue to adopt similar approaches over time. As this occurs, the production volume of
the individual cells that compose the several pack types produced from those cells would
increase dramatically, even though pack volume of any single pack type may remain relatively
low. This increased cell volume may recapture much of the economy of scale reflected at the
pack level in the 450,000 unit assumption.

   There is also some evidence to suggest that economies of scale may be achieved at much
smaller pack production volumes than 450,000, even without necessarily adopting a flex plant
approach. According to the ANL flex plant study, the benefits of a flex plant over a dedicated
plant for reducing the cost of BEV batteries levels off past a production level of about 60,000
units per year, suggesting that 60,000 units would approach maximum economy of scale for a
dedicated plant. The 2015 NAS report (p. 4-42), in noting that agencies' projected costs for 2012
"seem reasonable" despite the large volume assumed, cites as a possible explanation a TIAX
study (referred to  as Sriramulu & Barnett 2013 in a National Research Council report on
Overcoming Barriers to EV Deployment222) that also suggests a 60,000 unit volume at which
economies of scale would be  realized. This level of production is much closer to the technology
penetration levels predicted by the agencies.  Individual manufacturers such as Nissan and Tesla
are already approaching similar production levels, with Nissan having sold more than 30,000
Leaf EVs in North America in 2014, and Tesla projecting a similar amount in 2015. The BMW
i3 and i8 PHEVs are also approaching a global production level of 30,000 units per year.

   There is also evidence that actual battery pack costs experienced by some manufacturers are
already lower than the agencies' FRM estimates, at a much lower production volume than
450,000. As discussed in more detail below, General Motors has cited its rapidly falling battery
cell costs from supplier LG Chem as evidence of their being "able to achieve lower costs earlier
with much less capital and volume dependency" than presumably had been expected. The cell-
level costs cited by GM for the Chevy Bolt are lower than the BEV pack costs projected by the
agencies in 2012.  Because it appears to suggest a currently contracted price applicable  at the
very beginning of the Bolt product cycle, it therefore is likely to be based on an annual
production level of far less than 450,000 packs.  Production of the 2017 Bolt has been
characterized as capable of serving a demand of around 50,000 units per year.320

   The way the agencies apply the BatPaC-generated costs also treats them conservatively.
Although the cost estimates generated by BatPaC are intended by its authors to represent
technology being used in the year 2020, the agencies assign these costs to the year 2025 when
applying reverse-learning to generate year-by-year cost estimates for earlier years.  Although this
was a practical choice in order to cover the full time frame of the standards which run to 2025, it
has the effect of making the projected costs more conservative by assuming that the technology
projected by the BatPaC authors will not take effect for an additional five years.

   Consolidation among battery cell suppliers may also improve the ability for individual
suppliers to begin approaching the production volumes assumed in the analysis. Since the FRM,
there has been significant consolidation among battery manufacturers.321'322'323 For example,
A123 Systems, which at one time competed against LG Chem to supply battery cells for the
Chevy  Volt and was later chosen to supply the Fisker Karma and Chevy Spark, filed for
bankruptcy in late 2012 and was sold to Chinese auto supplier Wanxiang in 2013.324 Wanxiang
has since refocused A123's efforts toward smaller HEV and stop-start batteries as well as grid
                                             5-120

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
storage. Johnson Controls, which was ranked in second place as an industry leader by one
analysis firm in 2013,323 also has refocused its effort on smaller batteries.  As of late 2015, three
xEV cell suppliers appear to have been particularly successful at developing OEM partnerships:
LG Chem, Panasonic, and Samsung SDL325 LG Chem has grown its customer list to include not
only GM but also Renault, Volvo, Daimler, Volkswagen, Audi, and Tesla.326 Panasonic is also a
dominant player through its ongoing partnership with Tesla, as well as supplying smaller
contracts with Ford and Volkswagen. Samsung SDI is a supplier to BMW and in 2015
announced plans to acquire the battery division of Magna International.327 Nissan's joint-venture
arm Automotive Energy Supply Corporation (AESC) is also an important player through its
battery production for Nissan and Renault vehicles, including the Nissan Leaf. In 2015 it was
reported that Nissan is also considering a partnership with LG Chem for its future BEV
batteries.328  Even Tesla, which has long-term plans to source cells from its so-called
Gigafactory, is said to be investigating the possibility of sourcing cells from  other leading
suppliers in order to meet expected  demand for the Model 3 in  a timely manner.329

   For the reasons discussed above, and in view of the evaluation of 2012 FRM battery cost
projections (described in Section 5.2.4.4.9 below), EPA believes that an assumed manufacturing
volume of 450,000 was appropriate as a BatPaC input for the purpose of generating battery pack
cost estimates for the 2012 FRM analysis.

5.2.4.4.8     Potential Impact of Lithium Demand on Battery Cost

   Controversy has periodically arisen about the adequacy of known lithium reserves to service
the potential demand generated by the electrified vehicle industry. However, lithium appears to
be plentiful enough at this time to suggest that its availability will not be a constraint in the near
term.330'331

   At circa-2010 prices, the cost of lithium content was said to be only about 1 percent of total
material cost at the battery pack level331 or perhaps 2 percent at the cell level.332 Lithium
comprises a  similar percentage by mass, and at time of manufacture resides primarily as ions in
the cathode active material and the electrolyte solution.

   Lithium used in cell manufacturing is most commonly sourced as lithium carbonate.333
Lithium carbonate is primarily recovered from ancient continental brines underlying salt lake
deposits. These are widespread in the southern Andes (primarily Bolivia, Argentina, Chile) and
western China and Tibet, with deposits identified in the southwest United States as well.
Lithium may also be recovered from some oilfield brines in the western U.S.  Because industrial
applications  for lithium were relatively few and scattered prior to its use in batteries, known
reserves may not be as well enumerated as for other commodities, and may have potential to
increase as demand increases and previously unidentified or unexploited sources are recognized.

   Recently, concerns about lithium prices have been renewed  by a significant increase in the
price of lithium, thought to be resulting in part from increased demand for use in electrified
vehicles.334  Pressure also appears to be increasing on manufacturers to secure lithium  sources
that will be needed to supply increased production capacity.335  A study released by Carnegie-
Mellon University in May 2016336 addressed this issue directly by examining the sensitivity of
battery cell manufacturing cost to the price of lithium carbonate and lithium hydroxide. The
study concluded that the effect on battery pricing would be minimal (never more than  10
percent) even for the most extreme lithium price fluctuations considered (about four times the
                                             5-121

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
historical average). The researchers also suggested that the primary difficulty imposed by such
fluctuations would be felt by cell manufacturers in maintaining profit margins, rather than by
vehicle manufacturers or consumers.

5.2.4.4.9     Evaluation of 2012 FRMBattery Cost Projections

   In the 2012 FRM, the agencies adopted a bottom-up, bill-of-materials approach to projecting
the future DMC of xEV batteries by using the ANL BatPaC battery cost model.137 As discussed
in the Technical Support Document (TSD)136 accompanying the 2012 FRM, battery pack costs
projected by this model were shown to compare favorably with cost projections provided by
suppliers and OEMs that were interviewed during development of the rule. In the 2015 NAS
report (Finding 4.4, p. 4-43), the committee found that "the battery cost estimates used by the
agencies are broadly accurate," providing further support for the use of this model.

   At the time of the FRM, few public sources were available to further validate these
projections.  Since that time, several sources have emerged that provide additional information
on the evolution of battery costs since the FRM and potential future trends.

   In 2015, a peer-reviewed journal article (Nykvist and Nilsson, 2015) appeared that provides a
comprehensive review of over 80 public sources of battery cost projections for BEVs.142 Based
on a statistical analysis of these estimates, it was shown that industry cost estimates for lithium-
ion batteries for BEVs have declined 14 percent annually between 2007 and 2014, and that pack
costs applicable to leading BEV manufacturers have  followed a cost reduction curve of about 8
percent per year, with a learning rate of between 6 percent and 9 percent. The authors concluded
that the battery costs experienced by market leading OEMs are significantly lower than
previously predicted, and that battery costs may be expected to continue declining.

   Figure 5.37 compares the full population of cost estimates reviewed by Nykvist and Nilsson
to the battery pack cost projections of the 2012 FRM analysis. Because BatPaC does not
produce cost estimates for multiple years, the 2012 FRM analysis applied a learning curve to
generate costs for  the years 2017 through 2025, with BatPaC output costs assigned to the year
2025. The learning-adjusted FRM costs shown in the figure include those for PHEV40, EV75,
EV100 and EV150, which have relatively large capacities similar to those likely included in the
review.  The plot shows that the battery costs projected in the 2012 FRM fit well with the
reviewed estimates, and lie on a similar cost reduction curve.
                                             5-122

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
1600
1400
3f 1200
.2, 1000
a
c
5, Hnn
jit per kWh
s § i
0
9f¥1

2C






DO 20
n D
° BD
DBD
n n
0n

B n

05 2010


L

=





n B|
n L

2C



, 	
i !::!
- 1 * * K
I'

15 20
Year



|
1 • • • •
i • i i
III)
J
20 20




1 |

25 20




]

30 20






35
                                                                           n Nykvist/Nilsson

                                                                           • 2012 FRM
Figure 5.37 Comparison of 2012 FRM Projected Battery Cost Per kWh to Estimates Reviewed by Nykvist &
                                         Nilsson

   Cost estimates and projections are most useful when they can be validated by comparison to
actual costs. Unfortunately, information about actual battery costs paid by manufacturers for
production vehicles is rarely disclosed publicly. However, in October 2015, General Motors
publicly commented on its battery costs for the Chevy Bolt EV, providing an opportunity to
evaluate the FRM projections of BEV battery costs.

   At the General Motors Global Business Conference on Oct. 1, General Motors described to an
investor audience its current and projected cost per kWh (on a cell basis) for battery cells for the
Chevy Bolt EV.  Citing partnership with cell manufacturer LG Chem, Executive Vice President
of Global Product Development Mark Reuss stated, "When we launch the Bolt, we will have a
cost per kWh of $145, and eventually we will get our cost down to about $100. We believe we
will have the lowest cell cost with much less capital and volume dependency."337 An
accompanying chart shows the $145 cost continuing to 2019, dropping to $120 per kWh in 2020
and to $100  per kWh in 2022.338'339

   It is important to note that the costs described above are cell-level costs and not pack-level
costs. To compare them to the pack-level costs projected by the agencies requires converting
them to that basis using an appropriate methodology. Also, although the context of the
announcement suggests that the costs are comparable to a direct manufacturing cost, their exact
basis is unknown. Although these factors introduce some uncertainty in comparing the
announced costs  to the FRM projections, a qualified comparison is possible.

   Several sources exist that suggest a cost conversion factor from cell-level costs to pack-level
costs for lithium-ion batteries.340'269'248'341'342'343  These are summarized in Table 5.6.  Most of
these sources suggest a conversion factor of about 1.25 to 1.4 may be appropriate.

   Table 5.6 also shows two estimates derived from the ANL BatPaC model for a liquid-cooled
BEV-sized pack  at a production volume of 50,000 to 100,000. Outputs from this model suggest
                                             5-123

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
that the ratio of pack-level cost to cell-level cost for the pack format modeled by BatPaC may
range from about 1.5 for a 16 kWh pack to about 1.3 for a 32 kWh pack, and continuing to
decrease for larger pack capacities.

                Table 5.6 Examples of Conversion Factors for Cell Costs to Pack Costs
Source
Kalhammeretal.340
Element Energy269
Konekamp248
USABC341
Tataria/Lopez342
Keller343
BatPaC, 16 kWh
BatPaC, 32 kWh
Low
1.24
1.6
High
1.4
1.85
129BB
1.25CC
1.26DD
1.2EE
1.5
1.3
   On the basis of the BatPaC-derived ratios of 1.3 to 1.5, the 2015-2019 cell-level figure of
$145 per kWh would translate to approximately $190 to $220 per kWh on a pack level. The
future projections of $120 and $100 per cell kWh in 2020 and 2022 would translate to
approximately $156-$180 per kWh and $130-$150 per kWh at the pack level, respectively.

   On this pack-converted basis the GM cell costs agree well with the BatPaC cost projections
that the 2012 FRM analysis applied to 2025. Table 5.7 summarizes the  estimated pack-level
equivalents of the cell costs disclosed by GM and compares them to the EV150 pack-level
BatPaC output costs of the FRM analysis.  The pack-converted GM projection for 2020, at $156-
$180 per kWh, compares well to the FRM BatPaC output costs for EV150FF for 2025, which
ranged from $160 to $175 per kWh (at 450,000 units annual volume). The pack-converted GM
projection for 2022 at $130-$150 per kWh is significantly lower than the agencies' projection for
2025. This suggests that the 2012 FRM cost projections, at least for EV150, may have been
quite conservative.
      Table 5.7 Comparison of GM/LGChem Pack-Converted Cell Costs to FRM EV150 Pack Cost

Source of Estimate
EV150 in FRM
GM/LG Global Business Conference

Year Applicable
2025
2015-2019
2020
2022
Pack Cost/kWh (2015$)
Low
$160
$190
$156
$130
High
$175
$220
$180
$150
   Figure 5.38 compares the pack-converted GM costs to the year-by-year learning-adjusted
costs used in the 2012 FRM for Small, Standard, and Large Car EV150. It can be seen that the
BB Cell cost = 620 Euros* 16 modules = 9,920 Euros; pack cost = 12,800 Euros; 12,800/9,920 = 1.29.
cc USABC 2020 goals for advanced EV batteries cite a cost of $125/kWh at pack level and $100/kWh at cell level =
  1.25.
DD For a 40 kWh pack, cell costs estimated at $258/kWh; pack-related costs at $2,626, or $66 per kWh;
  (258+66)7258=1.26.
EE Cites one goal of 21st Century Truck Partnership as "Cost of overall battery pack should not exceed cost of the
  cells by more than 20% by 2016" (slide 6).
FF The Chevy Bolt is anticipated to offer a 200-mile driving range, potentially comparable to the real-world 150-mile
  range of the EV150 that the agencies modeled in the FRM.
                                              5-124

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
range of the pack-converted GM costs is lower than the costs predicted by the 2012 FRM
analysis.
     400

     350

     300
   £250
   10
   -200
   _c

   -^  150
      5:

M--M
 * FRM

-A— GM/LGlow
       D
        2015  2016  2017 2018  2019  2020  2021  2022  2023  2024  2025  2026
                                     Year

    Figure 5.38 Comparison of Estimated GM/LG Pack-Level Costs to 2012 FRM Estimates for EV150

   At the time of the FRM, the agencies' battery cost estimates appeared to be lower than costs
being reported by many suppliers and OEMs at the time, and also lower than some independent
estimates said to be applicable to the time frame of the rule. The agencies chose to place
confidence in the peer-reviewed ANL BatPaC model due to its rigorous, bottom-up approach to
battery pack costing, and the expertise of leading battery research scientists that contributed to its
development. The comparisons described above suggest that this approach was effective and
may in fact have been conservative not only with respect to characterizing the pace of reductions
in battery cost that have taken place in the time since the FRM but also to projecting future costs
for the 2020-2025 time frame. Up to and including the development of this Draft TAR analysis,
the agencies have continued to invest significant resources into understanding developments and
emerging trends in battery technologies so that these critically important projections of xEV
battery cost may be as reliable as possible.

   While other public examples of battery costs to manufacturers remain elusive, several
suppliers and manufacturers have made battery-related product announcements since the FRM.
Some of these include information suggestive of battery costs or pricing.  Some manufacturers
have published pricing for battery replacement parts or upgrades available to authorized service
providers. Others have  offered  different options, such as battery size or purchase method, the
relative pricing of which may suggest a relationship to battery cost. Finally,  stand-alone non-
automotive Li-ion battery packs are beginning to become available to end users and their pricing
may be informative. While the  agencies recognize that the pricing of these early-stage product
offerings may be subsidized by  their manufacturers for competitive and marketing reasons, these
announcements may still be relevant to understanding the evolution of battery pack costs as these
products increase their presence in the market.
                                            5-125

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   In 2013-2014, Tesla Motors offered the Model S in two battery pack sizes, 60 kWh and 85
kWh, at retail prices of around $69,900 and $79,900, respectively. Assuming no content
difference between the two versions, the retail price differential would suggest a battery cost of
$10,000 / 25 kWh = $400/kWh. An alternate analysis presented by Nykvist et al.344 subtracts the
estimated value of added content found in the 85 kWh version (Supercharger, premium tires, and
associated markup), resulting in a net price difference of $8,500 or $340 per kWh.

   In July 2014, Nissan announced the replacement cost of a 24-kWh battery for the Nissan Leaf
at $5499 with core return, which amounts to about $229/kWh net. Although Nissan requires
return of the original battery  (core), a $1000 credit is then applied for the core, suggesting a full
retail price of $6499, or $271/kWh.345'346'347  Later the same month, Nissan followed up by
pointing out that the quoted price is in fact subsidized by Nissan, although they declined to report
the amount of subsidy or the actual manufacturing cost.348 Nissan does not allow purchase of the
battery except as a Leaf battery replacement.

   In 2015, an independent vendor of OEM parts listed the 2011 Chevy Volt battery pack at
$10,208 list price, discounted to $7,228, with no mention of core exchange.  Assuming a 16 kWh
capacity, these prices would value the battery at $638/kWh and $452/kWh, respectively.
Although the product was listed and priced by the vendor, it was on restriction from ordering for
reasons that remain unclear.349'350

   In January 2015, it  was reported that the MSRP for a BMW i3 battery pack module was listed
at $1,805.89, each module being 2.7 kWh (21.6 kWh total divided by 8 modules). This module
price would equate to  $669/kWh.  A specific dealer was reported to be offering the module at a
price of $1715.60, or $635/kWh.351

   In September 2015, Tesla announced the price for a range-increasing battery pack upgrade for
the Tesla Roadster at $29,000, including installation and logistics. Tesla indicated that the
quoted price is meant to be equal to Tesla's expected  cost in providing the pack, and disclaimed
any intention to make  a profit.  Tesla also indicated that the price per kWh is higher than for a
Model S battery due to the low volume production expected for the Roadster upgrade pack (only
approximately 2,500 Roadsters were produced). Tesla did not list the kWh capacity of the
upgrade pack, but describes it as having approximately 40 percent more energy capacity than the
original Roadster pack, which is commonly listed as 56 kWh. This suggests that Tesla's cost for
low volume production of this pack is around $29,0007(56*1.4) = $370 per kWh.352 In October
2015, Tesla further announced that the Roadster upgrade packs would be provided through a
partnership with LG Chem.353 This suggests that the price of the pack may not reflect
anticipated savings from the Panasonic-Tesla "Gigafactory" partnership.

   In August 2013, the Smart ED was offered with a  17.6 kWh battery, with the option to either
purchase the battery with the car, or lease it separately. The vehicle price was  $5,010 lower
without the battery when the battery was leased at a price of $80/mo.  If the $5,010 differential
was taken to represent the incremental cost of the battery, it would value the battery at
$285/kWh. Of course, the present value of the lease payments would also contribute value to the
transaction, and it is possible that marketing considerations could also be represented in the
pricing.354'355'356

   In September 2015, Nissan announced pricing in the UK for the 2016 Nissan Leaf. In a press
release from Nissan, equivalent versions of the Leaf having a 30 kWh pack instead of a 24 kWh
                                             5-126

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
pack were priced at a difference of 1,600 British pounds. This would amount to approximately
267 British pounds per kWh, or U.S. $411 per kWh (assuming an exchange rate of 1.54 U.S.
dollars per pound). It should be noted, however, that although the two versions of the pack
appear to be designed to install into the same footprint and volume, any cost comparison is
potentially complicated by differences in chemistry and construction of the two versions.357

   In 2014, Tesla Motors began construction of a so-called "Gigafactory" in Nevada in
partnership with Panasonic. This factory is commonly cited by Tesla as enabling a potential 30
percent reduction in battery pack costs from the levels  Tesla currently pays. According to one
analysis,358 Tesla's current cost is estimated at about $274 per kWh.  A 30 percent reduction on
that figure would bring costs to about $192 per kWh.

   In April 2015, Tesla announced a home battery pack product called Powerwall, pricing a 7
kWh version at $3,000 ($428/kWh) and a 10 kWh version at $3,500 ($350/kWh).  Although
designed for stationary home use, the pack design bears similarities to automotive packs, being
liquid-cooled and using similar chemistries. The 7 kWh version employs NMC chemistry
similar to many production BEVs, while the 10  kWh version employs the NCA chemistry like
the Tesla Model S. Tesla also announced a similar product called Powerpack for commercial
use. Powerpack was said to be priced at $25,000 for 100 kWh capacity, or $250/kWh.  These
products are expected to take advantage of much of the cell output of the Gigafactory, suggesting
that these products may be priced in anticipation of the cost reductions it is expected to achieve.
Table 5.8 summarizes the estimated cost or pricing information derived from the foregoing
examples.
             Table 5.8 Summary of Published Evidence of Battery Pack Cost and Pricing

Source of Evidence
Tesla Model S 60 kWh vs 85 kWh comparison
Nissan 24 kWh replacement pricing
Vendor pricing for 2011 Volt pack
Dealer pricing for BMW i3 module
Tesla Roadster upgrade pricing
Smart ED lease vs buy pricing
Nissan UK price differential 30 kWh vs 24 kWh
Tesla Lux Research estimate
Tesla Lux Research estimate modified by Gigafactory
Tesla Powerwall
Tesla Powerpack

Year Applicable
2013-2014
2015
2015
2015
2015
2013
2015
2014
2017
2015-2016
2015-2016
Pack Cost or Price
per kWh
High
$340
$229
$432
$635
Low
$400
$271
$638
$669
$370
$285
$411
$274
$192
$350
$428
$250
   It is important to remember that the figures derived from these examples should be interpreted
with caution. The agencies' cost projections represent direct manufacturing costs and not retail
pricing. Also, as previously noted, retail pricing of these early-stage product offerings may be
subsidized by their manufacturers and may reflect competitive and marketing considerations that
further obscure their true manufacturing cost.  Furthermore, some of the estimates are derived
from full-product comparisons that may or may not accurately represent the battery portion of
the comparison. It should also be noted that the examples presented here represent current
pricing, while the FRM applies its BatPaC cost projections to the year 2025.
                                             5-127

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   On the other hand, the existence of these examples shows that the industry has progressed
considerably since the FRM, when such examples were almost entirely unknown.  The
identification and packaging of specific battery products for upgrade, replacement or standalone
use is a significant development and suggests that the industry is continuing to gain in maturity
and is growing along multiple paths.  The establishment of MSRPs for many of these products
also suggests that manufacturers are beginning to gain confidence in their understanding of the
cost structure of battery products. The examples and estimates derived from this analysis, even
if approximate,  can serve to ground the various cost estimates and projections that have
previously been the primary source of battery costing information (and will continue to play an
important role going forward).

5.2.4.5 Fuel Cell Electric Vehicles

5.2.4.5.1     Introduction to FCEVs

   Fuel Cell Electric Vehicles (FCEVs) are another potential technology option for
implementing electrified drive to achieve zero tailpipe emissions, like the BEV technology
presented in Section 5.2.4.3.5. Like BEVs, FCEVs use electricity to turn electric motors onboard
the vehicle that  provide the motive power for driving. However, unlike a BEV, the FCEV also
produces this power onboard. It achieves this by harnessing the energy produced in an
electrochemical reaction that combines hydrogen and oxygen to form water.  This process occurs
within the fuel cell itself, a device that  shares a basic structure with batteries; namely, it consists
primarily of an anode, a dividing electrolyte, and a cathode. Hydrogen from an onboard tank
enters the fuel cell's anode and is separated into its constituent electron and proton. The electron
is directed to an external circuit, where it ultimately provides power to the electric motors driving
the wheels. The proton is transferred across the fuel cell's electrolyte membrane to the cathode,
where it combines with oxygen from air entering the cathode and electrons returning from the
external circuit to form water.  Thus, the basic reaction in the fuel cell is H2 + VaCh -^EhO, with
usable electric power (and  some amount of heat) produced in the process.

   State and national policies have increasingly adopted the perspective that FCEV and BEV
technologies will be complementary vehicle technologies that will likely both be needed in order
to achieve long-term GHG reduction goals. Well-to-wheel GHG emissions for FCEVs and BEVs
vary depending on the method of production for their various fuels (electricity for BEVs and
hydrogen for FCEVs), but both technologies hold  promise for significant reduction below
current and projected future ICE vehicle GHG emission rates (see Chapter 9, Infrastructure
Assessment for  a more complete presentation of GHG emissions from hydrogen production).
Hydrogen energy storage, the conversion of electrical energy into hydrogen gas through the
process of electrolysis, has recently gained significant attention for its potential to enable
increased renewable penetration in the  electric grid, thus potentially playing a significant role in
decarbonizing multiple industries in the full US energy  system. Although there is potential for
FCEVs to play a significant role in reducing GHG emissions, the technology is still relatively
new (the first mass-produced vehicles entered the  market in 2014) and costs have historically
been higher than other options. For this reason, FCEVs  were not included in the projections of
the future vehicle fleet in the 2012 FRM.

   The 2010 Technical Assessment Report (TAR) covered developments and state-of-the-art
technology for the FCEV at the time. Since then, researchers and developers in government,
academia, and industry have continued to advance the technology's performance capability and
                                             5-128

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
cost-competitiveness. This has enabled a transition in recent years away from a pre-commercial
technology demonstration phase to the early phases of full commercial product introductions.
Additionally, the year 2015 was a critical year in meeting national goals for the development of
FCEV technology advancement and commercial deployment. The year has long been an
aspirational goalpost, as captured in the Energy Policy Act of 2005.359

   "To enable a commitment by automakers no later than year 2015 to offer safe, affordable, and
technically viable hydrogen fuel cell vehicles in the mass consumer market."

   ".. .to enable a commitment not later than 2015 that will lead to infrastructure by 2020 that
will provide— (A) safe and convenient refueling; (B) improved overall efficiency; (C)
widespread availability of hydrogen from domestic  energy sources.

   The above provisions in the Act directly applied  to the US Department of Energy (DOE), but
have in actuality enlisted active participation by auto manufacturers, state and federal
governments, national labs, academic researchers, fuel and energy firms, engineering firms and
consultants, hydrogen production and distribution companies, public-private partnerships, and an
array of other industry participants. Based on these requirements, the Department of Energy has
long set cost and performance targets for FCEVs, hydrogen storage, and hydrogen fueling
technologies, and adjusted these goals in accordance with  developments in the state-of-the-art
technology.

   At the time of the 2010 TAR, the FCEVs that were on the road were part of auto
manufacturers' research and demonstration programs. Although many of these cars were
operated by private lessees, the models were not fully commercial products and the release of the
vehicles was much more carefully managed than full commercial sales. As of 2015, a great deal
of progress has been made towards the commercialization goals and the directives of the Act.
Two auto manufacturers, Hyundai and Toyota, have begun selling and/or leasing FCEVs directly
to the mass market. The first Hyundai  Tucson Fuel Cell Crossover vehicles were delivered to
customers in June 2014360 and the first Toyota Mirai sedans began delivery in October of
2015.361 Other auto manufacturers have announced  imminent plans for release of their own
mass-market, mass-produced FCEVs; Honda has made indications that it will be the next auto
manufacturer to bring a vehicle to market with its Clarity Fuel Cell expected sometime in
2016.362

   Commercial releases of mass-produced FCEVs intrinsically rely on the availability of a retail
hydrogen fueling network to support the needs of the FCEV drivers. California has had the
longest experience with deploying and operating fueling stations. However, at the time of the
2010 TAR, the network in California included only  a handful of stations with public access, and
these stations were primarily research and/or technology demonstration stations. Many retail
features were not included in these early stations. Recent progress in the development of station
technology and deployment has moved infrastructure development in California towards retail
service stations. The recent commercial vehicle releases have been well-timed to the
development of this retail fueling infrastructure network; in California there is now a network of
51 stations currently funded and in development, with continued annual State funding expected
beyond 2020.363 For a more complete review of the  status  of hydrogen fueling infrastructure
development, see Chapter 9, Infrastructure Assessment.
                                             5-129

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Challenges remain in FCEV and fueling station technology and implementation, but progress
since the agencies reported in the 2010 TAR has helped the industry mature out of a pre-
commercial and demonstration phase into the first stages of the retail, mass-market phase. Many
of the previous targets have been met or exceeded and new targets, tied to production volumes
rather than specific timeframes, are now in place at the Department of Energy. These
developments have also allowed an escalation of the industry-wide dialogue of plans for
deployments and development nationwide, as opposed to the singular focus that has historically
been placed on the demonstration and nascent pre-commercial market in California alone. Cost
remains one of the major challenges for both the vehicles and fueling  infrastructure. Federal and
State financial incentive programs are currently in place to help meet the cost challenge, and it is
likely that these incentives will need to remain and expand as the commercial market develops,
similar to the  national experience with BEVs.

5.2.4.5.2     FCEV Cost Estimation

   Since FCEVs are electric-drive vehicles, they share many of the same types of components as
hybrid vehicles and full BEVs. In fact, it is anticipated that auto manufacturers that choose to
pursue multiple drive train technologies among these three options may  implement similar, if not
exactly the same, components whenever possible among FIEV, BEV,  and FCEVs in order to take
advantage of manufacturing efficiencies and benefits of scale in the supply chain. However,
there are three main subsystems that the FCEV does not share with other vehicles: the fuel cell
stack, air and  fuel delivery sub-systems, and the hydrogen  storage system. Although exact direct
manufacturing costs for individual  auto manufacturers' designs are proprietary information, the
Department of Energy has for a number of years supported work estimating the direct
manufacturing costs of these components. This work was cited in the  2010 TAR, published
through Directed Technologies, Inc.364 Since that time, Directed Technologies has been acquired
by Strategic Analysis, Inc. (SA), who continues to publish  annual updates to their estimates.
These estimates are a critical resource in estimating the potential costs of FCEVs, much in the
way that BatPaC is used to estimate the direct manufacturing costs of xEV batteries for the
purposes of the Draft TAR. In order to complete its analyses, SA adopts a Design for
Manufacturing and Assembly (DFMAGG) analysis method that captures optimized material and
processing costs at varying production rates.
  ' DFMA is a registered trademark of Boothroyd Dewhurst Incorporated.
                                             5-130

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
             Potential cost reduction based on DOE targets
      $60
      $50
      $40
      $30
      $20
      $10
       $0
c  £
*:  <->
                         *
                         -r
                         on
                         oo
       2014 System Cost
       at 500k Sys/Year
                                  c  s
                                  OJ -^
4-  LO
Q-  fS
c  --1
c  °
o  £
          -c
           QJ
          QC
                                            i/l
                                            O
                                            —
                                            U O
                               8
                               u
                               Q.
                               0.
                               CO
                                 ro
                                 w
                                  o
                                                      TJ
                                                      (V
                     "O
                     u
                     DC
           "value in SA 2014 baseline
                                                                c  *
                                    I Future Lower
                                      Cost System
       Note: DOE has since published an updated estimate for 2015 of $53/kW at 500k Sys/Year.

              Figure 5.39 Projection of Potential Cost Reductions for Fuel Cell System365

5.2.4.5.2.1    Fuel Cell System Cost

   The SA estimates allow the DOE to measure progress towards its cost reduction goals and
provide open and public analysis of the costs of materials and manufacturing processes for fuel
cell stacks, hydrogen storage tanks, and related balance of plant. The analyses provide detailed
information on the individual processes for nearly all components and estimated costs for
conventional and demonstration technologies. The 2014 analysis366 and 2015 update367 estimated
that current fuel cell system technologies, at high production volumes for a representative 80 kW
net power FCEV would cost $55/kW (not including the hydrogen storage system). With
advances currently  available or anticipated in the near-term, the cost can be potentially reduced
to $40/kW, meeting the DOE 2020 system cost target, which is based on achieving cost-parity
between FCEVs  and hybrid vehicles.368'369'370 Note: DOE has since published an updated
estimate for 2015 of $53/kW at 500k Sys/Year.
   Figure 5.39 Figure 5.39 provides an overview of the current system estimate and possible
steps to achieve the prospective lower-cost system.  The steps shown in the figure should not be
interpreted as the only or even the ideal route to a lower cost system; rather, it is a sample
                                             5-131

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
pathway for future development and other improvements may provide at least the same
improvement in cost. Figure 5.40 provides a breakdown of contributions to cost from raw
materials and individual manufacturing steps in the production of the catalyst, one of the most
expensive components of the fuel cell stack. The differing responses of each material and
process to increased volume production are apparent in the example.
             Material Cost (1,000 Systems/Year)

                3%
Material Cost (500,000 Systems/Year)
Ammonium
Persulfate
Iron (III)
Chloride
 Hydrochloric
 Acid
        Aniline
                                • Nitric Acid

                                • Carbon

                                • Iron (III) Chloride
                                • Water

                                • Hydrochloric Acid

                                : Aniline

                                 Ammonium Persulfate

                                 Sulfuric Acid
             Manufacturing Cost (1,000 Systems/Year)
                 1%
                               • Step 1: Carbon Activation

                               • Step 2: Catalyst Reaction

                               • Step 3: Belt Drying

                               • Step 4: Grinding

                               • Step 5: Rotary Kiln Pyrolysis

                               • Step 6: Acid Leaching

                                Step 7: Oven Pyrolysis
                   • Nitric Acid

                   • Carbon
                   • Iron (III) Chloride

                   • Water
                   • Hydrochloric Acid

                   • Aniline

                    Ammonium Persulfate
                   • Sulfuric Acid
Manufacturing Cost (500,000 Systems/Year)
           0.1%
              0.5%
                                                                     • Step 1: Carbon Activation

                                                                     • Step 2: Catalyst Reaction

                                                                     • Step 3: Belt Drying

                                                                     • Step 4: Grinding

                                                                     • Step 5: Rotary Kiln Pyrolysis

                                                                     • Step 6: Acid Leaching

                                                                      Step 7: Oven Pyrolysis
Figure 5.40 Cost Break-Down for Catalyst in an 80kw Fuel Cell System at 1,000 And 500,000 System Annual
                                      Production Rates371

   In addition to the detailed DFMA analysis, SA provided a simplified model of total fuel cell
system cost in its 2014 report, based on system design and operational parameters that could be
readily determined by a fuel cell system engineer.372 The simplified cost model was broken down
into fuel cell stack, thermal management system,  humidification management system, air
management system, fuel management system, and balance of plant contributions to total cost
(hydrogen storage costs are treated in a separate simplified model, discussed below).  Combined,
the simplified system's costs require the specification of 14 individual parameters. Baseline
values for these parameters that match the cost estimate for SA's 80kW representative system
can all be interpreted from the data within the report. However, there are certain details of the
80kW system that do not match well with  systems in FCEVs currently available or anticipated in
the next few years. Of particular note is the system net power; FCEVs coming to market are
nearly uniform in providing a system with lOOkW net power.

   To evaluate incremental costs for FCEV systems in this Draft TAR, CARB performed a study
of FCEV system costs based on the simplified cost model from SA with scaling and re-
parameterization in order to generate cost estimates for a lOOkW net power system. First, a linear
scaling relationship was  assumed between net power and many of the 14 variables in the
                                                5-132

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
simplified cost model. For example, cell active area was one of the variables assumed to scale
with power; however, neither the unit cost of platinum nor the peak air pressure in the system
was assumed to scale with power. Fuel cell system costs were then calculated for varying net
power and system production volumes (the effect of which was modeled after the trend between
the cost and production volume for the 80kW system presented in the SA report). System cost
was then parameterized according to best-fit relationships with production volume and net power
assumed as independent variables, the contributions of which were multiplicative. It was found
that a curve based on a power law relationship best fit the variation in system cost with respect to
production volume and an exponential curve best fit the variation with respect to system power.
It should be noted that these were derived from a parametric examination for best fit; no
underlying mechanism was assumed to lead to these relationships. Thus, system cost was
described in the form:
       Equation 1. Fuel Cell System Cost

                 Fuel Cell System Cost = A * ^production * exP(C * ^Vet)

   Where A, B, and C are best-fit coefficients, with A, C > 0, B < 0, and Vproduction is the annual
production rate, and PNet is the system net power.

   Figure 5.41 shows steps of the re-parameterization process,  including the variation in system
cost according to annual production rate at various system net powers, the complementary
parameterization (variation in system cost according to system net power at various annual
production rates),  and the surface of projected costs accounting for both variations. Note that
these costs are only for the power-producing fuel cell  system and its balance of plant
components; these costs do not include the hydrogen storage tank(s) and its balance of plant. Due
to the use of curve-fitting in the process (A = 70497.1, B  = -0.26055, and C = 0.0056), there is
some deviation for a specific system from the re-parameterization when compared to the original
SA data. However, for an 80kW system at 100,000 systems per year, the deviation is less than 5
percent. Additionally, the results demonstrate the need to re-parameterize the system costs in
order to be more in-line with technology seen in today's on-the-road FCEVs. For example, at
100,000 systems per year, an 80kW system is projected by this analysis to cost approximately
$5,500; a lOOkW system would cost approximately $6,200.
                                             5-133

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
         100     200     300     400     500
             Annual Production Volume (Thousands)
                System Net Power (kW):
       --60 -»-70 . 80 —90 —100  • 110   120
          40     60     80     100    120
              System Net Power (kW)
            Annual Production Volume:
-500000 -»-100000  A 80000 —30000 —10000  • 1000
                           System Net Power (kW)
                                                                     1000
                                                                   10000
                                                                30000
                                                             80000  Annua] production
                                                          100000       Volume
                                                110   i SOOOOO
Figure 5.41 Parameterization of SA Fuel Cell System Cost Analysis (Not Including Storage Tanks)
                     According To Production Volume and System Net Power
                                                  5-134

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                   700 Bar CF Pressure Vessel Storage System Cost
                                             $1.35
                                                           $0.70
                                                            5%
               2013 Record   ORNL Low Cost  2014 and Prelim.   PNNLMuIti-
                              PAN MA CF    2015 BOP Cost  Approach Prelim.
                                (2015)        Reduction     Cost Reduction
                                            (2014-2015)       (2015)
$12.99
                            Total of 23% cost reduction (prelim.)
  Goal of
  >15%
 reduction
  from
 Potential
Lower Cost
   Figure 5.42 Projection of Potential Cost Reductions for 700 Bar Compressed Hydrogen Storage Tank
                                       System373

5.2.4.5.2.2    Hydrogen Storage Cost

   SA also performs a complementary analysis of the costs for the on-board hydrogen storage
system and balance of plant. Like the fuel cell system analysis, SA performs a fully detailed
analysis of the predominant or conventional technology and provides estimates for emerging or
new technologies and compares the costs to DOE goals. As of 2014, SA estimates that 700 bar
compressed gaseous storage vessels made from carbon fiber-wrapped polymer cost $16.76/kWh
(approximately $660/kg storage capacity).374 With available or emerging technology
improvements, the cost could be reduced to $12.99/kWh (approximately $510/kg). This cost is
above the DOE 2020 target375, but is noteworthy for representing a reduction greater than the
DOE's hydrogen storage program's midterm milestone of 15 percent reduction from the 2013
cost estimate. The columns of incremental cost reduction in Figure 5.42 outline the technological
advances that may make this lower cost system possible.

   In addition to the DFMA analysis reported, SA has developed draft simplified cost models for
the hydrogen storage tank and storage balance of plant costs, parameterized according to the tank
volume and pressure (for tank costs) and the number of tanks (for storage system balance of
plant costs).  SA has shared these simplified cost models (for 10k, 200k, and 500k system annual
production rates) with CARS.376

5.2.4.5.2.3    Combined Fuel Cell and Hydrogen Storage Systems Cost

   The cost models for fuel cell and hydrogen storage systems were combined for a FCEV
system cost model.  CARB adopted the point estimates from the SA work directly and assuming
                                            5-135

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
piecewise linear fits between estimates for the 10k to 200k and 200k to 500k portions of the cost
curves, separately. CARB then performed a parametric analysis for FCEV costs (stack, tank, and
their respective balance of plants) of possible systems within the SA model domain for net
power, production volume, number of tanks, and total kg storage to investigate the possible
range of costs across the design space available to FCEV system engineers.  The ranges for all
variables are provided in the "All Possible Designs" column of Table 5.9.
   Table 5.9 FCEV System and Production Rate Input Parameters for Assessment of Potential Costs For
                          CARB-Modified SA Simplified Cost Models

Parameter
System Net Power (kW)
Annual Production Volume (lOOOs/year)
Number of Tanks
Total Storage (kg)
All Possible Designs
Minimum
60
1
1
0.4
Maximum
120
500
4
11
TAR Representative Designs
Minimum
100
3
1
4
Maximum
100
50
2
5
   Ranges for some of the variables specified in Table 5.9 are wider than realistically expected
for production vehicles; however, the wider ranges provide a fuller perspective of the potential
sensitivity of total FCEV costs. Calculated full FCEV system cost ranges and average values
(incorporating the fuel stack costs shown in Figure 5.41  and the SA-provided tank and tank BOP
costs) are provided in Figure 5.43 as a function of annual production rate. The costs shown are
indicative of a system with 2014 technology; the range of production volumes are similar to
today's volumes on the lower end and on the high end may be greater than volumes expected in
2025 (as will be discussed further below).  As in the SA estimates, there is a strong dependency
of total system cost on the annual production volume. Additionally, there is a fairly significant
difference between the cost estimates of the most and least expensive vehicle designs; at all
production rates, the most expensive system design costs approximately 30 percent more than the
least expensive  option. However, the distribution of prices at a given production rate was also
more heavily weighted towards the higher costs, given that the mean was consistently closer to
the maximum rather than the minimum (though this association decreased with increasing
system production volume).
                                             5-136

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
     $30
      $0
                Max: $25.58
                Avg: $24.07
                Min:$19.96
                                                                                Max: $8.49
                                                                                 Avg: $7.76
                                                                                 Min:$6.66
                55      105      155     205     255     305     355
                                  Annual Production Volume (Thousands)
                           405
                                   455
                                 Maximum
Minimum   Mean
   Figure 5.43  Combined Fuel Cell and Tank System Cost Estimates across Design Space of All Possible
                       Systems within Domain of SA Simplified Cost Models
   Although the values presented in Figure 5.43 are useful for understanding the potential
sensitivities in FCEV system cost to system design parameters and production rates, the
estimates are not quite representative of vehicles expected in the near term. For example, no
vehicles are yet designed with storage divided between four cylinders; two tanks is the current
industry norm.  Inclusion of non-representative system designs may skew the aggregate
estimates, providing misleading system cost estimates. Therefore, CARB  performed a secondary
analysis with a narrowed system design space to vehicles more closely matching current
expectations, as shown in the "TAR Representative Designs" column of Table 5.9. Figure 5.44
provides the cumulative mean costs from this much narrower set of system designs. In contrast
to Figure 5.43, Figure 5.44 does not include the range of values since the variation at a given
production volume was very small  due to the smaller design space. Additionally,  Figure 5.44
provides individual  costs for the tank, tank balance of plant, and fuel cell system (inclusive of
stack and its balance of plant).  According to the parametric study, fuel cell system plus
hydrogen storage costs for representative vehicles range from just over $20,000 at 3,000 vehicles
per year to $6,730 at 500,000 vehicles per year.

5.2.4.5.2.4    Marke t Projections

   Multiple projections for regional and global FCEV sales (and by inference production) rates
have been presented in  past literature, including the ORNL377 and NAS378  estimates discussed in
the 2010 TAR and updated estimates based on continuing work.379'380  However, as the
commercial launch of vehicles has  neared and the potential growth rate in  necessary supporting
                                              5-137

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
infrastructure has become more apparent, new trends have emerged. In particular, CARS's
assessments of projected growth in infrastructure and FCEV population in California (one of the
larger anticipated early adopter markets) show significant differences from previous in-state
estimates like those presented in the California Fuel Cell Partnership Roadmap.381'382

     $25.0
   2" $20.0
   oo
   1
   1 $15.0
     $10.0
   3
      $5.0
      $0.0
                 55     105     155     205     255     305     355
                                 Annual Production Volume (Thousands)
405
455
              —Tank Costs     Tank BOP Costs   —Fuel Cell System Costs (not including tanks)

  Figure 5.44 Mean Costs for All Possible Delineated Systems With Up To Two Tanks, Between 4 and 5 kg
               Onboard Storage, lOOkW Net Power, And At Least 3,000 Units per Year
   Based on its analysis showing a potential power law growth in the California FCEV stock out
to 2021, CARB estimated the global early adopter market for FCEVs. First, the power law
presented in the report was extrapolated out to 2025 for California. Annual changes in on-the-
road vehicles were then assumed to be roughly equal to new car sales (strictly speaking the
CARB on-the-road analysis includes vehicle attrition, but at the small volumes for FCEVs the
absolute number of vehicles leaving the fleet is not very large).  The annual California-specific
FCEV sales were then compared to total light duty vehicle sales projections in CARB's
EMFAC2014383 on-road emissions inventory model. For every year from 2014 to 2025,
estimates were thus generated for the California FCEV share of new light duty vehicle
purchases, which grows to approximately 1.7 percent in 2025.

   California, the United Kingdom, Germany, Japan,  and Korea are the five main regions that
FCEV and hydrogen industry stakeholders generally agree are expected to  comprise the majority
of the global early FCEV adopter markets. This market identification is also supported  by
numerous government and industry announcements regarding prospective vehicle launches and
investments in supporting infrastructure. For the sake of this analysis, CARB assumed that the
FCEV market share would grow in each of these market areas at the same rate calculated for
                                             5-138

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
California. IHS384'385 and ACEA386 data and documentation were relied on to estimate the full
light duty vehicle sales projections in each region out to 2025. The California-based FCEV
market share growth curve was then applied to each region's new vehicle sales projection to
estimate the global FCEV sales.

   Figure 5.45 shows the CARB-estimated FCEV new vehicle sales in California and globally
from 2014 to 2025 and the share of total new sales that these FCEV projections represent in
California and globally. Global estimates are based on the IHS projection of new vehicle sales to
2021, and then extrapolated linearly from 2022 to 2025. IHS-based data predict global new auto
sales will increase from approximately 86 million in 2014 to 122 million in 2025. Over the same
period, California's annual FCEV sales are projected to grow from approximately 25387 to  nearly
37,000 in 2025; global FCEV sales will grow to approximately 273,000 in 2025. In 2021,
California and global annual sales are projected to be 10,000 and nearly 83,000 respectively. As
a point of comparison, Toyota alone has publicly announced a goal of producing 30,000 FCEVs
by 2020; with increasing participation from other manufacturers, the projections of 83,000 in the
same timeframe appear consistent. Assuming global production volumes for cost estimates,
using the data shown in Figure 5.44 above, 2021  direct manufacturing costs for FCEV systems
are projected to be approximately  $12,200 which represents a cost in addition to manufacturing
the remainder of the vehicle and its systems (such as the body, electric motors, battery, etc.; 2025
FCEV systems are projected to have direct manufacturing costs of approximately $8,000.
_ 1,000
l/>
   re
   3
   o
   c
   o
       100
        10
   U     1
  •^     1
  2   0.1
   re
   to
~  0.01
     0.001
                                                                           1.5%
                                                                              2.5%
                                                                           2.0%
re
to
in
                                                                                   re
                                                                                   to
                                                                              1.0%
                                                                                 op
                                                                                 '_!

                                                                                 LU
                                                                                 u
                                                                             0.5% 3
                                                                           0.0%
                                                                                   u
                                                                                   
-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Technological status of FCEV components and systems continue to advance as commercial
launches begin globally. In the 2010 TAR, then-current technological performance status was
presented alongside some of the key targets for 2015 technology as defined by DOE. The status
values from the 2010 TAR are reproduced in the "2010 TAR" column of Table 5.10, alongside
current status values and current DOE target values. The targets shown in Table 5.10 are those
most directly affecting FCEV system-wide performance; DOE additionally sets several more
detailed cost and performance metrics that are not shown.

   As shown in Table 5.10, the 2014 status demonstrates significant progress since the 2010
TAR. Notably, the previous vehicle range target has been met and exceeded; at the current time,
there is no updated range target as commercial FCEV range has achieved relative parity with
conventional vehicles.  Additionally, costs have improved in the intervening years without any
projected loss in system efficiency or durability. New targets have been set for fuel cell system
efficiency, indicating a push to achieve performance even beyond the original program goals and
maintain the goals' price and performance parity with future hybrid vehicles. Note that the
Ultimate DOE Targets are not strictly  defined according to a timeframe; they are goals to be
achieved in order for full  fleet penetration of FCEVs across various manufacturers, models, and
vehicle classifications.
   Table 5.10 Updated DOE Status and Targets for Automotive Fuel Cell and Onboard Hydrogen Storage
                                       Systems388'389'390

System Efficiency
System Cost
Fuel Cell System Durability
Vehicle Range
H2 Storage Costs

2010 TAR
53-59%
$61/kW (SSl/kw)1
2,500 hrs
254 miles
$20/kWh
2015 TAR
2014 Status
60%
$55/kW ($43/kW)'
3,900 hrs
312 miles"
$15/kWh (SlS/kWh)1"
2020 DOE Target
65%
$40/kW
5,000 hrs

$10/kWh
Ultimate DOE Target
70%
$30/kW
8,000 hrsiv

$8/kWh
Notes:
(i) 2010 TAR value includes the then-current 2009 reported status and the 2010 update in parentheses. The 2014
includes the reported current cost status and a potential reduced cost based on available or near-term technologies in
parentheses. DOE has additionally reported a 2015 updated estimate of $53/kW.
(ii) Based on US EPA rating for the 2015 Toyota Mirai.
(iii) September 2015 DOE records reports $15/kWh; contact at 2015 AMR indicated the potential for reduction to
$ 13/kWh in very short term with application of technologies within DOE's funded Program.
(iv) Based on March 2016 communication from DOE Fuel Cell Technology Office.

   The Hyundai Tucson Fuel Cell (known as the ix35 in the global market) became the first
mass-produced fuel cell vehicle to enter the market,391 indicating the development of
manufacturing techniques and methods sufficient for full-scale early production volumes.
Announcements from Honda indicate that it has continued to innovate for its planned vehicle
release in 2016 by increasing power density more than 60 percent compared to the previously-
released FCX Clarity392 which allows an overall 33 percent reduction in the fuel cell stack
volume. For the newly released Mirai vehicle, Toyota was able to eliminate the humidifier
necessary in conventional fuel cell system designs by developing a Membrane Electrode Gas
                                              5-140

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Diffusion Layer Assembly that promotes self-humidification.393 Key to the development was the
design and implementation of a 3-D Fine Mesh Flow Field on the cathode, and a counter-flow
field design for hydrogen and coolant on the anode, that promote the necessary exchange of
reactant gases and product water within the cell and eliminate the need for the external
humidifier.  Toyota and Honda have also announced that their vehicles will have the ability to
export power generated by the vehicle's fuel cell, allowing owners to power their homes when
grid power may not be available for extended periods of time394'395 and increasing the FCEV
customer value proposition.
5.2.4.5.4
Onboard Hydrogen Storage Technology
   Current FCEV designs rely on compressed gaseous hydrogen for onboard storage of the fuel.
In the past, two pressures had been pursued by the majority of auto manufacturers: 350 bar and
700 bar (equivalent to 35 MPa and 70 MPa, respectively). As development has progressed, the
auto industry has predominantly converged on designs for 700 bar storage, as this pressure
allows for increased FCEV range.  Cost status for onboard storage is presented in Table 5.10.
Table 5.11 provides further detail of the technical performance status of 700 bar compressed
hydrogen storage, along with other options and the current 2020 and ultimate targets specified by
DOE. Although 700 bar compressed storage does not yet meet cost and performance targets, it is
the most feasible among the options currently being developed and does provide  sufficient range
for vehicles. However, for many reasons (including system complexity of refueling stations and
reductions in overall fuel lifecycle efficiencies when compressing to high pressures), there is an
interest in developing technologies that can achieve the cost and performance targets while
avoiding some of the challenges of 700 bar compression. The metal hydride,  sorbent, and
chemical storage methods all show promise for achieving these goals but are much earlier in
their development and not yet implemented today.
    Table 5.11 Hydrogen Storage Performance and Cost Targets and Status for Various Technologies396
Storage Technology
2020 DOE Target
Ultimate DOE Target
700 Bar Compressed
350 Bar Compressed
Metal Hydride
Sorbent
Chemical
cost (S/kWh),[S/kg]
10, [333]
8, [266]
15
13
43
15-16
17-22
Gravimetric Density (kWh/kg),
[kgH^kg system]
1.8, [0.055]
2.5, [0.075]
1.5
1.8
0.4
1.2
1.1-1.5
Volumetric Density
(kWh/L), [kgH2/L]
1.3, [0.04]
2.3, [0.07]
0.8
0.6
0.4
0.6-0.7
1.2-1.4
5.2.4.5.5
FCEV Commercialization Status
   Currently, three automakers (Hyundai, Toyota, and Honda) have begun to offer fuel cell
vehicles to the mass consumer market or announced specific near-term plans for market launch.
Hyundai has offered its Tucson Fuel Cell for lease in select regions of southern California since
2014. Toyota offers its Mirai sedan in at least eight dealerships across both northern and southern
California with options for both lease and purchase. Honda has unveiled its production Clarity
Fuel Cell at the Tokyo Auto Show in October 2015 and announced plans for a 2016 release.
Other automakers are known to be involved in the development of FCEV technology and
                                             5-141

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
expected to be moving towards commercial production, but have not yet made public
announcements of production models or release dates.

   In addition to the release of the first three mass-market FCEVs, many automakers have made
public announcements of other activities related to FCEVs. A number of automakers have signed
agreements to cooperatively work on development of their fuel cell systems and vehicles. BMW-
Toyota, Daimler-Ford-Nissan, and GM-Honda partnerships have been announced.397'398'399
Lexus, Toyota's luxury brand label, recently announced that its LF-LC concept is the precursor
to the next LS model and is expected to include a fuel cell-powered all-wheel drivetrain.400 This
development is notable for possibly being the first  announcement of a brand's flagship vehicle as
an FCEV. BMW recently unveiled a fuel cell prototype of its i8 sports coupe.401  Audi
announced a fuel cell version concept, the A-7 Sportback h-tron Quattro, which is unique among
current developments for being a fuel cell-powered plug-in hybrid.402

   Collectively, these releases, partnerships, and announcements signal progress and
commitment from the automotive industry towards the launch of a mass-consumer FCEV
market. Many automakers and industry  experts often caution that the eventual success of the
FCEV market will depend heavily on the successful and widespread implementation of hydrogen
fueling infrastructure. Automaker FCEV launches and production rates are likely to be closely
tied to the deployment rates of fueling infrastructure and will require that fueling infrastructure
development precede vehicle launches.  There is currently broad support for this strategy,
especially among regions where the first adopter market is anticipated to be large (California,
UK, Germany, Japan, and Korea). Public and private actions have in recent years helped to
accelerate much-needed activity in the fueling infrastructure industry. A more thorough
discussion of this dynamic is presented  in the Chapter 9 section on Hydrogen Infrastructure.

5.2.4.5.6     Outlook for National FCEV Launch

   Compared to the status reported in the 2010 TAR, FCEVs have progressed substantially,
transit!oning from a demonstration and  pre-commercial phase into the inception of commercial
launches. This has been aided by the technological and business advancements discussed above
(as well as many more) and has been reinforced by supporting policy actions, public-private
partnerships, and broad stakeholder initiatives toward cleaner transportation choices.
California's ZEV Mandate, Alternative  and Renewable Fuel and Vehicle Technology Program,
and multiple renewable energy and GHG reduction goals have and will continue to incentivize
the adoption of FCEVs alongside other  alternative  vehicle options like BEVs. Nationally,
California's ZEV regulations have been adopted by an additional 7 states (Connecticut,
Maryland, Massachusetts, New York, Oregon, Rhode Island, and Vermont), collectively
developing an Action Plan with the goal of enabling 3.3 million cumulative sales of ZEVs and
PHEVs within those states by 2025.403 Additionally, California, Connecticut, Maryland,
Massachusetts, Rhode Island, and Vermont joined with The Netherlands, Norway, the United
Kingdom, Quebec, and other jurisdictions in forming the International ZEV Alliance.404 The
Alliance has broad goals of accelerating global adoption of ZEVs, including FCEVs.

   Through these actions, the west coast and the northeast states are leading early market
adoption efforts for ZEVs broadly. In addition, California's AB 8 ensures funding is available
(up to $20 million a year) specifically for investments in hydrogen infrastructure to encourage
the role of FCEVs in meeting ZEV goals.  A more thorough  discussion is presented in Chapter 9,
Infrastructure Assessment.  Stakeholders have also begun developing plans to support the
                                            5-142

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
necessary infrastructure for an FCEV launch in the northeast states.  Connecticut has offered
grant funding for up to two stations near Hartford and multiple states in the region have
leveraged resources available through the DOE-initiated public-private partnership H2USA to
develop detailed infrastructure network planning. Well-planned growth of infrastructure in local
early markets, that anticipates integration into larger regional and ultimately national networks,
will be essential for ensuring FCEVs significantly contribute to the goals outlined by the multiple
ZEV-related State initiatives.

5.2.5  Aerodynamics: State of Technology

5.2.5.1 Background

   Aerodynamic drag accounts for a significant portion of the energy consumed by a vehicle, and
can become the dominant factor at higher speeds. Reducing aerodynamic drag can therefore be
an effective way to reduce fuel consumption and GHG emissions.

   Aerodynamic drag is proportional to the frontal area (A) and coefficient of drag (Cd) of the
vehicle. The force imposed by aerodynamic drag increases with the square of vehicle velocity,
accounting for its dominance at higher speeds.

   The coefficient  of drag Cd is a dimensionless value that essentially represents the aerodynamic
efficiency of the vehicle shape.  The frontal  area A is the cross-sectional area of the vehicle as
viewed from the front. It acts with the coefficient of drag as a sort of scaling factor, representing
the relative size of the vehicle shape that the coefficient of drag describes. Because the two
values are related in this way, the aerodynamic performance of a vehicle is often expressed as the
product of the two  values, CdA (also known as drag area).

   Cd and A are determined by the design of the vehicle, and so represent the primary design
paths for reduction of aerodynamic drag. The greatest opportunity for improving aerodynamic
performance is during a vehicle redesign cycle, when the best opportunity exists to make
significant changes to the shape or size of the vehicle. Incremental improvements may also be
achieved mid-cycle as part of a model refresh through the use of revised exterior components
and add-on devices. Some  examples of these technologies include revised front and rear fascias,
modified front air dams and rear valances, addition of rear deck lips and underbody panels, and
low-drag exterior mirrors.

   Aerodynamic technologies are  divided into passive and active technologies. Passive
aerodynamics refers to aerodynamic attributes that are inherent to  the shape and size of the
vehicle, including any components of a fixed nature. Active aerodynamics refers to technologies
that variably deploy in response to driving conditions. These include technologies such as active
grille shutters, active air dams and active ride height adjustment.

   Significant variations in CdA can be observed across vehicle classes and among individual
vehicles within a class.405'406'407 Within a class, drag coefficients tend to vary more than frontal
areas. Frontal areas are in part a function of interior passenger and cargo space, and therefore
tend to track with the interior space expectations associated with a vehicle class. In contrast,
drag coefficients are largely a function of body styling and may vary significantly with relatively
small changes in shape and exterior treatment.

5.2.5.2 Aerodynamic Technologies in the FRM
                                              5-143

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Based in part on the 2011 Ricardo study and public technical literature, the FRM analysis
projected that a 10 to 20 percent fleet average reduction in aerodynamic drag should be
attainable. Based on EPA vehicle modeling and the Ricardo study, each 10 percent reduction
was associated with an incremental reduction in fuel consumption and CCh emissions of 2 to 3
percent for both cars and trucks.

   The FRM considered two levels of aerodynamic improvements, called Aerol and Aero2.  The
first level, Aerol, represented a 10 percent reduction in drag from the baseline by means of
passive body features such as front/rear bumper air dams, front and/or rear wheel tire spats/dams,
minimal underbody  panels, and redesigned mirrors or rear spoilers. Aerol was estimated to
result in an effectiveness of 2.3 percent for all vehicle classes. The agencies estimated the DMC
of Aerol at $41 (2010$) applicable in MY2015. The second level, Aero2, represented a 20
percent reduction from the baseline (nominally 10 percentage points incremental to Aerol), and
included active technologies such as active grille  shutters and active ride height, as well as
passive technologies such as rear visors, larger under body panels and low-profile roof racks.
Aero2 was estimated to provide an effectiveness of 4.7 percent relative to a baseline vehicle.
The agencies estimated the DMC of Aero2 at $123 (2010$) incremental to Aerol, applicable in
MY2015.

   In the FRM analysis, fleet penetration of Aerol was uncapped for 2012 through 2025. Fleet
penetration of Aero2 was capped at 80 percent for 2021 and uncapped thereafter.

   Because the full benefit of active aerodynamic technologies may fail to be reflected in
standard test cycles, the agencies provided for active aerodynamic technology to be eligible for
credit under the Off-Cycle Credit Program. Off-cycle credits are discussed in a separate chapter
of this Draft TAR.

5.2.5.3 Developments since the FRM

   Since the FRM, the agencies have taken several steps to further evaluate the feasibility, cost
and effectiveness assumptions of Aerol  and Aero2. We followed industry developments and
trends in application of aerodynamic drag technologies to light-duty vehicles. We did this by
gathering input from stakeholders through meetings with OEMs, suppliers and other interested
parties, and also by attending conferences and trade shows and regularly monitoring the press
and technical literature.

   EPA also participated in a joint test program with Transport Canada, Environment and
Climate Change Canada, and National Research Council Canada to examine the aerodynamic
performance and effectiveness of various aerodynamic devices and strategies. This program was
conducted in four phases over three years, and examined aerodynamic technologies as currently
implemented in a selection of production vehicles, and the effectiveness of potential
improvements that have yet to be implemented. Results of this program also were used to
evaluate the 2012 FRM assumptions about off-cycle benefits of active aerodynamic technologies
and the associated default credit values.

   Additionally, EPA coordinated with California Air Resources Board (CARB) to share the
results of a research study performed for CARB by Control-Tec, a company that specializes in
automotive data analytics. This study is described in more detail in Appendix A, "CARB
Analysis of Vehicle Load Reduction Potential For Advanced Clean Cars." The study provided
                                             5-144

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
information helpful to assess the penetration of aerodynamic and other road load technologies in
the MY2014 fleet as represented nationally and in California. EPA also began a process to
compare the fleet aerodynamic performance of MY2014 vehicles as represented in the study to
those of MY2008 by using EPA certification data to estimate aerodynamic performance of the
2008 fleet (the baseline MY used for the 2017-2025 final rule). EPA also examined the
coefficients of drag reported in the Control-Tec data to determine if any vehicle categories are
experiencing difficulties in progressing toward the assumed aerodynamic improvements.

   The 2015 NAS report (p. 6-3, and Finding 6.1, p. 6-51) also examined the agencies'
assumptions for feasibility, cost, and effectiveness of Aerol and Aero2,  and concluded that the
assumptions appear to be reasonable for the 2020-2025 time frame (National Research Council,
2015).  The additional analyses outlined above further informs this conclusion.  Also, the
agencies considered redefining the specific technologies assumed for each level to better align
with what has been learned about actual fleet implementation since the 2012 FRM.

5.2.5.3.1      Industry Developments

   Since the 2012 FRM, the industry is seeing high levels of implementation of many passive
aerodynamic technologies. In addition, active aerodynamic technologies are seeing increasing
implementation, primarily in the form of active grille shutters, which are now offered by a
number of manufacturers.  Although relatively low penetration of other active technologies (such
as active ride height and wheel shutters) has occurred, this may be the result of a natural focus on
the most cost effective technologies in the early years of the program. These active technologies
will remain available for implementation in the future as other aerodynamic technologies begin
to reach maximum penetration.

   In January 2015, EPA staff attended the 2015 North American International Auto Show
(NAIAS) in order to gather information about  the state of implementation of various
aerodynamic technologies in the vehicles represented at the show. A total of 76 vehicles that
appeared to employ aerodynamic devices were viewed, across more than a dozen manufacturers.
A memorandum408 describing this informal survey is available in EPA Docket EPA-HQ-OAR-
2015-0827. Although the sample was casually collected and therefore was not random, the
information gathered informs our understanding of industry activity in application of
aerodynamic technology to production vehicles.  Table 5.12 shows  a breakdown of the
aerodynamic devices and technologies that were observed in these vehicles:
                                            5-145

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
       Table 5.12 Aerodynamic Technologies Observed in Vehicles Investigated at the 2015 NAIAS
Technology
Active Grill Shutters
Underbody Panels
Wheel Dams

front (full)
front
(partial)
middle or
side
rear
Front
Rear
Front Bumper Air Dam
Total vehicles inspected
Number of
vehicles
equipped
14
28
22
27
2
56
59
18
Percentage
equipped
18%
37%
29%
36%
3%
74%
78%
24%
76
   Based on this assessment, it is clear that manufacturers are choosing to implement passive and
active aerodynamic devices as permitted by the various levels of vehicle redesign or model
refresh represented in the displayed vehicles. Because many of the vehicles displayed at the
show are not completely new designs, the bulk of these aerodynamic improvements were likely
added in a non-optimized fashion; that is, added to an existing design rather than fully integrated
into a new vehicle design.  As a result, it is likely that opportunity for better-optimized
application of both passive and active aerodynamic technologies will  continue to exist as these
vehicles gradually enter redesign phases and entirely new designs are introduced.

   One example of the potential for optimized application of aerodynamic technologies can be
seen in the redesigned MY2015 Nissan Murano.  The exterior of this vehicle was completely
redesigned from its MY2003-2014 generation with the goal of minimizing aerodynamic drag by
combining passive aerodynamic devices with an optimized vehicle shape. 409 The primary
passive devices employed include optimization of the rear end shape to reduce rear end drag, and
addition of a large front spoiler to reduce underbody air flow and redirect it toward the roof of
the vehicle, thus augmenting the rear end drag improvements. Other passive improvements
include plastic fillet moldings at the wheel arches, raising of the rear edge of the hood, shaping
of the windshield molding and front pillars, engine under-cover and floor cover, and air
deflectors at the rear wheel wells. An active lower grille  shutter also redirects air over the body
when closed.  Together, these measures give the 2015 model a drag coefficient of 0.31,
representing a 16 to 17 percent improvement over the 0.37 Cd of the previous model.

   Another example of aerodynamics improvement can be found in the redesigned 2015 Acura
TLX Sedan. According to a 2015 presentation by Acura,410 this vehicle was redesigned with the
help of computational fluid dynamics (CFD) as well as wind tunnel and real-world coastdown
testing to achieve a 15 percent lower  CdA compared to the 2012 model year Acura TL. The
frontal area was described as having been reduced by  1.5 percent, suggesting that Cd alone was
improved by about 13.7 percent to achieve this result. Some of the methods used included
eliminating welds from the forward and rearward edges of the wheel arches by use of a roller
hem wheel arch design in place of spot welds, and smoothing transitions between body panels in
this area. These results were said to be achieved with no  compromises in interior space or crash
safety by Acura.
                                             5-146

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Another example of the aerodynamic improvements available in a full redesign is seen in the
all-new Ford F-150 pickup truck. An article in Motor Trend411 highlighted seven distinct tactics
by which drag was reduced, including: air ducts added under the headlamps to reduce wheel-
generated air wake; trim pieces strategically placed to avoid trapping air; box geometry modified
for better airflow without reducing the cargo volume; adding spoiler features to the tailgate;
angling of rear and front corners; and a flush mounted windshield. The 2015 model is touted as
being slightly larger than the previous model, indicating that the benefit  of these improvements
was achieved without loss of cargo space.

5.2.5.3.2     Joint Test Program with Transport Canada

   In 2013 a Joint Aerodynamics Assessment Program was initiated between Transport Canada
(TC), Environment and Climate Change Canada (ECCC), National Research Council (NRC) of
Canada, and EPA.405 The participating organizations and their respective programs share mutual
interests in the primary goals of the program, which are: (a) to quantify the aerodynamic  drag
impacts of various OEM aerodynamic technologies, and (b) to explore the improvement
potential of these technologies by expanding the capability and/or improving the design of
current state-of-the-art aerodynamic treatments.

   This program provides an important contribution to the agencies' technical assessment by
offering an opportunity to further validate the feasibility and effectiveness estimates for the
passive and active aerodynamic technologies assumed for Aerol and Aero2.

   The program also provides an opportunity to further validate off-cycle credits that were
assigned to active aerodynamics in the 2012 FRM. Two active aerodynamic technologies were
identified for pre-defined credit availability of specified amount: Active  Grille Shutters and
Active Ride Height. 86.1869-12 (b)(l)(iv). The default value for these credits offered were
determined in large part by analysis using an early version of the EPA ALPHA model to
simulate aerodynamic improvements for varying Cd inputs.  A key assumption in development of
these credits was that active technologies only affect the coefficient of drag, which is assumed to
be constant over the speed range of the test.  Further validation of this assumption, and of the list
of creditable active technologies assumed to be available in production vehicles during the time
frame of the rule, would strengthen the basis of the program. A total of four project phases
consisting of twenty-five test vehicles in all EPA vehicle classes was undertaken by the project
partners.406

   Active technologies evaluated by this program  include: active grille shutters (opened,  closed,
intermediate positions, speed effects, yaw effects,  leakage effects); a detailed sealing study (i.e.
grille shutter sealing; external grille shutter concept);  and  an active ride height concept (i.e.
manual ride height adjustment on vehicles not necessarily equipped to do so from factory).
Passive technologies include: Air dams (front bumper and wheels); active front bumper air dams
(concept/prototype); underbody smoothing panels (both OEM and idealized prototypes);  larger-
than-baseline wheel/tire packages; wheel covers (i.e. solid hubcaps); and miscellaneous
improvements (including front license plates, decorative grille features and smoothing, tailgates
(opened/closed/removed), and tonneau covers).  Significantly, NRC facilities include a 9 meter x
9 meter rolling road/moving floor wind tunnel that allows testing of full  scale vehicles for
accurate comparison of aerodynamic performance with and without active technologies.  Listed
technologies were not evaluated on every vehicle due to stock configuration, timing and funding.
                                             5-147

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   One valuable outcome of this testing was further validation of the default credit menu values
established in the 2012 FRM for active aerodynamic technologies under the off-cycle credit
program. Phase 1 of the Joint Program evaluated the aerodynamic performance of eleven (11)
vehicles (3  small cars, 5 midsize cars, 2 sport utility vehicles and 1 pickup truck). The
conclusions of the Phase 1 study indicated that the active aerodynamic technologies studied are
within the range of the default menu credit values anticipated in the 2017-2025 GHG rule TSD136
for active aerodynamic off-cycle credits.

   The Phase 1  study also concluded that the benefit of active grille shutters is constant across
the operating speed range, confirming one key assumption in the FRM analysis. In addition, it
concluded that passive technologies may each improve the aerodynamics of future vehicles by 1
to 7 percent depending on the passive technology employed and overall vehicle design. This
conclusion  was  based on individual component installation,  and does not account for synergistic
component effects, nor the effect of integrating passive technologies into an overall vehicle
redesign.

   Depending on stock vehicle equipment, sometimes  it was necessary to fabricate prototype
components to make an A to B comparison possible. Prototype components were constructed by
study partners Roechling Automotive and Magna International, both of which are Tier 1
suppliers of various aerodynamic technologies to the industry.

   Effectiveness values identified in Phase 1 of the Joint Program are shown in Table 5.13.
     Table 5.13 Aerodynamic Technology Effectiveness from Phase 1 of Joint Aerodynamics Program
Aero Feature (A-B Testing)
Fixed Air Dam-Bumper
Active Air Dam - Bumper
(Conceptual)
Fixed Air Dam-Wheels
Underbody Panels
Increased Tire Size
Wheel Covers
Front License Plates
Decorative Grille Optimization
Pick-up Tailgates Open
Removed
Pick-up Tonneau Cover
Aero Drag Reduction (%)
1 - 6%
4 - 9% (fixed air dam + 3%)
1% (front)/4.5% (front &
rear)
1-7% (stock OEM)
-2.0 - 3.2%
1.5 - 3%
+/- 0.3%
1.6%
-5.2%
-7.5%
3.7%
Comments
OEM stock components
Fixed, prototype parts w/ lowest
deployment height used

Addtn'l 0.5%-4% w/full body panels. Dodge
Ram prototype: 8%
17"/18" stock OEM rims vs. 22" optional
OEM rims
Solid wheel covers only; brake cooling
affects not considered
Negligible impact
Smoothing of grille features; function vs.
styling trade-offs

Open tailgate + 2. 3%

   Phase 2 of the Joint Program412 investigated similar technologies using the same methodology
of Phase 1. Vehicles studied in Phase 2 included nine vehicles including one small car (2014
                                             5-148

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
Chevy Spark BEV), one midsize car (2014 Chevy Impala mild hybrid), one large car (2014 Ford
Taurus SEL), one minivan (2014 Honda Odyssey), and five SUV/crossovers (2014 Subaru
Crosstrek Hybrid, 2014 Ford Edge SE EcoBoost, 2014 BMW X5, 2015 Nissan Pathfinder, and
2015 Chevy Tahoe LS).  Active technologies studied included: active grille shutters (including
yaw sweep) and active ride height (stock and conceptual). Passive technologies included:
underbody panels and air dams, and optional wheel packages.  Other technical assessments
included turbulent flow impacts and yaw sweep impact.  To take into account the fact that
vehicles are generally traveling in a windy environment from potentially all wind azimuth
angles, the wind averaged drag area was calculated for all cases where a yaw sweep was carried
out.

   Results of the Phase 1 and Phase 2 studies further support the conclusion that the Aerol and
Aero2 goals appear to be attainable, with many individual technologies that have not yet been
implemented on a majority of light-duty vehicles showing capability for significant
improvements in drag area.

   Phase 3 involved the testing of 4 vehicles: one sedan (2014 Nissan Versa Note Plus), one
minivan (2015 Toyota Sienna), and two sport utility vehicles (2014 Jeep Cherokee, 2015  Nissan
Murano)407. Phase 4 involved the retesting of previous vehicles with a focus on turbulent flow
including a small car (2014 Chevrolet Spark) and a pick-up truck (2015 Ford F-150)HH.

   One significant outcome of the study was the identification of several high-impact areas for
drag reduction. For example, the study found that lowering the ride height while pitching the
vehicle nose down could provide significant drag reduction. Also, it was shown that certain
combinations of technologies (such as active grille shutters with  air dams) often acted with
positive synergy (i.e. more than additive) to result in greater reductions in overall drag than the
individual technologies alone would suggest.

   It should be noted that the Phase 1 and Phase 2 studies found that some technologies could
potentially increase drag area if poorly applied, and that some individual technologies did not
appear to be fully additive when combined with certain others. For example, presence of active
air dams was seen in some cases to reduce the effectiveness of adding underbody coverings.
Further, combination of active air dams or underbody coverings with active ride height tended to
reduce the effectiveness of active ride height. This latter result corroborates with information
related to EPA in an OEM meeting that suggested that vehicles that already have underbody
coverings are not as highly responsive to adjustments in ride height.  On the other hand,
combining certain aerodynamic technologies (for example, active grille shutters with air dams)
often demonstrated higher total drag reduction than individual additive measurements would
have suggested.

   All phases of the study found that lowering ride height while pitching the vehicle at highway
speeds (for example, 40mm in the front and 20mm in the rear) provided significant drag
reduction for all vehicles. The highest reduction was observed for the Large Car classification.
Additionally, underbody panels that are extended to cover the entire surface area underneath the
vehicle (full underbody cover) proved to be an efficient way to reduce drag.
  1 The Phase 4 report was not yet finalized at the time of Draft TAR publication.
                                             5-149

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   It was also found that yaw angle had a significant effect on measurement. Some technologies
that perform well at 0° wind angle were found to perform relatively poorly at different wind
angles (for example, at 8° to 10°, the differences were quite significant). It was also found that
some technologies that tend to work well for one class of vehicle may not perform well for
another vehicle class (for example, air dams in turbulent flow conditions were shown to perform
better on SUVs than on Large Cars.

   In an effort to better represent real-world aerodynamic performance of aerodynamic
technologies, the study also investigated the effect of turbulent flow conditions on aerodynamic
measurements. The study produced an extensive data set comparing steady smooth and turbulent
flow performance for most of the vehicle classes. The study found that both turbulent flow and
yaw angle can be important to understanding the effectiveness of aerodynamic technologies in
real-world use.

5.2.5.3.3     CARB Control-Tec  Study

   In 2013, the California Air Resources Board (CARB) issued a Request for Proposal413 to
solicit research on the potential for vehicle road load reduction technologies to reduce the CCh
emissions of future vehicles.  The work was proposed to support the mutual interests of the
California Advanced Clean Cars Program and the agencies' midterm review effort.  An
automotive research firm called Control-Tec LLC was contracted by CARB to perform this work
and the work was completed in March 2015414.

   The objectives of the research included: determining vehicle load reduction technologies
included in or applicable to the California light-duty fleet; identifying the extent to which these
technologies have been applied to this fleet; developing a "what-if' scenario by applying best-in-
class load reduction technologies to the future fleet; and conducting projections to determine the
potential GHG reductions if all future vehicles were to adopt the best-in-class technologies.
Because aerodynamic technology is one of the components of road load technology, the results
of this study are very relevant to evaluating the feasibility and effectiveness of aerodynamic
technologies assumed in the FRM.

   As described in Appendix A (CARB Analysis of Vehicle Load Reduction Potential For
Advanced Clean Cars), the study defined the best-in-class application of aerodynamic
technology in the MY2014 fleet as being represented by the 90th-percentile drag coefficient
observed in that fleet within a given vehicle class.  Depending on vehicle class, this represented
an 8 percent to 12 percent improvement in drag coefficient over the median vehicle in the class.
Applying this degree of improvement to all of the vehicles in each respective class resulted (by
simulation) in an improvement  of about 5 g/mi in CCh emissions for the fleet overall, or about 2
percent, relative to a 2014 baseline value of 263 g/mi. It should be noted that the study was by
its nature limited only to consideration of aerodynamic technologies that existed in the MY2014
fleet, and therefore did not consider any more advanced examples of drag reduction technology
that may now be present in MY2015  or 2016 vehicles, nor any further improvements that may be
achieved by 2022-2025.

5.2.5.3.4     EPA Study of Certification Data

   The CARB/Control-Tec project created additional opportunities for EPA to study
aerodynamic technology implementation since the FRM. Control-Tec had based its analysis
                                             5-150

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
upon a large database of performance attributes of about 1,350 MY2014 vehicles, including
aerodynamic attributes such as drag coefficients and frontal area.  This database included CdA
values from various sources such as publicly available information and manufacturer reports.
Control-Tec had estimated values that were unavailable from these sources by a proprietary
methodology that estimated CdA values by mathematical analysis of coastdown coefficients
found in manufacturer certification data. This resulted in an unusually comprehensive and
inclusive picture of aerodynamic performance characteristics of MY2014 light-duty vehicles.

   A similar methodology might be used to help track adoption of technologies over time by
making it possible to generate fleet-wide estimates of CdA for any model year using
manufacturer certification data as a basis. This would provide a means to estimate the degree of
aerodynamic improvement that has been implemented since the 2008 model year baseline, by
using such a methodology to generate a database of fleet aerodynamic performance for MY2008
and comparing it to that of MY2014.

   While the Control-Tec methodology for estimating drag characteristics from test data is
proprietary, an understanding of the basic physics principles involved allowed EPA to study the
possibility of developing a similar methodology for estimating drag performance from
coastdown performance data contained in certification records. Figure 5.46 shows a frequency
distribution of CdA values for MY2008 and MY2014 derived from a preliminary exploratory
analysis. While  some improvement in drag performance appears to have occurred, the overall
magnitude of change is quite small; particularly noting that estimated CdA has increased from
0.942 in 2008 to 0.996 in 2014.
                     Cd*A Frequency Distribution for Model Years 2008 versus 2014
2003
Total Records
Average CaA/m'2!
Mm
Max
Standard Deviation
13421
0.9'2
0.218
4.1S6
0.345
2014
Total Records
Average CdA,'™-*2!
Mm
Max
Standard Devyt'c-
13493
0.996
0.351
4.371
0.623
    S 3000
         Q.2C] 030 Q40 Q50 ft BO 0.70 D.80 ft 90 100 1.10  1.20 130 1.41  150 1.60 1.70  1BD 1.90 ZOO 2.10 120 Z30 2.« 150
    Figure 5.46 Distribution of Estimated CdA for MYs 2008 and 2014 Derived from Certification Data

   Since the Control-Tec database relies largely on manufacturer-reported or publicly available
information as well as analytically derived figures, EPA sought paths to further validate the
proprietary methodology behind the figures.  EPA recognized that the Joint Aerodynamic
Assessment Program (previously described) could provide a sample of accurately measured CdA
values that could be used to validate the Control-Tec methodology in this application. Although
this analysis was not completed in time for publication of this Draft TAR analysis, results may
become available to further inform the agencies' analysis.
                                              5-151

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   EPA also plans to more closely examine the Control-Tec database to look at various vehicle
categories and examine the span between the best and worst aerodynamically performing
vehicles, using CdA values as a metric. The size of the span as it exists in a given category of
current MY vehicles might be suggestive of the remaining potential for aerodynamic
improvement within that category. Although this analysis was not completed in time for the
publication of this Draft TAR analysis, results may become available to further inform the
analysis in the future.

   In general, it appears that manufacturers are aggressively pursuing improvements to
aerodynamic drag across a wide range of vehicles, particularly for vehicles where the efficiency
improvement is highly cost effective.  For example, in 2015 Toyota announced155 that the 2016
Prius would have a drag coefficient of 0.24, which not very long ago was considered to be an
extremely low value for a production vehicle. This value is expected to be eclipsed by vehicles
such as the Tesla Model 3, which has been described as targeting a drag coefficient of about
0.21.  Examples such as these further support the attainability of the aerodynamic technology
cases Aero 1 and Aero 2.

5.2.5.3.5      Conclusions

   In summary, the agencies  evaluated the feasibility, cost and effectiveness of the two levels of
aerodynamic technology (Aerol and Aero2) by the efforts described above. The agencies'
analysis of industry developments shows that manufacturers are already implementing many
passive and active aerodynamic technologies in MY2015 vehicles, with significant opportunity
remaining to further apply these technologies in a more optimized fashion as vehicles enter
redesign cycles in the future.  The findings of the Joint Aerodynamics Assessment Program and
the Control-Tec analysis also lend support to the feasibility of the 10 percent and 20 percent
effectiveness levels assumed  for Aerol and Aero2.  The NAS report likewise generally
supported the assumptions for Aerol and Aero2 as being applicable to the 2020-2025 time
frame.

   Some tradeoffs and interactions among specific aerodynamic technologies were identified that
suggest there could be value in refining the specific combinations of technologies that are
assumed to make up the Aerol  and Aero2 packages.

   For the cost and effectiveness assumptions the agencies  are adopting for the GHG Assessment
and CAFE Assessment for this Draft TAR analysis, see Sections 5.3 and 5.4.

5.2.6  Tires: State of Technology

5.2.6.1 Background

   Tire rolling resistance is a road load force that arises primarily from  the energy dissipated by
elastic deformation of the tires as they roll. Deformation, and  hence rolling resistance, for a
given tire design is largely a function of vehicle weight and is fairly constant across the normal
range of vehicle speeds. Rolling resistance therefore carries an ever-present and often quite
significant effect on fuel economy and CCh emissions.

   Tire design characteristics (for example, materials, construction, and tread design) have a
strong influence on the amount and type of deformation and the energy it dissipates. Designers
can select these characteristics to minimize rolling resistance.  However, these characteristics
                                             5-152

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
may also influence other performance attributes such as durability, wet and dry traction,
handling, and ride comfort.

   Low rolling resistance tires are increasingly specified by OEMs in new vehicles, and are also
increasingly available from aftermarket tire vendors.  They commonly include attributes such as:
higher inflation pressure, material changes, tire construction optimized for lower hysteresis,
geometry changes (e.g., reduced aspect ratios), and reduced sidewall and tread deflection.  These
changes are commonly accompanied by additional changes to vehicle suspension tuning and/or
suspension design to mitigate any potential impact on other performance attributes of the vehicle.

5.2.6.2 Tire Technologies in the FRM

   In the 2012 FRM, the agencies considered two levels of low rolling resistance technology,
known as LRRT1 and LRRT2. The first level, LRRT1, was defined as a 10 percent reduction in
rolling resistance from a base tire, made possible by methods such as increased tire diameter and
sidewall stiffness and reduced aspect ratios (coupled with reduction in rotational inertia).  The
second level, LRRT2, was defined as a 20 percent reduction in rolling resistance from a base tire.
LRRT2 was associated with more advanced approaches such as use of advanced materials and
complete tire redesign.

   Based on the 2011 Ricardo study, the agencies estimated the effectiveness of LRRT1 as 1.9
percent and the effectiveness of LRRT2 as 3.9 percent for all vehicle classes.  This represents a
2.0 percent incremental effectiveness increase from LRRT1 to LRRT2.

   In the 2012 FRM, NHTSA assumed that the increased traction requirements for braking and
handling for performance vehicles could not be fully met with the LRRT2 designs in the MYs
2017-2025 timeframe. For this reason the CAFE model  did not apply LRRT2 to performance
vehicle classifications. However, the agency did assume that traction requirements for LRRT1
could be met in this timeframe and thus allowed LRRT1 to be applied to performance vehicle
classifications in the MYs 2017-2025 timeframe.

   In the 2012 FRM, the agencies estimated the incremental DMC for LRRT1 at an increase of
$5 (2007$) per vehicle, adjusted to 2010 dollars11. This included costs associated with five tires
per vehicle: four primary and one spare tire. There was no learning applied to this technology
due to the commodity based nature of this technology. The agencies considered LRRT1 to be
fully learned out or "off the learning curve (i.e., the DMC does not change year-over-year) and
have applied a low complexity ICM of 1.24 through 2018, and then 1.19 thereafter, due to the
fact that this technology is already well established in the marketplace.

   Prior to the FRM, EPA, NHTSA, and CARB met with a number of the largest tire suppliers in
the United States to analyze the feasibility and cost for LRRT2. The suppliers were generally
optimistic about the ability to reduce tire rolling resistance in the future without the need to
sacrifice traction (safety) or tread life (durability). Suppliers all generally stated that rolling
resistance levels could be reduced by 20 percent relative to then-current tires by MY2017.  As
11 We show dollar values to the nearest dollar. However, dollars and cents are carried through each agency's
  respective analysis. Thus, while the cost for lower rolling resistance tires in the 2012-2016 final rule was shown
  as $5, the specific value used in that rule was $5.15 (2007$) and is now $5.40 (2010$). We show $5 for
  presentation simplicity.
                                             5-153

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
such, the agencies agreed, based on these discussions, to consider LRRT2 as initially available
for purposes of the FRM analysis in MY2017, but not widespread in the marketplace until MYs
2022-2023. In alignment with introduction of new technology, the agencies limited the phase-in
schedule to 15 percent of a manufacturer's fleet starting in 2017, allowing complete application
(100 percent of a manufacturer's fleet) by 2023.

   EPA projected fleet penetration of low rolling resistance technology based on penetration of
LRRT2.  Because LRRT1 and LRRT2 technology are both defined as incremental to a baseline
vehicle, increased use of LRRT2 would displace use of LRRT1. LRRT2 technology was
projected to essentially replace LRRT1 technology by the later years of the rule. Penetration of
LRRT2 was projected to achieve 73 percent fleet penetration by 2021  and 97 percent by 2025.

   For this Draft TAR analysis, the agencies continue to believe that this schedule aligns with the
necessary efforts for production implementation, such as system and electronic system
calibration and verification.

   At the time of the 2012 FRM, LRRT2 technology did not yet exist  in the marketplace, making
cost estimation challenging without disclosing potentially confidential business information. To
develop a transparent cost estimate, the agencies relied on LRRT1 history, costs, market
implementation, and information provided by the 2010 NAS report. The agencies assumed
LRRT1 first entered the marketplace in the 1993 time frame with more widespread adoption
being achieved in recent years, yielding approximately  15 years to maturity and widespread
adoption. Then, using MY2017 as the starting point for market entry for LRRT2 and taking into
account the advances in industry knowledge and an assumed increase  in demand for
improvements in this technology, the agencies interpolated DMC for LRRT2 at $10 (2010$) per
tire, or $40  ($2010) per vehicle. This estimate was seen to be generally fairly consistent with
CBI suggestions by tire suppliers. The agencies did not include a cost for the spare tire because
we believe manufacturers are not likely to include a LRRT2 as a spare given the $10 DMC. In
some cases  and when possible pending any state-level requirements, manufacturers have
removed spare tires replacing them with  tire repair kits to reduce both cost and weight associated
with a spare tire.  The agencies continued to consider this estimated cost for LRRT2 to be
applicable in MY2021. Further, the agencies considered LRRT2 technology to be on the steep
portion of the learning curve where costs would be reduced quickly in a relatively short period of
time. The agencies applied a low complexity ICM of 1.24 through 2024, and then 1.19
thereafter. The ICM timing for LRRT2 was different from that for LRRT1 because LRRT2 was
not yet being implemented in the fleet.

   For the 2012 FRM,  the agencies also considered introducing a third level of rolling resistance
reduction, LRR3, defined as a 30 percent reduction in rolling resistance from the baseline, but
ultimately declined to do so.  See 77 FR  and the 2012 TSD, p. 3-210.

5.2.6.3 Developments since the FRM

   The 2015 NAS report (p. 6-35, and Finding 6.10, p. 6-53) examined the agencies' assumptions
for feasibility, cost, and effectiveness for the two levels of rolling resistance, LRRT1 and
LRRT2.  The report concluded that the feasibility and effectiveness projected by the agencies for
a 20 percent reduction in rolling resistance in the 2020-2025 time frame appears to be
reasonable.  With regard to costs, the Committee substantially agreed with the costs projected by
the agencies, while noting that the problem of maintaining tread wear  and traction requirements
                                            5-154

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
while reducing rolling resistance continues to present engineering challenges that could affect
tire costs.

   Since the FRM, the agencies have taken several additional steps to further validate the
feasibility, cost, and effectiveness assumptions of LRRT1 and LRRT2.

   We followed industry developments and trends in application of low rolling resistance
technologies to light-duty vehicles.  We did this by gathering input from stakeholders through
meetings with OEMs, suppliers and other interested  parties, and also by attending conferences
and trade shows and regularly monitoring the press and technical literature.

   EPA is coordinating with Transport Canada (TC) and Natural Resources Canada (NRCan) on
a study of the rolling resistance and traction characteristics of low-rolling resistance tires.  TC
and NRCan  originated this study in part to support the development of a Canada consumer
information  program for replacement tires.  The program will study the correlation between
rolling resistance performance and safety performance (traction) for winter and all-season tires.
As such, it promises to provide concrete input on any tradeoffs between rolling resistance  and
traction in current production tires, and so will inform the safety concerns noted by NHTSA and
the NAS report. A total of 50 randomly selected all-season tires and 5 all-weather tires will be
tested under this program.  The study is scheduled for completion by December 2016, with
testing to be completed earlier that year. Although the analysis was not complete in time for the
publication of this Draft TAR analysis, its findings will be incorporated into the agencies'
analysis as they become available.

5.2.6.3.1     Industry Developments

   Tires that achieve the level of improvement of LRRT1 are widely available today, and  since
the FRM appear to have continued to comprise a larger and larger portion of tire manufacturers'
product lines as the technology has continued to improve and mature. Improvements that  would
reach the level of LRRT2 have also seen significant  progress in the industry, with indications of
increased availability, improved traction and performance characteristics, and additional cost
information.

   Since the 2012 FRM and even before, the tire industry has become increasingly focused on
improving tire performance. Recent industry momentum in this direction was captured well in a
quote by Kurt Berger of Bridgestone, in a 2014 article in Automotive News.415 "A low-rolling-
resistance tire of 2010 would not be considered a low-rolling-resistance tire today. We've really
been pushed in a short time to reduce rolling resistance further." Several typical examples of
industry research and implementation efforts are outlined in a 2015 report by Auto World416.
One example of a specific product embodying lower rolling resistance technology is the Falken
Sincera SN832 Ecorun Tire, with a 22 percent improvement over its immediately previous
generation, while maintaining a 27 percent improvement in braking distance. According to a
Continental  spokesperson cited in the Auto World report, "... improvements of more than 20
percent from one generation to the next [are possible] by introducing rolling resistance optimized
tires. ... an additional 5 percent improvement generation-to-generation  is possible." According
to Indraneel  Bardhan, Managing Partner of EOS Intelligence, so-called "green tires" have
achieved a global market share of about 30 percent.
                                             5-155

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
   The Automotive News article cited above also discussed ongoing challenges for low rolling
resistance tires, including issues such as wet traction, tread wear, and the magnitude of real world
benefits in comparison to customer expectations. Customers were said to be relatively
indifferent about the fuel economy benefits of low rolling resistance tires, but the perception of
differences in handling performance between these tires and traditional tires appeared to be
stronger. Due to these perceptions, it was suggested that although original equipment fitments of
low rolling resistance tires have been increasing, consumers may tend to replace them with more
conventional tires after the original tires wear out, potentially reducing the lifetime impact of this
tire technology on fuel economy.

   Despite the typical perception that reducing rolling resistance sacrifices traction performance,
tire designers can exercise a variety of design options to preserve traction characteristics while
maintaining low rolling resistance. For example, as shown in Figure 5.47, preliminary results of
the Transport Canada/Natural Resources Canada study show that winter tires are available with a
wide variety of rolling resistance and wet grip characteristics, including tires with both low
rolling resistance and good wet grip. For instance one tire had a rolling resistance coefficient less
than 9.0, and a wet grip index greater than 1.1.

                   1.4
                   1.3

                   0.9 -
                   0.8
                     7.00        8.00       9.00       10.00
                                     Rolling Resistance Coefficient
ll.C
12.1
 Figure 5.47 Relationship between Wet Grip Index and Rolling Resistance for Winter Tires from Transport
                                    Canada/NRCan Study

   One example of the potential for careful design to maintain traction in a low rolling resistance
tire is seen in the Bridgestone "ologic" design, which appears on the BMW i3 electric vehicle.
This tire has a relatively large diameter coupled with a narrow width, reducing rolling resistance
by maintaining low deformation through a stiffer belt tension. The larger diameter and unique
construction increases the length of the contact patch, which serves to provide improved braking
performance and wet and dry traction. An advanced rubber compound and special tread design
also contributes.417 The relatively narrow design is also said to improve aerodynamic
performance.416 The trend toward larger diameter tires with narrower cross-sectional width is
also associated with lower tire noise levels, and have been described as one of the likely tire
design trends that will continue into the future, particularly for BEVs that value both energy
efficiency and quiet performance416.  As another example, the tire manufacturer Pirelli has
                                              5-156

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
projects focusing on development of new tire polymers through joint ventures with chemical
suppliers416.

   Research data presented at the 2014 U.S. DOE Merit Review strongly suggests that
significant rolling resistance improvements are accessible to much of the tire market.  A project
involving Cooper Tires, funded by the U.S. Department of Energy, targets a 30 percent reduction
in rolling resistance and a 20 percent reduction in tire weight, while maintaining traction
performance.418  By investigating new materials and methods for reducing rolling resistance in
ways that maintain wet traction and tread wear capabilities, this project has suggested that
potential improvements in rolling resistance of 10 to 20 percent are achievable by selection of
appropriate materials and construction, with examples of reduction in rolling resistance from a
prevailing 0.08 to 0.10 down to 0.064 to 0.08.

5.2.6.3.2      Control-Tec Analysis of Trends in Tire Technologies

   As discussed under Aerodynamics (Section 5.2.5.3.3) and also in Appendix A (CARB
Analysis of Vehicle Load Reduction Potential For Advanced Clean Cars), an analysis performed
by Control-Tec for the California Air Resources Board413'414 resulted in a large database of
estimated road load parameters for many current vehicles, including estimates of tire rolling
resistance. Many of these estimates were analytically derived from input data such as
dynamometer road load coefficients. To derive tire rolling resistance, factors representing
driveline drag and aerodynamic drag were subtracted from the total road load force, with the
remainder being  taken as representative of tire rolling resistance.

   As described in the CARB Analysis, the study defined the best-in-class application of rolling
resistance technology in the MY2014 fleet as being represented by the 75th-percentile rolling
resistance coefficient observed in that fleet within a given vehicle class. Depending on the tire
category, this represented an 11 percent to 14 percent improvement in rolling resistance over the
median vehicle in the class. Applying this degree of improvement to all of the vehicles in each
respective class resulted (by simulation) in an improvement of about 5 g/mi in CCh emissions for
the fleet overall,  or about 2 percent, relative to a 2014 baseline value of 263 g/mi. It should be
noted that the study was limited only to consideration of rolling resistance technologies
represented in the MY2014 fleet, and therefore did not consider more advanced technologies that
may now be present in MY2015 or 2016 vehicles, nor any further improvements that may be
achieved by 2022-2025.

   EPA plans to  more closely examine the Control-Tec database for its potential to characterize
the penetration of tire rolling resistance technologies in the 2014 fleet.  Although this analysis
was not completed in time for the publication of this Draft TAR analysis, any results that become
available may be used to further inform the agencies' analysis.

5.2.6.3.3      Canada Tire Testing Program

   EPA is coordinating with Transport Canada (TC) and Natural Resources Canada (NRCan) on
a tire testing program that will  provide a large amount of test data relating the rolling resistance
of tires to their wet and dry traction performance.  The tire testing program was initiated by
Transport Canada as part of a Canadian initiative to develop a tire consumer information
program to inform consumer selection of aftermarket replacement tires. EPA partnered with the
Canadian agencies due to mutual interests in supporting the midterm evaluation.
                                             5-157

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   A major goal of the testing program is to study the correlation between tire rolling resistance
and safety performance of winter tires, all-weather tires, and all-season tires. The program will
also examine various approaches to the characterization of rolling resistance of tires operating in
cold ambient temperatures, a consideration of particular interest to the Canadian market.

   To date, a random selection of 23 winter tires have been tested and a random selection of 50
all-season and 5 all-weather tire models have been acquired and are undergoing testing. The
previously presented plot of tire rolling resistance and traction performance (Figure 5.47) was
derived from preliminary data provided by this program.

   Although this testing project will not be completed in time for the June 2016 publication of
this Draft TAR analysis, a final report is expected to be completed by the end of 2016 and may
be available to further inform the agencies' analysis.

5.2.6.4 Conclusions

   In summary, the agencies have revisited the feasibility of the two levels of rolling resistance
reduction (LRRT1 and LRRT2) through the efforts described above. The 2015 NAS  report
generally supported the cost, effectiveness, and feasibility assumptions for LRRT1 and LRRT2
as being appropriate for the 2020-2025 time frame. The agencies' analysis of industry
developments shows that tire manufacturers are aggressively pursuing rolling resistance
technology capable of achieving a 10 percent and 20 percent reduction in rolling resistance,
while OEMs are increasingly specifying low rolling resistance tires in original  fitments  of their
products.  Although there is some evidence that consumers have associated low rolling resistance
technology with reductions in traction, the ability  of tire designers to exercise many design
parameters in pursuit of traction performance makes it unclear whether this will continue in the
future.

   For the cost and effectiveness assumptions the  agencies are adopting for the GHG  Assessment
and CAFE Assessment for this Draft TAR analysis, see Sections 5.3 and 5.4.

5.2.7   Mass Reduction: State of Technology

5.2.7.1 Overview of Mass Reduction Technologies

   Mass reduction remains a key technology that vehicle manufacturers are expected  to continue
to apply to meet the light-duty GHG standards.  The reduction of overall vehicle mass can be
accomplished through several different techniques. Techniques include CAE optimization of
designs, adoption of lighter weight materials, and part consolidation. The cost of reducing
vehicle mass is highly variable.  Design optimization, consolidation of components along with
adoption of secondary mass savings opportunities can result in some cost savings. Secondary
mass reduction is weight reduction opportunities that are available as the base vehicle becomes
lighter. A smaller engine block, transmission and brakes are examples of secondary mass
reduction technologies. Cost increases are often the result of changing from a high density,
lower cost material like steel, to a lower density, higher cost material such as certain advanced
high strength steels, aluminum, magnesium or composites. The cost for each mass reduction
solution depends on the approach and material used. In some cases, the cost savings can offset
the cost increases. Benefits from adopting mass reduction technologies, also include increased
performance such as improved vehicle dynamics and responsiveness.
                                             5-158

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   For this Draft TAR, which is reviewing technologies for the 2022-2025 standards, EPA
reevaluated all aspects of mass reduction including the methodologies described above, mass
reduction cost, the FRM conclusions and the amount of mass reduction in the baseline fleet. To
support this Draft TAR, EPA and NHTSA have also completed new work including research,
stakeholder meetings, supplier meetings, technical conferences and literature searches.  Public
information from these sources are contained in this section and are the basis of the development
of new mass reduction cost curves for technology package modeling.  Section 5.3 describes the
specific data and assumptions that were used for modeling mass reduction for this assessment
and includes the 2014 baseline fleet mass reduction estimates including mass allowances for
safety and footprint changes between the 2008 and 2014 vehicles, cost curve development and
application, and effectiveness.  Specific material (steel, aluminum, magnesium, plastic, glass
fiber and carbon fiber composites, glass) and application details addressing Feasibility, Cost,
Mass Reduction, Safety and Research, are included in Part B of the Appendices.

   The relationship between mass reduction and safety has also been an important consideration
and NHTSA performed an updated analysis for which a description and results can be found in
Chapter 8.

   Current industry trends in mass reduction are to adopt mass reduction technologies in various
degrees. From vehicles that have adopted large amounts of lower density materials in their body
in white (BIW), as with the MY2015 Ford F150 and MY2014 BMWi3, to vehicles that have
adopted smaller changes in vehicle design such as an aluminum hood or a steel  clamshell  control
arm in the suspension such as the MY2014 Silverado 1500. The EPA 2015 Trends report
illustrates, in Figure 5.48, how in overall sales weighted basis, vehicles have not yet achieved a
notable decrease in curb weight or have continued the trend of using mass reduction to offset
increased vehicle content or larger footprint as the mass difference has remained constant over
the past 10 years. The detail within the report notes 2014 results show a 0.5 percent mass
increase for cars and  0.7 percent mass decrease for trucks, each on a sales weighted basis.

                   Change in Adjusted Fuel Economy, Weight, and Horsepower for MY 1975-2015
                       100%-
                           1975  19BO  19B5  1990  1995 2000  2005  2010 2015

                                          Model Year
       Figure 5.48 Change in Adjusted Fuel Economy, Weight and Horsepower for MY1975-20154
                                             5-159

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   One reason for the current trend of curb weight changes may be the desire to make significant
mass reducing design changes during major vehicle redesigns, hence slowing down the release of
lightweight vehicles.  Other reasons may include the idea that the standards for MYs 2014/2015
don't require high levels of MR, different manufacturers have different compliance strategies, or
some vehicles are prioritized for mass reduction for the ancillary benefits that mass reduction
provides. Recent announcements, as listed in Table 5.14, indicate that the adoption of mass
reduction technologies, and resultant lower curb weights, will continue into the future as vehicle
design cycles are revisited and material costs are lowered. One  example is the announcement of
the MY2017 Acadia by GMC in which it was stated as having a 7001b mass reduction through
adoption of high strength steels, smaller engine offering and smaller footprint.420 The January
2016 announcement of the  2017 Chrysler Pacifica also touted 2501bs of mass reduction through
"extensive use of advanced, hot-stamped/high-strength steels, application of structural adhesives
where necessary and an intense focus on mass optimization." Magnesium is also used in the
instrument panel and the inner structure of the Pacifica's liftgate, the rest of which is
aluminum.421

   To understand the general  trend in the use of lightweight materials we have included Figure
5.49 which shows a comparison of metal material adoption from 2012-2025 included in the 2014
Executive Summary for the Ducker Study.422 The study notes that there was a slight increase in
the use of light-weight materials for BIW and closures between  2012 and 2015. The use of
AHSS/UHSS grew from 15 percent to 20 percent of the vehicle body  and closure parts.
Aluminum sheet also grew from 1 percent to 4 percent and aluminum extrusions made it onto the
pie chart in 2015. Overall,  the analyses expects that steel will still remain the dominant material
in BIW and closures. According to IHS increases in plastics are expected to grow to be
350kg/average car in 2020  which is up  from 200 in 2014, as shown in Figure 5.50.  Auto
manufacturing use of carbon fiber is expected to increase from 3,400 metric tons in 2013 to 9800
metric tons in 2030. According to Ducker Worldwide the use of magnesium is expected to
increase through 2025  as over the next  10 years magnesium castings are expected to grow
significantly. "Growth is highlighted within "large tonnage" parts like closure inners, IP
structures etc. and other body/structural parts."
                                             5-160

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
           Executive Summary
              The material mix for body and closure parts will change dramatically over the next ten years.
              On a weight basis, aluminum will grow to 19% of the weight for body and closure parts by 2025.
                             2012
                                   • Mild and HSLA
                                   • AHSS/UHSS
                                   Aluminum Sheet

                                   3 Aluminum VD Castings

                                   1023 Ibs.
                            2020
                                 • Mild and HSLA
                                 • AHSS/UHSS
                                 -Aluminum Sheet
                                 • Aluminum Extrusions
                                 • Aluminum VO Castings

                                   941 Ibs.
                                                                  2015
                                                                 2025
• Mild and HSLA
• AHSS/UHSS
E Aluminum Sheet
• Aluminum Extrusions
• Aluminum VD Castings

 993 Ibs.
• Mild and HSLA
• AHSS/UHSS
 Aluminum Sheet
• Aluminum Extrusions
« Aluminum VD Castings
                                                                      888 Ibs.
Figure 5.49 Estimated Vehicle Material Change over Time 2012-2025 - Ducker Worldwide422
                      Automotive Market Summary
12005 2010 2015 2020 2024 Forecasted |
%CAGR 1





Vehicle Unit
Deliveries
Total Vehicle
Wt (MT)
Total FGRP
Structures (MT)
Total CFRP
Structures (MT)
Total CF
Demand (MT)
66.SM
30.2M
79,212
3,921
3,666
77.9M
35.4M
102,371
3,771
3,526
93.0M
37.2M
131,310
13,060
10,056
100.7M
41.1M
167,588
37,085
23,456
109.0M
44.5M
203,704
47,011
47,011
1.8096
1.80%
4.4996
16.6796
13.7396





         Figure 5.50 Forecast of Automotive Market Consumption of Composites
                                                                                 423
                                               5-161

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
        2010      2011      2012      2013      2014     2015 (e)    2020 (f)    2025 (f)

                                Low   Pounds per Vehicle  • Hi
                                                                           ,424
          Figure 5.51 Magnesium Growth Expectations through 2025 (Ducker Worldwide)

   While a significant amount of work and resources have already been devoted to developing
and implementing mass reduction technologies by OEM's and suppliers, the research for new
materials and processes continues and some of the research is included in the Appendices'
material summaries. The agencies expect that innovative mass reduction solutions will continue
to be developed and adopted through 2025 and that mass reduction will be less costly than it is
today.  Advancements expected include the development of lower cost high strength steel alloys
for body structures (3rd generation steels), lower cost and higher quality product (for Class A
surfaces) from the aluminum Micromill sheet manufacturing processes and advancements in
engineered plastics and composites for structural applications. Additional anticipated
developments in design include further development and use of CAE design tools to characterize
new material properties and behaviors which will result in material use advances including
optimized load pathway analyses in BIW geometries or consolidation of multi-part components
resulting in the achievement of mass reduction in the most cost effective way. The agencies will
continue to follow the progress of lightweight material adoption.

5.2.7.2 Developments since the 2012 FRM

   Since the publication of the FRM, the agencies have been able to gather additional
information on technological advancements and application of mass reduction technologies
through a variety of resources including conferences, public reports, material association
meetings, academic research work, online articles and CBI discussions and materials from
manufacturers and suppliers. A snapshot of publicly available information on lightweight
materials is included in the Appendices.  The agencies also generated two new holistic
lightweighting studies for mass reduction and cost data on light duty pickup trucks (MY2011 and
MY2014) and updated existing passenger car (EPA Midsize CUV and NHTSA Passenger car)
holistic lightweighting studies completed in 2012. The light duty truck holistic reports join the
projects currently described in the FRM on a midsize CUV, one funded by EPA and one by
ARE, and a passenger car, funded by NHTSA. The Aluminum Association also conducted
several projects including a project with EDAG, Inc. to evaluate the EPA Midsize CUV high
strength steel BIW CAE model with aluminum material replacement.
                                             5-162

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   DOE also joined forces with Ford/Magna to develop a multi-material lightweight vehicle,
through vehicle build and durability tests. In addition to vehicle lightweighting, research
projects were performed on the mass adds due to safety requirements by IIHS small overlap test
(2012) for their Top Rated Safety Pick. NHTSA funded a CAE passenger car evaluation and
Transport Canada funded a CAE light duty truck study evaluation which included a crash test of
the baseline vehicle. With respect to mass reduction efficiency, the Aluminum  Association
funded a  study on the impact of mass reduction on fuel economy for various vehicles with
Ricardo,  Inc. on which the 2015 NAS report comments were based.  The EPA and NHTSA
(through  ANL) also re-evaluated the effectiveness of mass reduction on CCh  and fuel
consumption reductions for several vehicle classes, including standard car and light duty truck.
The studies on efficiency will be addressed in Section 5.3.

   The following section provides a description of the multi material approach to lightweighting
being used by OEM's and presents some examples of current vehicle designs that have adopted
notable mass reduction which resulted in curb weight changes. Further sections present an
overview of the various holistic mass reduction and cost studies that have been  completed since
the FRM. The studies provide technology, primary and secondary mass reduction, and cost
information in order to create cost curves for application of mass reduction technology for a
passenger car and light duty pickup truck.

5.2.7.3 Market Vehicle Implementation of Mass Reduction

   Trends of slightly decreased curb weight in the new vehicle fleet are starting to be seen in the
data.  The 2014 EPA Trends report in Figure 5.48, illustrates that the overall sales weighted
vehicle weight has remained  steady over the past 10 years.  The information in  Figure 5.52
illustrates that in 2008, the sales weighted vehicle weight was 4085 Ib at 48.9 sq ft while the
2014 sales weighted vehicle weight was 4060 at 49.9 sq ft which is a decrease of 251bs and an
increase of one square foot. At the same time mass increases from additional safety regulations
and are accounted for in the 2014 weights. In order to achieve the results of increased size and
decreased weight, lightweight technologies/approaches have had to be incorporated into vehicle
designs.
                                             5-163

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
        Table 2.1
        Adjusted COi Emissions, Adjusted Fuel Economy, and Key Parameters by Model Year1
                             Adj   Adj Fuel
                   Production   CO:   Economy  Weight
         Model Year    (000)    (g/mi)   |MPG)     (Ib)
HP
                           Alternative
                           Fuel Vehicle
Footprint     Car      Truck     Share of
 (sq ft)   Production  Production  Production
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
X1S (prelim!
15,892
IS, 104
15,276
13,898
9,316
11, HE
12,018
13.448
15.19S
15.512

447
to
-131
424
397
£M
397
375
366
366
160
19.9
20.1
20.fi
21.0
22.4
21.G
21.4
23.7
24.3
24.3
21-7
•'.05-:
4067
4093
40BS
3914
4001
4126
3979
4003
4060
4076
an
213
217
219
208
214
230
222
226
230
233

-
-
48.9
48.1
48.5
49.5
48.8
49.1
49.7
49.9
55.6%
57.9%
SS.9%
S9.3%
67.0%
62.8%
57.8%
64.4%
64.1%
SS.3%
59.6*
44.4%
42.1%
41.1%
40.7%
33.0%
37.2%
42.2%
35.6%
35.9%
40.7%
40. tX
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.1%
0.4%
0.7%
0.7%
1.1%

        1 Adjusted CO2 and fuel economy values reflect real world performance and are not comparable to automaker standards compliance
        levels. Adjusted CO2 values are, on average, about 2S% higher than the unadjusted, laboratory C02 values that form the starting point for
        GHG standards compliance, and adjusted fuel economy values are about 20% lower, on average, than unadjusted fuel economy values.
                 A	

                   Figure 5.52  Footprint (square feet) Change and Weight 2007-2014

   Table  5.14 lists a number of vehicle lightweighting efforts that have been announced over the
past few years.  Some vehicles adopted high strength  steel solutions, up to 2 GPa tensile strength
steels, in  their BIW such as in the Audi Q7, Acura TLX, Nissan Murano and Cadillac CIS
redesigns. The MY2015 F150 and the MY2014 Range Rover by Land Rover have both adopted
a number of lightweighting components including aluminum body and cabin structure, aluminum
closures,  etc.
                                                   5-164

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Table 5.14 Examples of Mass Reduction in Selected Recent Redesigns (Compared to MY2008 Design)J
Vehicle Make
Acura MDX
Audi Q7
Land Rover Range Rover
Silverado 1500 Crew Cab
4x4
Ford F150
2.7L EcoBoost, 4x2
Supercrew
Nissan Murano
Cadillac CIS
Honda Pilot
Chevy Cruze425
Chevy Malibu426
GMCAcadia
Chrysler Pacifica
Cadillac XT5427
2008 Model Year
curb Weight (kg)
2070
2320
2400
2422
2446
1500
1833
4367
1425
1552
2120
2110
1893
Model Year
2014
2014
2014
2014
2015
2015
2015
2016
2016
2016
2017
2017
2017
Change in Vehicle
Curb Weight (kg)
238
325
336
86
318
30
110
131
114
136
318
114
82
% Change
11.5%
14%
14%
3.6%
13%
2%
6%
3%
8%
9.2%
15%
5.4%
4.5%
% Footprint
Change
+0.5%
0
+5.2%
n/a
n/a
n/a
+1.6%
+6.1%
n/a
+0.3%
-7.8%
+8.2%
+2.7%
   The press release by Audi428 represents the engineering perspective that is needed to achieve
notable mass reduction: "Although it (Q7) is shorter and narrower than its predecessor, the cabin
is longer and offers more head room.  20 years of experience with lightweight construction flow
into the new Audi Q7. Equipped with the 3.0 TDI engine, the new Audi Q7 tips the scales at just
1,995 kilograms (4,398 lb.), which is 325 kilograms (716.5 Ib.) less weight.... The Q7 with the
3.0 TFSI engine is even lighter, weighing just 1,970 kilograms (4,343.1 lb.).  Lightweight
construction has been applied in all areas, from the electrical system to the luggage compartment
floor. The key is the body structure, where a new multi-material design reduces its weight by 71
kilograms (156.5 lb.)... .Ultra-high-strength parts made of hot-shaped steel form the backbone of
the occupant cell.  Aluminum castings, extruded sections and panels are used in the front and rear
ends as well as the superstructure. They account for 41 percent of the body structure. Other parts
made entirely of aluminum are the doors, which  shave 24 kilograms (52.9 lb.) of weight, the
front fenders, the engine hood and the rear hatch. Audi uses new manufacturing methods for the
production and assembly of the parts. The crash  safety and occupant protection  of the new Audi
Q7 are also on the highest level."

   In order to achieve the fullest amount of mass reduction from lightweighting efforts, vehicle
design and planning are important in order to determine additional secondary mass that may be
reduced from the vehicle. Secondary mass savings are identified as a result of primary mass
reduction savings. Primary mass savings are those items which are not dependent on a lighter
overall vehicle and include such items as aluminum closures and lightweight seats. The most
identifiable secondary mass is the adoption of a smaller engine in the light weighted vehicle.
Ford mentioned in a 2010 International Magnesium Association article that "Strategic use of
lightweight and down-gauged material allows a vehicle's powertrain to be smaller and more
11 Some vehicles were redesigned twice from 2008 and so the changes aren't exactly the same as noted in the articles,
  from which some of the information was taken, for the table references differences between 2008 and 2014.
                                             5-165

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
fuel-efficient. Combining magnesium with aluminum for the MKT liftgate's panels instead of
steel saves 22 pounds in vehicle weight. When coupled with other weight-saving measures, re-
matching the vehicle with a smaller powertrain - known as right-sizing of power to weight — is a
key factor in achieving greater fuel economy."429 The adoption of Ford's EcoBoost engines
allow Ford to realize the benefits of secondary lightweighting.

   Downsizing is an option not considered for this analyses for lightweighting and not
commonly seen in the marketplace to date.  GMC designed the MY2017 Arcadia to be 6.4 inches
shorter in wheelbase and 3.5 inches narrower than its predecessor and adopted some lightweight
solutions for a 7001b reduction in mass in addition to being designed to meet the IIHS small
overlap test.430 The new vehicle achieves 22 city and 28 highway, a 22 percent increase over the
original 17/24, with its mass reduction, aerodynamics, new 2.5L Ecotec engine and stop/start
technology.  "The original Acadia was very truck-inspired, but the new model has a decidedly
SUV influence conveyed in sculptural details, softened corners and a sleeker windshield
angle."431

5.2.7.4 Holistic Vehicle Mass Reduction and Cost Studies

   EPA and  NHTSA's feasibility assessments for the 2012 Light Duty FRM incorporate mass
savings and  related costs.  The 2017-2025 FRM Joint Technical Support Document contained a
linear mass reduction cost curve for direct manufacturing costs (DMC) in the expression of DMC
($/lb.)=$4.36(percent-lb.) x Percentage of Mass Reduction level (percent) as shown in Figure
5.53. This equation starts at $0/kg for no mass reduction and increase  at a constant rate of
$4.36/( percent-lb.) for each percent mass reduction (ex: $0.44/lb. for 10 percent MR on a 4,000
Ib. vehicle and $0.66/lb. for 15 percent on same) and was applied to all 2008/2010 MY vehicles.
This cost curve expression was based on a number of available data sources on mass reduction
which included a number of papers on individual components.
$1.00
^$0.80
.Q
^$0.60
tt
& $0.40
D $0.20
$0.00
0
Mass Reduction Cost




^



^^
X



X^



_4*
X

s

x^


ope =
_X
^


= 4.36
,X



^
^














% 5% 10% 15% 20% 25%
Percent of Mass Reduction
 Figure 5.53 Mass Reduction Cost Curve ($/lb.) for 2017-2025 LD GHG Joint Technical Support Document

   In order to capture a more complete picture of the potential for mass reduction and related
costs, the agencies (EPA, NHTSA, ARB, and DOE) have committed significant resources to
acquire mass and cost information through a number of holistic vehicle studies as listed in
                                             5-166

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Table 5.15. The projects were performed with constant performance in mind and hence the
benefits of all lightweighting efforts were put into improving fuel efficiency and lowering CCh
emissions. Each project includes many steps including baseline vehicle teardown,
component/system examination for mass reduction technologies, direct manufacturer cost
estimation for mass reduction technology and related tooling, CAE safety crash evaluation, NVH
assessment and durability analyses. Mass reduction technologies for these studies are found in a
variety of sources including those found on other vehicles, technologies in development at
suppliers and material companies, technologies developed in other government funded projects,
etc. Cost estimates were made by the project contractors based on their extensive automotive
experience and industry contacts. The DOE/Ford/Magna joint project itself did not include a
cost study for its two evaluations - Mach 1 (25 percent MR) and Mach 2 (50 percent MR).
However DOE did fund two independent cost studies related to this work. One for a 40-45
percent mass reduction vehicle whose results were presented at the DOE Annual Merit Review
(AMR) in 2015 and a second independent study was also funded by DOE for a 20-25 percent
mass reduced vehicle and results are expected sometime in 2016. The Mach 1 work also
included several  additions which included the buildup of seven lightweight vehicles for a number
of durability and crash analyses as well as testing of some of the project's new technologies.
Two other studies provided insights into the mass add for meeting the IMS small overlap test
which is required in order to achieve the IIHS rating of Top Safety Pick. NHTSA funded a
follow-up study on their 2012 passenger car work and Transport  Canada funded a follow-up
study on the EPA 2015 light duty pickup truck.  The studies provided a revised final cost and
mass reduction to the original works. The agencies also greatly appreciate and acknowledge the
work of many individual companies, academia representatives, and material associations to
provide information on lightweighting technologies, both in production and in research, to the
agency contractors for the holistic vehicle studies.  This information was also used as the basis
for material information contained in the Appendices to address topics of feasibility, mass
reduction, cost, safety, research and recycling. In addition, the agencies greatly appreciate the
feedback from OEM's and others on the results of the holistic vehicle studies which formed a
basis for revisions to the individual study cost curves for  this analysis.
                                             5-167

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                Table 5.15 Agencies Sponsored Mass Reduction Project List since FRM

Pass
Car/CUV
Studies
Light
Duty
Truck
Studies
Agen
cy
US
EPA
ARB
NHTS
A
DOE/
Ford/
Magn
a
NHTS
A
EPA
NHTS
A*
Trans
port
Cana
da
Description
Phase 2 Midsize
CUV
(2010 Toyota
Venza)
Low Development
(HSS/AI focus)
Phase 2 Midsize
CUV
(2010 Toyota
Venza)
High Development
All Aluminum
Passenger Car
(2011 Honda
Accord)
Passenger Car
(2013 Ford Fusion)
Mach land Mach
2 projects
Cost Study for 40-
45% Mass
Reduction
Passenger Car
small overlap mass
add
2011 Silverado
1500
2014 Silverado
1500
IIHS small overlap
mass add on LOT
(EPA)
Completion
Date
2012
2012
2012
2015
2016
2015
2016
2015
Reference
Final Report, Peer Review and SAE Paper
https://www3. epa.gov/otaq/climate/documents/420rl20
26.pdf
https://www3.epa.gov/otaq/climate/documents/420rl20
19.pdf
SAE Paper 2013-01-0656
Final Report and Peer Review
http://www.arb.ca.Rov/msproR/levproR/leviii/final arb p
hase2 report-compressed.pdf
http://www.arb.ca.Rov/msproR/levproR/leviii/carb versio
n lotus project peer review.pdf
Final Report, Peer Review, OEM response, Revised Report
ftp://ftp.nhtsa.dot.Rov/CAFE/2017-25 Final/811666.pdf
http://www.nhtsa.gov/Laws+&+Regulations/CAFE+-
+Fuel+Economy/ci.NHTSA+Vehicle+Mass-Size-
Safety+Workshop. print
http://www.nhtsa.gov/staticfiles/rulemakinR/pdf/cafe/81
2237 LiRhtWeiRhtVehicleReport.pdf
http://energy.gov/sites/prod/files/2015/06/f24/lm072_sk
szek_2015_o.pdf
http://enerRV.Rov/sites/prod/files/2014/07/fl7/lm072 sk
szek 2014 o.pdf
http://enerRV.Rov/sites/prod/files/2014/07/fl7/lm088 sk
szek 2014 o.pdf
http://avt.inl.Rov/pdf/TechnicalCostModel40and45Percen
tWeiRhtSavinRs.pdf
SAE papers include
2015-01-0405..0409
2015-01-1236.. 1240
2015-01-1613.. 1616
Final Report
http://www.nhtsa.Rov/staticfiles/rulemakinR/pdf/cafe/81
2237 LiRhtWeiRhtVehicleReport.pdf
Final Report, Peer Review and SAE Paper
https://www3.epa.Rov/otaq/climate/documents/mte/420
rl5006.pdf
SAE Paper 2015-01-0559
Final Report (in peer review)
Final Report and Peer Review
https://www.tc.gc.ca/enR/proRrams/environment-etv-
summary-enR-2982.html
Peer Review (EPA docket)432
Note:
*Completion expected May-June 2016
                                              5-168

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   The holistic vehicle studies in Table 5.15 are nearly all focused on 2008/2010 design era
vehicles due to the fact that one purpose of the Draft TAR is to evaluate the assumptions utilized
in the FRM and perform an updated assessment based on the information available today. The
majority of vehicles have not yet incorporated significant mass reduction technologies due to the
fact that many vehicle designs were already underway when the rulemakings were finalized and
the lead time required to achieve such a transition is influenced by a three year lead time433 for
acquiring aluminum sheet in volume.  The MY2014 new generation light duty pickup truck
evaluated by NHTSA was a 'next step' approach to evaluate the mass  save and cost from
converting from a more high  strength steel approach (compared to the 2008 design) to other
lightweight materials  including aluminum and CFRP. It should be noted that the cost curve
expression for the EPA and NHTSA projects take different approaches as will be discussed
throughout the following sections.

   The agencies are using the information in the publicly available government sponsored
studies in its modeling of mass reduction and related costs for all the vehicles sold in the US.
The vehicles for the holistic vehicle projects were chosen based on their representation of high
sales volume vehicles, as the  Honda Accord and Chevy Silverado 1500, and/or representative of
new vehicle designs that were showing increasing popularity, as the Toyota Venza.  The projects
were conducted over the past 6 years and were multi-million dollar efforts. The same detailed
information collected in these projects were not readily available from any other source -
especially cost information and secondary mass effects. Additional mass comparison
information was found to be available through the A2Macl  vehicle databases and that
information has been  used to  supplement our analyses on mass differences - especially on mass
add for vehicle footprint increases. Ducker Worldwide executive summaries have also provided
insights into aluminum and steel material trends.

   To understand how the results from our projects relate to real world lightweighting efforts,
staff from the agencies, EPA, NHTSA and ARE, met with OEMs and attended many technical
conferences over the past four years. It was observed that there are some cost savings to be
achieved from lightweighting MY2008/2010 design vehicles and more is expected as costs are
reduced through material recycling and optimization of material use.  The agencies agree that
some mass reduction  technologies will add cost, however recent developments in material
processing, as with development of 3rd generation steels and Alcoa's Micromill for aluminum,
indicate that these costs may be less than that utilized in the studies.  In addition, the decrease in
metal material pricing over the past year has not been included in most of the holistic vehicle
studies.  The agencies understand that OEM's have typically utilized mass reduction technologies
to offset the weight of added features or safety measures to remain competitive.

5.2.7.4.1      EPA Holistic Vehicle Mass Reduction/Cost Studies

   The U.S. EPA funded two holistic vehicle mass reduction/cost studies for the Midterm
Evaluation between 2010 and 2015. The first study was the Phase 2 low development (steel
BIW) lightweighting  study on a Midsize CUV performed by EPA with FEV North America,
Inc., EDAG, Inc. and Munro  and Associates, Inc. and was focused on achieving 20 percent mass
reduction which resulted in a high strength steel structure with aluminum closures amongst other
technologies. This was a follow up to the Phase 1 paper study on the  Midsize CUV performed
by Lotus Engineering and includes in-depth analyses on cost and CAE safety analyses of the
vehicle. The second study was a lightweighting study on a 2011MY light duty pickup truck and
                                             5-169

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
was performed by the same contractors using a similar methodology however added in the
dynamic vehicle analyses and a number of component evaluations performed in CAE space. The
result was an aluminum intensive vehicle with high strength steel/aluminum ladder frame.

   EPA's cost curve development methodology for both projects is based on a cumulative
additive approach of the best $/kg rated technologies. Primary mass reduction technologies,
technologies not dependent on mass savings in other areas of the vehicle, are listed along with
the related costs and mass savings.  The $/kg for each technology is calculated and then the order
of the technologies are sorted from lowest $/kg to highest. The original mass and costs are then
each added in a cumulative manner and then the resultant $/kg is calculated at each technology
and a related percent mass reduction. Secondary mass savings, those mass savings which are
dependent on  other mass savings within the vehicle, are noted on a component evaluation basis,
summed,  and then applied at the solution point for the project.  Since the secondary mass savings
are based on the size of the component - hence material basis - then this can be proportioned
across the whole range of primary mass reduction curve. The cost savings are also proportioned.
Two assumptions work into this costs curve methodology: 1) OEM's will adopt cost saving mass
reduction technologies first; and 2) secondary mass savings, such as a resized engine, will occur
at all percent mass reduction points.  This methodology works into EPA's mass reduction
modeling methodology for this Draft TAR, however is different from NHTSA's cost curve
methodology and assumptions which is described separately.

   Other related studies to the Phase 2 Low Development Midsize CUV include the Phase 2
High Development study funded by ARB. ARB hired Lotus Engineering to compete an in-depth
look into the aluminum intensive (High Development) Midsize CUV and included CAE safety
analyses and an in-depth cost analyses. Both of the Phase 2 studies, High Development and Low
Development, are follow-up studies to the Phase 1 paper study by Lotus Engineering on the
Midsize CUV. Following the Phase 2 studies, the Aluminum Association Automotive
Technology Group contracted with EDAG, Inc. to evaluate aluminum material replacement
within EPA's CAE model of the Midsize CUV BIW. A cost analyses was also performed by
EDAG for this project.

5.2.7.4.1.1    Phase 2 Low Development Midsize CUV Updated Study and Supplement

   The Phase 2 Low Development (steel BIW) Midsize  CUV lightweighting study was
completed in August of 2012.  The results of this work were peer reviewed through an
independent contractor as well as through the SAE paper publication process. Feedback was
received by OEM's and others independent of the official peer review process.

   The MY2010 Toyota Venza was chosen as the base vehicle for this  work and vehicle
teardown and  coupon testing revealed that the base vehicle BIW included high strength steel
components made of HSLA 350, HSLA 490, DP500, a 7000 aluminum rear bumper and HF1050
B pillar and side roof rail. After consideration of nearly  150 lightweighting ideas, the project's
final lightweighting results  stated that 18.5 percent mass reduction was achieved for a cost
savings of $0.47/kg.  The report also stated that if aluminum doors were included then the mass
save would be 20.2 percent with a cost savings of $0.1 I/kg. To make the non-compounded cost
curve, the primary lightweighting ideas were listed with the lowest $/kg to the highest $/kg
which reflects an approach where the OEM's would choose the less expensive, or cost saving,
technologies first. Then the mass and cost data were individually cumulatively added and a
cumulative $/kg was determined at each technology addition to create the non-compounded
                                            5-170

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
curve.  The compounded curve was developed by determining the secondary mass savings at the
primary solution point and then the mass savings were ratio'd across the primary cost curve to
yield the final cost curve with compounding.  A short summary of this work and the cost curve,
see Figure 5.54, were included in the 2012 FRM. The compounded cost curve was not included
in the cost curve development in the FRM as the study was not completed in time for the FRM
analysis.
Cost/Kibgrant of Mass Reduction
Vehicle Level Cost Curve



0






V
_f*>*
; 5% 10%p*15%^V^ 20% 25%
J&**

wff' )\\ Optimized Vehicfe Solution
Jr {-$0.47/kg, 13-26%)
J
*
% Vehide Mass-Reduction
       Figure 5.54 Original Phase 2 Low Development Midsize CUV Lightweighting Cost Curve434

   Additional consideration was given to the feedback EPA and FEV received on the study as
well to methodology updates which were made during the 2011MY light duty truck
lightweighting study after the FRM. Modifications made to the data for the original curve, shown
in Figure 5.54, included adding in the aluminum doors as a lightweight technology, and
removing several features including the magnesium engine block and the cost savings for some
of the light weighted plastic components. Several customer features were put back into the
vehicle including the lumbar and active head rest for the back seat and  the cargo cover. A mass
and cost allowance for NVH was added as well as the related cost savings for the secondary
mass which had not been accounted for in the FRM methodology. The revised cost curve is
shown in Figure 5.55 and is 17.6 percent mass reduction at +$0.50/kg.  Also included are the
$/kg and percent mass reduction solution points for two aluminum BIW Midsize CUV studies.
First is the work funded by ARB from Lotus Engineering on the Phase 2 High Development
Midsize CUV aluminum intensive project which utilized an aluminum  BIW  design and results
came in at -$0.64/kg for 31 percent MR,440 per our calculations of study results. Second is the
aluminum intensive point from the Aluminum Association work of 27.81 percent mass reduction
at $1.12/kg, in which EDAG utilized the  same CAE baseline model developed for the EPA
Phase 2 Low Development Midsize CUV work.442
                                            5-171

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
       $4
                 EPA Final Vehicle Solution (HSS BIW)
                 ($0.50/kg @ 17.6% Mass Reduction)
                                     V = 3968.3X3 - 1282.6X2 + 160.78x - 9.9319
             I—I—I—I—|—I—I—I—I—|—I—I
ARB/Lotus Engineering Aluminum
Intensive Desien (-$0.64/ke (est) at 31%)
                                                               I — I — I — I — | — I — I
    20%
                                                                        25%
                                                                              _
                                            Aluminum Association Inc. Published Data Point
                                            Developed from EPA Venza Analysis and EDAG Al
                                            Intensive BIW ($1.12/kg 27.81% Veh  MR)
                                      % Vehicle Mass Reduction
             Figure 5.55 Revised Cost Curve for the Midsize CUV Light Weighted Vehicle

   This cost curve, in Figure 5.55, is clearly different from the 2012 FRM cost curve for mass
reduction, in Figure 5.53, in which all mass reduction points were associated with positive costs.
The EPA Phase 2 Low Development Midsize CUV holistic vehicle study is a whole vehicle
study which examines nearly every component in the vehicle for mass reduction potential and
calculates a related cost and mass save for each and reviews them from most cost/kg save to
most costly cost/kg.  This methodology was chosen based on the understanding that OEMs will
choose the cost saving technologies first and that some cost mass reduction technologies will be
paid for by the cost save mass reduction technologies. A vehicle cost curve similar to the FRM
expression could be achieved if cost technologies were listed first in the cumulative adding
approach and hence losing the appearance of the cost saving technology ideas. However, this is
not the approach OEM's are utilizing for lightweighting. For example, a 2016 publication by
CAR  contains an illustration and caption which  states that "(Figure 5.56) illustrates a generic
cost curve for lightweighting that is broadly supported.436 GM has also claimed publically to its
potential investors that over $2B435 was saved in material costs reveals that costs can be saved
with mass reduction ideas over several passenger vehicles.  It is very likely that some of this
savings was due to the  decreased material  costs over the past year in addition to the cost saving
lightweighting approaches.
                                              5-172

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
          Figure 7: General Auto Manufacturer Cost Curve to Lightweight Vehicles
                        COST
                       INCREASE
                        so




                        COST
                       SAVINGS
                                                      Marginal
                                 Cost Savings
                                                                Expensive
                                           % Mass Reduction
          ©Center for Automotive Research
Page | 12
Figure 5.56 Cost Curve Figure from CAR: "A Cost Curve for Lightweighting That Is Broadly Supported'
5.2.7.4.1.2   Light Duty Pickup Truck Light- Weighting Study

   The U.S. EPA NVFEL contracted with FEV North America to perform this study utilizing the
methodology developed in the Midsize CUV lightweighting effort (2012) and the study was
completed in 2015.  The results of this work went through a detailed and independent peer
reviewed as well as through the SAE paper publication process. Feedback was received by
OEM's and others independent of the official peer review process.

   For this study a 2011 Silverado 1500 was purchased and torn down.  The components were
placed into 19 different systems. The components were evaluated  for mass reduction potential
given research into alternative materials and designs. The alternatives were evaluated for the
best cost and mass reduction and then compared to each other.  CAE analyses for NVH and
safety was completed for the baseline and the light-weighted aluminum  intensive vehicle. A
high strength steel structure with aluminum closures was the first choice of a solution for this
project; however, this was not fully completed for the decision was made by the project team to
change course and pursue the aluminum structure solution due to the expected introduction of the
aluminum intensive F150 into the marketplace. Durability analyses on both the baseline and
light-weighted vehicle designs were performed through data gathered  by instrumenting a
Silverado  1500 light duty pickup truck and operating it over various road conditions.  Included in
the durability analyses are durability evaluations on the light weighted vehicle frame, door and
other components in CAE space. The crash and durability CAE analyses allowed for gauge and
grade determinations for specific vehicle  components. Load path redesign of the light duty truck
structure (cabin and box structure and vehicle frame) was not a part of this project.

   As shown in Figure 5.57, the most mass reduction was achieved in the Body System Group -
A- (Body  Sheet metal) in which the cabin and box structure and the closures, etc. were converted
                                             5-173

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
to aluminum.  The suspension system is the second highest system for mass reduction and
includes composite fiber leaf springs. Mass reduction technologies with cost save examples
include 1) material and design optimization in the connecting rods, 2) material and design
through use of vespel thrust washer versus roller bearings, 3) material processing in the Poly one
and Mucell applications, 4) material substitution in the thermoplastic vulcanizates (TPV) vs.
EPDM static and dynamic weather seals, 5) material and part consolidation in the passenger side
airbag housings, and 6) design and processing through incorporation of the half shafts and the
Vari-lite® tube process by U.S. Manufacturing Corporation. A complete listing of vehicle
technologies can be found in the online report437 and Figure 5.57 shows that there was a 50kg
and $150 allowance for NVH considerations.

3

15
v 	
2
3
4
5
6
7
8
9
10
11
1?
13
14
15
16
17
18
19
20
to
a.
o>
u
)OS
*"oi 	
r 	
03A
03B
03C
03D
' 04
" 05
' 06
' 07
' 09
10
' 11
' 12
r"l5~
17
ris
00
Description

E!£.s .CM^^^^
Engine System
Transmission System
Body System Group -A- ( Body Sheetmetal)
Body System Group -B- (Body Interior)
Body System Group -C- (Body Exterior Trim)
Body System Group D (Glazing & Body Mcchatronics)
Suspension System
Driveline System
Brake System
frame and Mounting System
Exhaust System
Fnfil System
Steering System
Climate Cunliul System
Information, Gage and Warning Device System
Electrical Power Supply System
In-Vehicle Entertainment System
Lighting System
Electrical Distribution and Electronic Control System
Fluids and Miscellaneous Coating Materials
a. Analysis Totals Without NVH Counter Measures — >
b Vehicle NVH Counter Measures (Mass & Cost ) --
c. Analysis Tulals Willi NVH Cuunlei Measures — >

Mass Reduction Impact by Vehicle System
(Includes Secondary Mass Savings)
Base
Mass
"kg"

	 2393 	
145.3
574.7
247.0
40.5
50.9
301.2
183.8
101.0
267.6
38.4
?fi 3
32.5
20.3
1.6
21.1
2.2
9.6
33.0
116.8
2454.4
0.0
2454.4

Mass
Reduction

	 sTs 	
39.4
207.1
34.0
2.1
4.5
105.4
20.4
45.8
23.7
6.9
7 3
8.b
1.9
0.2
12.8
0.0
0.4
8.0
0.0
560.9
-50.0
510.9
(Decrease)
Cost
Impact
NIDMC
"$" (2)

-92.83
-96.57
-1194.86
-127.23
2.73
2.30
-154.90
38.01
-148.92
-54.42
-13.69
11 9?
-14/.46
14.71
066
-172.73
0.00
-2.00
01.44
0.00
-2073.82
-150.00
-2223.82
(Increase)
Cost/
Kilogram
NIDMC
"W (2)

	 -Z92 	
-2.45
	 -577 	
-3.74
1.28
0.51
-1.47
1.86
-3.25
-2.30
-1.97
1 6?
-1/44
7.59
2.66
-13.49
0.00
-5.18
7.20
0.00
-3.70
n/a
-4.35
(Increase)
Cost/
Kilogram
NIDMC +
Tooling
"$/kg" (2)
-2.63
-2.47
	 -5.77 	
-378
1.28
0.51
-1.48
1.89
-335
-2.30
-1.97
1 77
-1/.45
7.59
297
-1344
000
-5.18
727
0.00
-3.69
n/a
-4.35
(Increase)
System
Mass
Reduction

	 1373%
27 1%
	 3676% 	
13.8%
5.3%
8.9%
35.0%
11.1%
45.4%
0.9%
18 1%
?7 9%
260%
9.5%
15.7%
60.6%
0.0%
4.0%
25.2%
0.0%
n/a
n/a
n/a

Vehicle
Mass
Reduction

	 ?"3% 	
1.6%
	 8.4% 	
1 .4%
0.1%
0.2%
4.3%
0.8%
1 .9%
1.0%
0.3%
03%
0.3%
0.1%
0.0%
0.5%
0.0%
0.0%
0.3%
0.0%
22.9%
n/a
20.8%

 (1) Negative value (i.e., -X.XX ) represents an increase in mass
 r(2) Negative value (i.e., 4X.XX) represents an increase in cost

                 Figure 5.57 Light Duty Pickup Truck Lightweighting Study Results

   The individual technology mass and cost saving used to develop the system summaries listed
in Figure 5.57 were used to develop EPA's cost curve for the light duty pickup truck
lightweighting study, as shown in Figure 5.58.  It should be noted that the blue squares are
                                              5-174

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
individual solutions and are not based on the cost curve technology points which lead to the red
square solution point.
f.
3
- 
S
5 tn
.1 o
3 54
tj
* S8

Aluminum Intensive Body and HSS
Intensive Frame w/ Mass Compounding J\^
+>•* ^- "
A At •*"
* ^IP M^fc ^
' ^A^ '
.A*
"***
Aluminum Inten5ive Bodyand
Frame w/ Mass Corn pound ing
20% 25%
^j^ HSS Intensive Bodyand Frame w/
^^r Mass Compounding
^

1
*
% Vehicle Mass Reduction
^w/o Compound ing
Aw/Compounding

                  Figure 5.58 Light Duty Pickup Truck Lightweighting Cost Curve

   The curve without compounding in Figure 5.58 (green curve) includes primary mass
reduction ideas which do not depend on the vehicle being made lighter. The mass reduction
ideas based on a resultant lighter vehicle are called secondary mass saving ideas and are based on
components decreasing in size and hence material.  In this study the engine was able to be
downsized 7 percent due to the mass reduction in the vehicle design and still maintain the current
towing and hauling capacities. The other systems that were reduced in size, while considering
truck performance  characteristics, included the transmission, body system group A (bumpers),
suspension, brake,  frame and mounting systems, exhaust, and fuel systems. The systems
considered for secondary mass are included in Figure 5.59 and show the total 83.9kg mass save
at $68.74 savings.  Overall, the secondary mass savings are  17.6KK percent of the primary.  The
compounded curve in Figure 5.58 is the EPA light duty truck cost curve utilized in the
development of the overall cost curve for light duty trucks described in section 5.3.
  % Secondary Mass = 560.9 compounded-83.9secondary =477kg primary, 83.9/477 = 17.6% secondary.
                                             5-175

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment

s
15
1
2
3
7
9
10
11
12
0)
S.
1
D
DOS
01
02
03A
04
06
07
09
10
Description
eries Chevrolet Silve
Engine System
Transmission System
Body System Group -
A- ( Body Sheetmetal)
Suspension System
Brake System
Frame and Mounting
System
Exhaust System
Fuel System
a. Analysis Totals Without NVH
Counter Measures — >
Secondary Mass Savings (SMS) Impact by Vehicle System
Base
Mass
"kg"
rado Pick-
239.9
140.3
574.7
301.2
101,0
2676
38.4
263
1694.5
Mass
Reduction
with SMS
Jp Truck
31.8
39.4
207.1
105 4
45.8
23.7
6.9
7.3
467.5
Mass
Reduction
without
SMS

23.8
34.2
190.7
83 1
43.9
00
6.3
1.6
383.6
Incremental
Mass
Reduction
from SMS
"kg" (1)

80
0.2
16.4
224
2.0
237
0.6
5.7
83.9
Cost
Impact
NIDMC will I
SMS
"$" (?)

-9283
-90.07
-1194.86
-154.90
-148.92
-54.42
-13.69
11.92
-1744.26
(Increase)
Cost
Impact
NIDMC
without
SMS
"$" (2)

-114.63
-128.20
-1125.15
-26084
-167.87
0.00
-19.54
3.25
-1813.00
(Increase)
Savings
riom SMS
T (2)

21 81
31.04
-69.71
105.94
18.95
-54.42
5.85
8.67
68.74
Cost/
Kilogram
NIDMC with
SMS
"$/kg" (?)

-292
-2.40
-5.77
-1.47
-3.25
-230
-1.97
1.62
-3.73
(Increase)
Cost/
Kilogram
NIDMC
without
SMS
"$/kg" (2)

-4.82
-3.70
-5.90
-3.14
-3.83
000
-3.08
2.02
-4.73
(Increase)
Cost
Savings/
Kilogram
NIDMC
from SMS
"$/kg" (2)

1.90
1.30
0.13
1.67
0.58
-230
1.11
-0.40
0.82
 (1) Negative value (i.e., -X.XX ) represents an increase in mass
 '(2) Negative value (i.e., -$X.XX) represents an increase in cost
5.2.7.4.2
Figure 5.59 Light Duty Pickup Truck Lightweighting Study Secondary Mass

NHTSA Holistic Vehicle Mass Reduction/Cost Studies
   NHTSA funded two holistic vehicle mass reduction/cost studies for the Midterm Evaluation.
The first lightweighting study was performed on a 2011MY Honda Accord as the base vehicle,
with Electricore, Inc., George Washington University and EDAG, Inc. and was completed in
2012438. EDAG was also rehired to re-evaluate the public study feedback received from Honda
on the project as well as evaluate the mass add for IIHS Small Overlap for the passenger car.
This study was completed in February 2016.  The second was a lightweighting study on a
2014MY light duty pickup truck, Silverado 1500 as the base vehicle, and was performed by
EDAG, Inc. using a similar methodology to the passenger car work and is expected to be
completed in 2016.
5.2.7.4.2.1
 Updated Midsize Car Lightweight Vehicle Study
   At the time of the original 2012 passenger car lightweighting study438, NHTSA did not
consider IIHS small overlap test performance as part of overall safety assessment of light-
weighted vehicle. Honda commented on the above light-weighted study and highlighted some of
the performance, build quality, platform sharing and other customer experience constraints that
should be taken into consideration.  NHTSA updated the above Honda Accord light-weighted
vehicle study in the new report "Update to Future Midsize Lightweight Vehicle Findings in
Response to Manufacturer review and IIHS Small-Overlap Testing."439  The mass and cost
                                             5-176

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
adjustments in response to Honda's comments resulted in 21.75kg less mass reduction from the
original light-weighted vehicle and an increase cost of $18.13.  Further, to address IIHS small
overlap test, the mass of the light-weighted vehicle had to be increased by another 6.9kg with
$26.88 increased cost on top of $18.13.  The resultant cost curve is displayed in Figure 5.60.
LWV1.0 is the original AHSS BIW and Aluminum Closures and Chassis Frames solution point.
LWV 1.1 includes the corrections based on the Honda feedback and LWV 1.2 includes the
Honda feedback as well as the IIHS SOL mass and cost add.  This cost curve has been further
revised for the Draft TAR as discussed below.

         n~.lcr~ v.hid. ) - S-
         J-« L 1/JHf  O.&*4
                          S.Ml       10 OH       IS OH       20 OH
                                Mass Reduction (% of Curb Vehicle Weight)
           Figure 5.60 NHTSA Passenger Car Updated Cost Curve (DMC($/kg) v %MR)442

   The final light-weighted vehicle (LWV 1.2 Solution) had mass reduction of 303.65kg
compared to the baseline vehicle at the cost of $364.01 after accounting for Honda's
recommendation and IIHS small overlap tests.  The green point in the cost plot in Figure 5.60
shows revised cost and mass reduction levels after consideration of Honda's recommendations
and the mass addition to meet IIHS small overlap test performance.  As explained in section
5.2.7.4.2, NHTSA developed cost curve based on the LWV solution point which is explained in
detail in section 5.4.

   NHTSA realized some limitations in the form of the cost curve in Figure 5.60. Since the cost
curve was derived more at the systems level, a more detailed cost curve was developed using
cumulative mass savings approach from each of the components considered for mass reduction
opportunities. Figure 5.61 shows the cost curve developed from Honda Accord light-weighted
vehicle. Table 5.16 shows the list of components considered for mass reduction. Note here the
LWV solution is represented as AHSS+AL solution point in the cost curve below.
                                             5-177

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5S


JT
•<"•.,


SO -
0
NHTSA Passenger Car Cost Curve





























X
,^&
If



5

















y

*?
x*i

Si








= 16C

^


11


2*c

-»


%


-70

-=^:




95x
I

• 	 .




+ 7
v: =

^^^^
~ 	




80.5
994

~~~^
• 	




r -



4.02



6j


0.25

^^


~^^.


22


— —
• AHSS
Solution
15%

Mass Reduction % (curb We
—#— Primary Mass Reduction • Optimized Solutions (Honda/501

21






















• .
!





U.
olut


%



25




an


%
C
S







tKP
aluti







Ml













30
W
ight)
	 Poly. (Primary Mass Reduction}
     Figure 5.61 NHTSA Revised Passenger Car DMC Curve (($/kg v %MR) and ($/vehicle v %MR))

   Table 5.16 shows list of components considered for mass reduction and used for constructing
the passenger car cost curve.
                                             5-178

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                           Table 5.16 Components for LWV Solution
Vehicle
Component/System
Front Bumper
Front Door Trim
Front Door Wiring Harness
Head Lamps
HVAC
Insulation
Interior Trim
Parking Brake
Rear Door Trim
Rear Door Wiring Harness
Tail Lamps
Tires
Wiring and Harness
Wheels
Rear Bumper
Instrument Panel
Body Structure
Decklid
Hood
Front Door Frames
Fenders
Seats
Rear Door Frames
Powertrain components
(Engine, transmission, Fuel
system, Exhaust system,
coolant system), Brakes
etc.
Cumulative Mass
Saving (kg)
3.59
4.93
5.23
6.94
9.54
12.74
15.77
16.76
17.89
18
18.63
23.08
27.38
28.82
32.33
41.78
96.18
101.39
108.86
124.26
127.53
147.56
159.02
302.92
Cumulative MR%
0.24%
0.33%
0.35%
0.47%
0.64%
0.86%
1.07%
1.13%
1.21%
1.22%
1.26%
1.56%
1.85%
1.95%
2.18%
2.82%
6.50%
6.85%
7.36%
8.40%
8.62%
9.97%
10.74%
20.5%
Cumulative Cost
(S)
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-$1.23
$0.53
$17.27
$173.13
$188.97
$211.49
$262.88
$274.98
$374.02
$428.47
364.37
Cumulative Cost
$/kg
-0.34
-0.25
-0.24
-0.18
-0.13
-0.10
-0.08
-0.07
-0.07
-0.07
-0.07
-0.05
-0.04
-0.04
0.02
0.41
1.80
1.86
1.94
2.12
2.16
2.53
2.69
1.20
5.2.7.4.2.2    Light Duty Pickup Truck Light- Weighting Study

   The Department of Transportation National Highway Traffic Safety Administration (NHTSA)
awarded a contract to an automotive design and engineering company EDAG, Inc., to conduct a
vehicle weight reduction feasibility and cost study of a 2014MY full size pick-up truck.  The
light weighted version of the full size pick-up truck (LWT) used manufacturing processes that
will likely be available during the model years 2025-2030 and capable of high volume
production.  The goal was to determine the maximum feasible weight reduction while
maintaining the same vehicle functionalities, such as towing, hauling, performance, noise,
vibration, harshness, safety, and crash rating, as the baseline vehicle, as well as the functionality
and capability of designs to meet the needs of sharing components across same or cross vehicle
                                             5-179

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
platform. Consideration was also given to the sharing of engines and other components with
vehicles built on other platforms to achieve manufacturing economies of scale, and in
recognition of resource constraints which limit the ability to optimize every component for every
vehicle. At the time of writing for this Draft TAR, the report is in peer review and will be
finalized by the NHTSA NPRM and EPA Proposed Determination in 2017.

   A comprehensive teardown/benchmarking of the baseline vehicle for engineering analysis
that included manufacturing technology assessment, material utilization and complete vehicle
geometry scanning was performed. The baseline vehicle's overall mass, center of gravity and all
key dimensions were determined.  Before the vehicle teardown, laboratory torsional stiffness
tests, bending stiffness tests and normal modes of vibration tests were performed on baseline
vehicles so that these results can be compared with the CAE model of the light weighted design.
After conducting a full tear down  and benchmarking of the baseline vehicle, a detailed CAE
model of the baseline vehicle was created and correlated with the available crash test
results.  The project team then used computer modeling and optimization techniques to design
the light-weighted pickup truck and optimized the vehicle structure considering redesign of
structural geometry,  material grade and material gauge to achieve the maximum amount of mass
reduction while achieving comparable vehicle performance as the baseline vehicle. Only
technologies and materials projected to be available for large scale production and available
within two to three design generations (e.g. model years  2020, 2025 and 2030) were chosen for
the LWT design. Three design  concepts were evaluated, a multi-material approach, an
aluminum intensive approach and a Carbon Fiber Reinforced Plastics (CFRP) approach. The
multi-material approach was identified as the most cost effective. The recommended materials
(advanced high strength steels, aluminum, magnesium and plastics), manufacturing processes,
(stamping, hot stamping, die casting, extrusions, and roll forming) and assembly methods (spot
welding, laser welding, riveting and adhesive bonding) are at present used, some to a lesser
degree than others. These technologies can be fully developed within the normal product design
cycle using the current design and development methods.

   The design of the LWT was verified, through CAE modeling, that it meets all relevant crash
tests performance. The LS-DYNA finite element software used by the EDAG team is an
industry standard for crash simulation and modeling. The researchers modeled the
crashworthiness of the LWT design under the NCAP Frontal, Lateral Moving Deformable
Barrier, and Lateral Pole tests, along with the IIHS Roof, Lateral Moving Deformable Barrier,
and Frontal Offset (40 percent and 25 percent) tests. All of the modeled tests  were comparable
to the actual crash tests performed on the 2014 Silverado in the NHTSA database. Furthermore,
the FMVSS No. 301 rear impact test was modeled and it showed no damage to the fuel system.

   The baseline 2014 MY Chevrolet Silverado was platform shares components across several
platforms. Some of the chassis components and other structural components were designed to
accommodate platform derivatives, similar to the components in the baseline vehicle which are
shared across platforms  such as GMT 920 (GM Tahoe, Cadillac Escalade, GMC Yukon), GMT
930 platform (Chevy Suburban, Cadillac Escalade ESV,  GMC Yukon XL), and GMT 940
platform (Chevy Avalanche and Cadillac Escalade EXT) and GMT 900 platform (GMC Sierra).
As per the National Academy of Sciences guidelines, the study assumes engines would be
downsized or redesigned for mass reduction levels at or greater than  10 percent. As a
consequence of mass reduction, several of the components used designs that were developed for
other vehicles in the weight category of light-weighted designed vehicles were used to maximize
                                            5-180

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
economies of scale and resource limitations. Examples include brake systems, fuel tanks, fuel
lines, exhaust systems, wheels etc.

   Cost is a key consideration when vehicle manufacturers decide which fuel-saving technology
to apply to a vehicle. Incremental cost analysis for all of the new technologies applied to reduce
mass of the light-duty full-size pickup truck designed were calculated. The cost estimates
include variable costs as well as non-variable costs, such as the manufacturer's investment cost
for tooling etc. The cost estimates include all the costs directly related to manufacturing the
components.  For example, for a stamped sheet metal part, the cost models estimate the costs for
each of the operations involved in the manufacturing process, starting from blanking the steel
from coil through the final stamping operation to fabricate the component.  The final estimated
total manufacturing cost and assembly cost are a sum total of all the respective cost elements
including the costs for material, tooling, equipment,  direct labor, energy, building and
maintenance.

   The information from the LWT design study was used to develop a cost curve representing
cost effective full vehicle  solutions for a wide range  of mass reduction levels. The cost curve is
shown in Figure 5.54. At lower levels of mass reduction, non-structural components and
aluminum closures provide weight reduction which can be incorporated independently without
the redesign of other components and are stand-alone solutions for the LWV. The holistic
vehicle design using a combination of AHSS and aluminum provides good levels of mass
reduction at reasonably acceptable cost.  The LWV solution achieves 17.6  percent mass
reduction from the baseline  curb mass. Further two more analytical mass reduction solutions (all
aluminum and all carbon fiber reinforced plastics) were developed to show additional  mass
reduction that could be potentially achieved beyond the LWV mass reduction solution point. The
Aluminum analytical solution predominantly uses aluminum including chassis frame and other
components. The carbon fiber reinforced plastics analytical solution predominantly uses CFRP)
in many of the components. The CFRP analytical solution shows higher level of mass reduction
but at very high costs. Note  here that both all-Aluminum and  all CFRP mass reduction solutions
are analytical solutions only and no computational models were developed to examine all the
performance metrics.

   An analysis was also conducted to examine the cost sensitivity of major vehicle systems to
material cost and production volume variations.
                                             5-181

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
 $20.00

 $15.00

 $10.00 -

  $5.00 -

  $0.00
)

 -$5.00

 -$10.00

 -$15.00

 -$20.00

 -$25.00
                          NHTSA Light Truck Mass Reduction Cost Curve
                     y = 172698X5 - 40530x4 - 2587.5x3 + 1022x2 - 25.202x + 0.1934
                                     R2 = 0.9965
                        4%   6%    8%   10%   12%   14%   16%   18%   20%   22%  24%   26%
                                                %MR
    Figure 5.62 NHTSA Draft Light Duty Pickup Truck Lightweighting (AHSS Frame with Aluminum
                            Intensive) Cost Curve (DMC $/kg v %MR)
   Table 5.17 Components for LWV Solution, below lists the components included in the
various levels of mass reduction for the LWV solution.  The components are incorporated in a
progression based on cost effectiveness.
                                                5-182

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                           Table 5.17 Components for LWV Solution
Vehicle Component/System
Interior Electrical Wiring
Headliner
Trim - Plastic
Trim - misc.
Floor Covering
Headlamps
HVAC System
Tail Lamps
Chassis Frame
Front Bumper
Rear Bumper
Towing Hitch
Rear Doors
Wheels
Front Doors
Fenders
Front/Rear Seat & Console
Steering Column Assy
Pickup Box
Tailgate
Instrument Panel
Instrument Panel Plastic Parts
Cab
Radiator Support
Powertrain
Cumulative
Mass Saving
1.38
1.56
2.59
4.32
4.81
6.35
8.06
8.46
54.82
59.93
62.96
65.93
77
102.25
116.66
128.32
157.56
160.78
204.74
213.14
218.66
221.57
304.97
310.87
425.82
Cumulative
MR%
0.06%
0.06%
0.11%
0.18%
0.20%
0.26%
0.33%
0.35%
2.25%
2.46%
2.59%
2.71%
3.17%
4.20%
4.80%
5.28%
6.48%
6.61%
8.42%
8.76%
8.99%
9.11%
12.54%
12.78%
17.51%
Cumulative
Cost
($28.07)
($29.00)
($34.30)
($43.19)
($45.69)
($45.69)
($45.69)
($45.69)
$2.57
$7.89
$11.04
$14.13
$28.09
$68.89
$92.53
$134.87
$272.57
$287.90
$498.35
$538.55
$565.06
$580.49
$1,047.35
$1,095.34
1246.68
Cumulative
Cost $/kg
-20.34
-18.59
-13.24
-10.00
-9.50
-7.20
-5.67
-5.40
0.05
0.13
0.18
0.21
0.36
0.67
0.79
1.05
1.73
1.79
2.43
2.53
2.58
2.62
3.43
3.52
2.93
   A fitted curve was developed based on the above listed mass reduction points to derive cost
per kilogram at distinct mass reduction points as shown in Table 5.18 below.
                      Table 5.18 Costs Per Kilogram at Various %MR Points
MR%
5.0%
7.5%
10.0%
15.0%
20.0%
$/kg
$0.97
$2.09
$2.98
$3.27
$5.75
                                             5-183

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   As explained above, the total direct manufacturing costs for the components listed above are
shown Figure 5.63 below.
NHTSA Light Truck Cost Curve (Direct Manufacturing Cost)
$3,500 -
$3,000 -
$2,500 -
$2,000 -
$1,500 -
$1,000
$500 -
$0 <
O1
-$500 -






E.. ^
._ — — — -
/o

















^^^




^,,--'
»•+— *-*4'"'"1




+ 	







i
/
	 "T\HSS +AL
Solution (LW




f-


fV)









5% 10% 15% 20% 25%
   Figure 5.63 NHTSA Light Truck Cost Curve (Direct Manufacturing Costs) $/vehicle vs
%MR

   Table 5.19 shows the direct manufacturing costs are distinct mass reduction levels.
                      Table 5.19 Direct Manufacturing Costs at MRO-MR5
LT
Baseline Curb Wt. 2432 kg
MRO
MRl-5%
MR2-7.5%
MRS -10%
MR4-15%
MRS -20%
Mass
Reduction
(kg)
0
122
182
243
365
486
DMC ($)LL
$0
$118
$381
$725
$1193
$2797
5.2.7.4.3
ARE Holistic Vehicle Mass Reduction/Cost Study
   The California Air Resources Board funded Lotus Engineering on further analysis of in-depth
cost and CAE, of the Phase 2 High Development of the Midsize CUV440. The project focused on
  Value calculated from best fit curve in previous figure, not from figure above table.
                                             5-184

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
the BIW design through CAE and more in-depth costing of the BIW. A full vehicle solution
point was developed by adding the cost and mass save results of the BIW analysis to the cost and
mass save information on the other vehicle systems from the Phase 1 work.441 The report
changed the original BIW design of 30 percent magnesium, 37 percent aluminum, 6.6 percent
steel and 21 percent composites to one of 12 percent magnesium, 75 percent aluminum, 8
percent steel and 5 percent composites, shown in Figure 5.65. The report states that its BIW
design reduced the number of parts from 419 parts in the baseline Venza to 169 parts in the low
mass design.  Specifically the report states "By factoring in the manufacturability of the materials
and designs into the fundamental design process, it is expected that... this type of design [will]
be production ready in 2020."

   The summary write-up for this work is contained within the LD GHG 2017-2025 FRM Joint
Technical Support Document. A cost curve was not developed for this work.  Values  of cost and
overall mass reduction were located in several areas of the report. The overall results, including
all of the mass reduction items in the Phase 1 report and including powertrain were taken from
Table 4.5.7.2.f. totaling 531.2kg reduced (31 percent of 1711kg) and the total cost was taken
from the 4.6.1.  Conclusions section of $342/vehicle cost save. The cost per kilogram for this
solution is calculated as -$0.64/kg cost save. This point, along with two other all aluminum
vehicle solution points - one by NHTSA and the other by the Aluminum Association, helps to
indicate the direction for additional mass reduction beyond the AHSS BIW/Aluminum closure
solution on which the cost curve for the passenger car/Midsize CUV is based.
                       Key:
                       Silver - Aluminum
                       Purple - Magnesium
                       Blue - Composite
                       Red - Steel
                   Figure 4.2.3.a: Body-in-white material usage front three-quarter view

                  Figure 5.64 Phase 2 High Development BIW - Lotus Engineering


5.2.7.4.4     Aluminum Association Midsize CUV Aluminum BIW Study

   The Aluminum Association funded a project with EDAG, Inc.442 in 2012 to perform an
aluminum substitution analysis in the BIW of the Midsize CUV work by EPA using the EPA
CAE baseline model for the work. The baseline model was also developed by EDAG, Inc.  The
analyses utilized CAE crash safety and NVH verifications when determining the specifics, gauge
and grade, of the aluminum to be utilized in the BIW (Figure 5.65).
                                             5-185

-------
                       Technology Cost, Effectiveness, and Lead-Time Assessment
              Toyota Venza AIV
                     Baseline Venza
                                                  Venza AIV
                                              Aluminum ,
                                    28% Mass     m TotalMass.
                                                   1,237 Kg
                                                         OriveAluminum.org
              Toyota Venza AIV | Material Selection
                       6022 T6 Alloy Sheet
                       5754 O Alloy sheet
                       6082 T6 Extrusion
                       Generic Casting
290 Mpa
117 Mpa
315 Mpa
160 Mpa
Figure 5.65 Midsize CUV Baseline vs Midsize CUV Aluminum Intensive Vehicle
                                        5-186

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Description
Body Structure Subsystem
UnderbodyAsy
Front Structure Asy
Roof Asy
Bodyside Asy
Ladder Asy
Bolt on BIP Components
Body Closure Subsystem
Hood Asy
Front Door Asy
Rear Door Asy
Rear Hatch Asy
Front Fenders
Bumpers Subsystem
Front Bumper Asy
Rear Bumper Asy
Totals
Estimated
Mass
Reduction
"Kg"

19.8
14.3
14.6
72.2
38.1
3.2

7.7
15.0
11.3
7.2
2.0

2.3
0.0
207.7
"+" = mass decrease, "-" = mass increase
"+" = cost decrease. "-" = cost increase

Estimated
Cost Impact
"$"

-67.56
•121.84
-44.81
•306.60
•235.53
•3.97

•27.70
•21.65
-19.31
-21.21
•16.22

•8.60
0.00
•895.01
Average
Cost/
Kilogram
"$/Kg"

•3.41
-8.49
-3.07
•4.25
•6.19
•1.23

•3.62
•1.44
-1.70
-2.93
•8.25

•3.82
0.00
•4.31

  Figure 5.66 Summary Table of Mass Reduction and Cost for Aluminum BIW and Closure Components

  Figure 5.66 lists the results from aluminum material substitution into the existing BIW and
closures. When combined with the remaining mass and cost saved identified in the U.S. EPA
Midsize CUV report, resulted in a $1.12/kg for 27.8 percent mass reduction for the entire
vehicle, as shown in Table 5.20. This data point is included in the overall cost curve shown in
Figure 5.55.

                     Table 5.20 Summary of the Automotive Aluminum 2025

Body and Closure MR
Total Vehicle MR
Cost Impact
Multi-Material
(MMV-EPAIowdev)
-14%
-19.2%
-$0.23/kg
Aluminum (AIV)
-39%
-27.8% (-476kg)
$1.12/kg (+$534)*
       *Note: Full Vehicle Mass Optimization
5.2.7.4.5
DOE/Ford/Magna MMLV Mach 1 andMach 2 Lightweighting Research Projects
   The Multi Material Lightweight Vehicle (MMLV) project was initiated in 2012 by the
Department of Energy and co-funded by Magna International and Ford Motor Corporation under
the project number DE-EE0005574. The objectives of the project included identifying 25
percent (Mach 1) and 50 percent (Mach 2) vehicle mass reduction packages. This work was peer
reviewed through the DOE AMR and the SAE publication processes.  The "Multi-Material
Lightweight Vehicles" presentation, which was a combination of the Mach 1 and Mach 2
projects, was peer reviewed at the 2015 DOE AMR in front of a panel of experts in the field and
the results of the peer review were included in the final report for the DOE AMR.443  The project
received a weighted average score of 3.77 out of 4.0 and was measured on reviewer questions
related to approach, technical accomplishments, collaborations, and future research.  The results
were also presented in a number of SAE papers and hence reviewed through the SAE publication
process.
                                            5-187

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
   The DOE/Ford/Magna developed the lightweight vehicle solutions off of a 2013MY Ford
Fusion platform (used to represent a 2002 Ford Taurus). Results include 23.5 percent for the
Mach  1 design. Seven vehicles were built and the vehicles, and certain components, were tested
under a series of durability tests. New technologies of composite fiber springs, carbon fiber
wheels, seat back frame, and the multi-material body structure were included in the durability
tests. For the Mach 2 design, 50 percent mass reduction is achieved however the vehicle is not
market viable due to extensive de-contenting and use of materials that are not yet ready for full
volume production including composite "tub" package tray and roof. A comparison of the
MMLV structures weight for BIW, Closure, Chassis and Bumper is displayed in Figure 5.67.
                                                                         &VEHMA
MMLV Structures Weight Comparison   ©ENERGY
BIW, Closure, Chassis, Bumper
                  Baseline
                 BIW   316.04 kg
              Closures    32.17 kg
               Chassis    89.07 kg
              Bumpers	20.38 kg
                Totals  517.ee kg
                             MMLV Mach I
                               BIW 231.33 kg
                            Closures  57.23 kg
                             Chassis  52.30 kg
                            Bumpers   11.13 kg
                              Totals 352.58 kg
                              31.9% Reduction
  MMLV Mach II
   BIW  17259 kg
Closures   45.16 kg
 Chassis   30.80 kg
Bumpers   11.13 kg
  Totals  259.67 kg
  49.3% Reduction
                                                                    3.4%
                                                                               12.7%
          • Aluminum Stampinjs ^Aluminum Extrusions  • Hot Stsmpin8s        .COMPOSITE         • MAGNESIUM
          • Aluminum Castings  • Steel Stampings     • Fasteners/Sleeves/Other  ,MASNESiUMWAi,MFoRMiN6 «S™G/FORi3IMe/SXTRUSiON

                                                                          • CAD WEIGHT
                             TBs presefHattxi does no) corteBi any p
          Figure 5.67  MMLV Structures Weight Comparison BIW, Closure, Chassis, Bumper44
   Gaps identified by the MMLV projects (I and II) include those listed in Table 5.21.
                                                5-188

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                         Table 5.21 Gaps Identified by MMLV Project
Topic
Steel
Aluminum
Magnesium
Carbon Fiber
Composites
Multi Material
Vehicles
GAP
Improved coatings on ultra-high strength steels for multi material applications
Increased die life and bi-metallic (inserts, etc.) for Al die castings plus low cost
7xxx aluminum sheet and extrusions
High volume warm forming, hemming, class A finish, plus improved die life and
bi-metallic inserts in high pressure vacuum die casting
Material characterization for CAE, joining, corrosion, paint, class-A finish
Corrosion mitigation strategy including universal equivalent of phosphate (or
eqiuv) bath for any mix of steel, aluminum and magnesium before e-coat and
paint
Joining methods with corrosion mitigation
Aluminum rivet, high hardness, high strength
Alternative NVH treatments for lightweight panels sheet metal and glazings
Design for disassembly, end of life, for reclaiming, recycling
   No cost analysis was performed for the Mach 1 study. A 40-45 percent MR cost analyses
from the base 2013MY vehicle was completed under a separate DOE project, through Idaho
National Laboratories performed by IBIS Associates Inc., and results indicate the cost of carbon
fiber must decrease in order to make the technology viable for mass market vehicles.445 This
project is described in 5.2.7.4.6.

5.2.7.4.5.1   Mach I

   The MMLV Mach I project achieved 364 kg (23.5 percent) mass reduction from the baseline
weight of the 2013 Ford Fusion (representing a 2002 Ford Taurus). Seven prototype vehicles
were built and these vehicles were used to conduct a number of test such as, corrosion,
durability, NVH (noise vibration harshness), and crash. Maintaining performance and
capabilities,  along with safety and durability were also goals of the MMLV. All parts used in the
MMLV are either low volume or high volume production capable up to 250,000 vehicles per
year. The Mach I mass reduction was achieved using materials such as aluminum, carbon fibers,
magnesium,  and high strength steels. Results of the Mach I project were presented in 14 SAE
papers.
446 447 448 449 450 451 452 453 454 455 456 457 458
   The Mach I project group presented an estimate of the fuel economy improvement at the 2015
SAE World Congress 2015 as being an increase to 34 mpg from 28 mpg. This change in fuel
economy was estimated by taking the fuel economy of a Ford Fiesta (which is the equivalent
weight of the lightweight Mach-I) and comparing to the 2013 Ford Fusion.  The fuel economy
numbers were from fueleconomy.gov. Key requirements of durability, safety, and Noise
Vibration Harshness (NVH) were also met within the Mach I design as illustrated in a report
presentation at the 2015 DOE AMR.459 All components of the MMLV were specifically chosen
for optimal weight reduction without shorting on performance or technicality.

   Five subsystems of the mach-i compared to the baseline 2013 fusion of full body mass
reduction.459
                                             5-189

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
          The body-in-white (BIW) and closures contributed 76 kg (4.9percent) to the overall
          vehicle mass reduction. The baseline 2013 BIW is 326 kg and the Mach-I BIW is
          250 kg. The 2013 Fusion BIW is steel intensive, and the Mach-I design included
          advanced high strength steels were integrated for use as primary safety structures like
          crush rails, B-pillars, and selected cross car beams.  Closures in the Mach 1 were
          aluminum intensive. The transition from steel to aluminum is also the primary design
          strategy for the light weighting of the deck lid, and front fenders, as well as the side
          door structures and hinges. Also, chemically foamed plastics were used in the door
          design as trim.
          Body Interior and Climate Control consists of the seats, floor components, instrument
          panel/ cross car beam (IP/CCB), and climate control system which contributed 28 kg
          (1.8 percent) to the overall vehicle mass reduction.  The IP/CCB decreased in part
          count from 71  to 21, new material design involved carbon fiber reinforce nylon from
          the baseline welded assembly  of steel stampings and tubes.  The material selection of
          the seat structures was carbon fiber reinforced nylon composite compared to the
          baseline steel stampings and tubes.
          Chassis subsystem reduced its total mass by 98 kg (6.3 percent) to the overall vehicle
          mass reduction. The major components identified in the Mach 1 subsystem include
          hollow coil springs, carbon fiber wheels, and tires with a tall and narrow design,
          hollow steel stabilizer bars, aluminum sub frames, control arms and links.
          The powertrain subsystem was reduced by 73 kg (4.7 percent) to the overall vehicle
          mass reduction. The baseline  engine is a 1.6 liter four-cylinder gasoline turbocharged
          direct injection (EcoBoost) with a six-speed automatic transmission. The Mach-I
          design has a 1.0 liter three-cylinder gasoline turbocharged direct injection (Fox
          EcoBoost) with a mass reduced six-speed automatic transmission. The use of carbon
          fiber within this subsystem encouraged mass reduction and include components such
          as the engine oil pan.
          The electrical subsystem achieved a 10 kg (0.64 percent overall vehicle mass
          reduction). A few adjustments were made to accomplish this number. The battery
          was switched to a lithium ion  12-volt start battery from the baseline lead-acid battery.
          The change of the battery achieved 5 kg mass reduction.  Also, copper electrical
          distribution wiring was replaced with aluminum conductors meeting a 4 kg mass
          reduction.  The remaining 1 kg mass reduction was  achieved by small adjustments to
          the speakers, alternator, and the starter motor.
   DESIGN AND FUNCTION VALIDATION: The Mach-I used computer aided engineering
(CAE) for many safety simulations due to low budget, however several vehicles were used to
perform a number of actual vehicle safety crashes. Many computer aided designs (CAD) and
CAE tests were performed initially before the vehicle components were manufactured and/or
physically tested.  Seven MMLV Mach-I vehicles were built and selectively tested. Seven
different validation tests were completed as listed in Table 5.22.
                                             5-190

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                        Table 5.22 Safety Tests Performed on the Mach-I.
VEHICLE
Test Buck
Durability A
Corrosion A Traditional Surface
Treatments
Corrosion B MMLV Alternative
Surface Treatments
Safety A
Safety B
NVH + Drives
TESTING
Body-in-White + Closures + Bumpers + Glazing + Front
Subframe - Body-in-Prime NVH modes, global stiffness,
attachment stiffness, selected Durability
DRIVABLE, full MMLV content with Fusion powertrain -
MPG Structural Durability, Square Edge Chuckhole Test
for Wheels and Tires
DRIVABLE, with alternative surface treatment and paint
process - MPG Corrosion R-343. Humidity soaks and
salt spray etc.
DRIVABLE, with traditional surface treatment and paint
process - MPG Corrosion R-343. Humidity soaks and
salt spray etc.
NON-Drivable, most MMLV content, without carbon
fiber instrument panel - Low Speed Damageability test
(front) Right Hand (passenger) side - IIHS Front ODB
40% Offset 40 mph, Left Hand (driver) side - Side Pole
Test on Right Hand (passenger) side (FMVSS 214)
NON-Drivable, most MMLV content, without carbon
fiber instrument panel - NCAP Frontal 35 mph rigid
wall, then 70% Offset Rear Impact (FMVSS 301)
DRIVABLE, full MMLV content with downsized and
boosted powertrain, 1.0-liter 13 EcoBoost, gasoline
turbocharged direct injection engine plus six-speed
manual transmission - Wind Tunnel, Rough Road
Interior Noise, Engine & Tire Noise, Ride & Handling
   The overall outcome of the safety and durability tests provided assurance a multi-material
lightweight vehicle was successful. Noise Vibration Harshness was tested in a high frequency
range of 200-10000 Hz and fell within acceptability but slightly short of requirements. Durability
test classified the Mach-I as a durable vehicle and showed no major cracking or durability
incidents in the test mileage. Frontal crash safety tests showed that nine parts withstood the test
at a good level. Table 5.23 is a list of the parts that performed the best. The carbon fiber wheels
had one issue in the durability test with the outer coating on the carbon fiber, however it was
solved and the wheel is currently planned for the Shelby Mustang.  The composite fiber springs
performed better than expected and it is understood that they are in production, or planned for
production, in the Audi A6 Ultra Avant and the Renault Megane Trophy RS vehicles.  The
durability issue for the composite fiber wheels was solved and the improved wheels are being
employed in the Shelby Mustang.  Some new discoveries were made including the near zero
mass add for NVH considerations  and corrosion concerns will be better addressed with a correct
amount of sealant and the proper choice of nuts and bolts in the multi material vehicle design.
                                             5-191

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
              Table 5.23  Mach-I Components to Maintain Frontal Crash Performance.
PART
Front bumper
Crush Can
Subframe
Shock Tower
Coil Spring
Wheel
A-Pillar joint node
Windshield
Seat frame
MATERIAL
Extruded aluminum
Extruded aluminum
Cast and extruded aluminum
Cast aluminum
Chopped glass fiber composite
Woven carbon fiber composite
Cast aluminum
Chemically toughened laminate
Woven carbon fiber composite
5.2.7.4.5.2    Mach2

   The goal of the Mach 2 project was to create a lightweight design that achieved 50 percent
mass savings from the 2013 Ford Fusion (representing a 2002 Ford Taurus). This amount of
mass reduction is forward looking and of limited use for the time frame considered for this Draft
TAR (2022-2025) which has a top application of 20 percent mass reduction.

   The project achieved 51.1 percent (798kg) mass reduction with a significant degree of mass
reduction using materials and processes that have some initial research but not ready for high
volume. Significant vehicle de-contenting was employed which included items from air
conditioning to thinning the windows and the resultant vehicle was not marketable.

   The vehicle technologies for the BIW and Closures includes carbon fiber and composites as
seen in Figure 5.68.  However the CAE inputs were not mature for the materials and as a result
the outputs were insufficient.  CAE information included cards for stiffness, durability, and
fatigue analyses. In terms of production, the composite material and manufacturing infrastructure
was also not mature for automotive volumes. The carbon fiber and composite panels were not
deemed acceptable for Class A surfaces and as a result aluminum or magnesium sheet products
were chosen for the BIW and closure applications.
                                             5-192

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
                          Table 5.24 Mach II Design Vehicle Summary459
System
Body and Closures
Interior & Climate
Control
Chassis
Powertrain
Electrical
Technology
Body
Closures
Windows
Seats
IP
Reduced content
Subframes
Coil Springs
Reduced
capacity
Engine
Transmission


Material/Approach
Composite intensive
Magnesium
Reduced Thickness
Carbon fiber seats with reduced function
Carbon fiber composite
No bins, center console, air conditioner, etc.
Cast magnesium
Composite
For reduced weight cargo and towing
l.OL 3 cyl naturally aspirated
Remove turbocharger and intercooler
Material change
Reduced capacity manual
Eliminate content and features
Reduced battery, alternator, wiring
            Mach II Design
            Mixed Material BIW & Closures
                                                                   Multi-Matenai Lightweight Vehicle
                      Body-in-White (BIW)   155 kg mass reduction from baseline (47.5%)
              ALUMINUM SHEET
             I ALUMINUM CASTING
I ALUMINUM EXTRUSION
I HOT STAMPING
I COMPOSITE
 MAGNESIUM SHEET
I MAGNESIUM EXTRUSION
I MAGNESIUM CASTING
                        CLOSURES  47.0 kg mass reduction from baseline (48.0%)
              51%
                             THIS presentation does not contain any propnetaiy, confidential, or omerwise restricted information
         Figure 5.68 Mach II Mixed Material BIW and Closure Design (brown is carbon fiber)459

   The Mach II design had a number of estimated performance impacts.  The CAE based
assessments were not complete due to insufficient  carbon fiber CAE modeling capabilities and as
a result there was low confidence in load cases.  There was a large degradation in all metrics for
sound and stiffness.  Corrosion capability was significantly challenged with mixed material
joints that included carbon fiber composites and magnesium. There are some unknown
processes for high volume production and challenges with joining, surface treatments, paint,
                                                5-193

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
thermal expansion and dimensions and tolerances. Areas identified needing additional research
include recyclability and vehicle repair.

5.2.7.4.6      Technical Cost Modeling Report by DOE/INL/IBIS on 40 Percent-45 Percent
Mass Reduced Vehicle

   The U.S. Department of Energy's Vehicle Technologies Office Materials Area funded a study
to provide cost estimates and assessment of a 40 percent and 45 percent weight savings on a
North American midsize passenger sedan based on the work of the Mach 1 and Mach 2
lightweighting projects. The title of the report is "Vehicle Lightweighting: 40 percent and 45
percent Weight Savings Analysis: Technical Cost Modeling for Vehicle Lightweighting"460.
This work was peer reviewed through the 2015 DOE AMR "Technical Cost Modeling for
Vehicle Lightweighting" presentation in front of a panel of experts in the field. Results of the
peer review were included in the final report for the DOE AMR.461 The project received a
weighted average score of 2.98 out of 4.0 and was measured on reviewer questions related to
approach, technical accomplishments, collaborations, and future research.

   The goal of the work was to achieve 40 percent-45 percent mass reduction relative to a
standard North American midsize passenger sedan at an effective cost of $3.42/lb. This study
utilized existing mass reduction and/or cost studies including those from FEV, Lotus
Engineering, DOE Mach 1 and Mach 2. The Executive Summary to this report states "The
analysis indicates that a 37 to 45 percent reduction in a standard mid-sized vehicle is within
reach if carbon fiber composite materials and manufacturing processes are available and if
customers will accept a reduction in vehicle features and content,  as demonstrated with the
Multi-Materials and Carbon Fiber Composite-Intensive vehicle scenarios."
                                             5-194

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
                                        Vehicle Lighlweighting Scenario Comparison
           $27.000
           $25,000
         8  $23,000
         o
         01
         I
         6
           $21,000
$19,000
           $17,000
                                                               A Carbon Stage 5
                                                                (45% Weight Reduction)
                                                  Carbon Stage 4
                                                  (40% Weight Reduction)
                     Multi-Material Stage 4
                     (Maximum Potential)
Carbon Scenario Meeting
Cost Target (S6/lb for materials
and S5/lb for processing)
                                                           Carbon Scenario Meeting
                                                           Cost Target (S4.20/lb for materials
                                                           and $5/lb for processing)
                                        Aluminum Intensive Stage 4
                                        (Maximum Potential)
           $15,000
                30%     32%     34%     36%     38%     40%     42%

                                                 % Weight Savings
                                                                     44%
                                                                            46%     48%
                                                                                           50%
      Figure ES-I. Costing results of advanced weight savings scenarios based ondit't'erent material systems. Carbon scenarios assume an optimistic,
      projected, carbon composite processing cost of $5/lband current carbon fiber price of $12 50/lb.


Figure 5.69  Technical Cost Modeling Results for 40 Percent to 45 Percent Lightweighting Scenario (Based on

                                 Mach 1/Mach 2 Project Technologies)
5.2.7.4.7
    Studies to Determine Mass Add for IIHS Small Overlap
   The lightweighting analysis within the Midterm Evaluation will give credit for mass adds due
to safety regulations and requirements.  One of the requirements of the IIHS Top Safety Pick is
to meet the IIHS small overlap crash test. The IIHS SOL test is designed to reproduce what
happens when the front corner of a vehicle hits another vehicle or an object like a tree or utility
pole. Estimating the mass impact to succeed this test can vary widely among different types of
vehicles.  The structure of the vehicle must be  redesigned in order to design load paths such that
the passenger compartment remains sound throughout the crash event.
                   Figure 5.70 Post-test Laboratory Vehicle of IIHS Small Overlap Test
                                                    5-195

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Two studies were funded to examine the mass add to existing vehicle study models.  NTHSA
funded the passenger car study using their LWV model and Transport Canada funded the light
duty truck study using the LDT model from the EPA light duty pickup truck study. All of the
CAE modeling, from the base studies to the IMS small overlap studies were performed by two
separate groups within EDAG, Inc.  The results of these studies are described in the following
sections.

5.2.7.4.7.1    NHTSA Mass Add Study for a Passenger Car to Achieve a "Good" Rating on the
IIHS Small Overlap

   The analysis of the IIHS Small Overlap resultant mass add for a variety of unibody passenger
car vehicle classes are included in the February 2016 report "Update to Future Midsize
Lightweight Vehicle Findings in Response to Manufacturer review and IIHS Small-Overlap
Testing."439 In order to improve the structural performance during the IIHS SOL test, several
options were considered and implemented using a detailed LS-DYNA crash model that was
originally part of the NHTSA LWV study. The CAE model was first updated to address the
concerns in performance as identified by Honda. Changes regarding the SOL test include
reinforcement of major areas in the body structure and were designed for easy manufacturability
and assembly into the body structure. The findings for the IIHS SOL solution was a mass add of
6.9kg and 26.88 in cost.

   The report also includes the IIHS mass add results for a range of unibody vehicle classes as
shown in Table 5.25 (MY2010) and Table 5.26 (MY2020). Although the IIHS SOL test came
out in 2012, the MY2010 refers to the baseline used in the NHTSA work in which it is assumed
that all vehicles have no mass reduction technology. The individual mass adds are based on
formulas determined for various vehicle classes with unibody design. The overall Light Duty
Vehicle Average is based on a  straight average of the values for each vehicle class. The report
also notes that estimated  mass increases for 'body on frame' vehicles should be further reviewed
due to a differing body structure design. This was done in Transport Canada's evaluation of the
2011 Silverado 1500 discussed in the section  following this section.
           Table 5.25 Estimated Mass Increase to Meet IIHS SOL for 2010 Vehicle Classes

Vehicle Class
Sub-Compact Car
Compact Car
Mid-Sized Car
Small SUV/LT
Large Car
Mid-Sized SUV/LT
Minivans
Large SUV/LT
Light Duty Vehicle
Average
2010 Vehicle Class Average
Curb Vehicle
Weight (kg)
1261
1345
1561
1592
1752
1916
2035
2391
1732
Test Vehicle
Weight (kg)
1411
1495
1711
1742
1902
2066
2185
2541
1882
Increase in mass to
meet IIHS SOL (kg)
7.4
7.8
8.9
9.1
9.9
10.8
11.4
13.3
9.8
Curb Vehicle Weight with IIHS
SOL Changes (kg)
1268
1353
1570
1601
1762
1927
2046
2404
1741
                                            5-196

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
            Table 5.26 Estimated Mass Increase to Meet IIHS SOL for 2020 Vehicle Classes

Vehicle Class
Sub-Compact Car
Compact Car
Mid-Sized Car
Small SUV/LT
Large Car
Mid-Sized SUV/LT
Minivans
Large SUV/LT
Light Duty Vehicle Average
2020 Vehicle Class Average
Curb Vehicle
Weight (kg)
1055
1119
1294
1318
1453
1632
1689
1962
1440
Test Vehicle
Weight (kg)
1205
1269
1444
1468
1603
1782
1839
2112
1590
Increase in mass to
meet IIHS SOL (kg)
6.3
6.6
7.5
7.7
8.4
9.3
9.6
11.0
8.3
Curb Vehicle Weight
with IIHS SOL
Changes (kg)
1062
1125
1302
1326
1462
1641
1699
1973
1449
5.2.7.4.7.2     Transport Canada Mass Add Study for a Light Duty Truck to Achieve a "Good"
Rating on the IIHS Small Overlap

   A body on frame 2013MY Silverado 1500 light duty pickup truck (designed in 2007) was
evaluated for necessary mass add in order to achieve a "Good" rating on the IIHS small overlap
crash test in both current and lightweighted designs. This information was needed in order to
evaluate the mass impact from compliance with the safety crash test.

   Transport Canada funded the project with EDAG, Inc.462 in which the light duty pickup truck,
utilized in EPA's light-weighting light duty pickup truck study through FEV437, was evaluated
for mass add in the light-weighted aluminum intensive design with the goal of achieving a Good
or Acceptable rating.  The report and models have been peer reviewed through EPA's peer
review process.

   The baseline original CAE model was prepped for the work and then an IIHS small overlap
crash test with  a 2013MY Silverado 1500 Crew Cab 4x4 was conducted with Transport Canada's
Motor Vehicle Test & Research Centre in Blainville, Quebec. This was performed in order to
assure that the  necessary components for the test were modeled correctly in the baseline model
and that the crash could be reproduced in CAE space. A more complete series of CAE tests were
conducted at each point in the process to assure that performance was maintained in all crash
requirements, NVH, etc.  The state of the truck from the barrier impact is shown in Figure 5.71.
A number of components were material tested through the assistance of Natural Resources
Canada's CanmetMATERIALS facility in Hamilton, Ontario. This was done in order to ensure
that the most accurate materials properties were being input into the baseline model at the start of
the process and in order that the CAE modeling could reproduce the video from the actual crash
test as closely as possible. The baseline model was modified with failure criteria and timing of
respective components involved in the IIHS small overlap test.  Figure 5.72 shows the baseline
model correlating to the baseline truck crash event.
                                            5-197

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
         Figure 5.71  MY2013 Silverado 1500 IIHS Small Overlap Test Crash Before and During
                    Figure 5.72 Converting the Actual Crash Event to a Model

   Development of the light duty truck design modifications to the baseline structure began with
research on existing IIHS crash results including those from the GM Equinox, Mercedes ML,
and design information on the 2014MY Silverado 1500 and the 2015MY Ford F150 which had
been released before the conclusion of this project.  A solution for a "Good" rating on the IIHS
small overlap crash test was determined for the steel intensive vehicle in order to highlight the
areas for improvement in the lightweight model. The mass add for this design was not optimized
for the minimum mass add that would still achieve a "Good" rating.

   To develop the lightweight model mass add to the "Good" rating on the IIHS small overlap,
the vehicle lightweighting ideas from the original U.S. EPA lightweight light duty truck project
were first adopted onto the vehicle.  The solution from the baseline vehicle was then optimized
and the mass add determined. The report states "Like the original EPA Project cab, the T5-LW
(light-weighted) cab exploited the low density and manufacturing methods specific to
Aluminum, .. .Extrusions and castings were used to meet and exceed the static bending  and
torsion requirements with mass efficient solutions."  The components in the area of the  crash
(including suspension and wheel) were not changed to aluminum for the failure information for
the aluminum components were not available. The resultant light-weighted model before and
after IIHS small overlap crash is illustrated in Figure 5.73.  The passenger compartment stays in
tact as shown.
                                             5-198

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
               Figure 5.73 Light Weighted Model in the IIHS Small Overlap Crash Test

   The accelerations for the dummies will change based on the stiffer passenger compartment
which doesn't allow the extreme intrusions in the baseline model. The report contains a
comparison of the Velocity (m/s) at CoG X-velocities for the T4-GA LDT model and other
production vehicles with "Good" IIHS small overlap results and the results are similar.  The T5-
LW results are very similar to the T4-GA results. The report concludes that "the pulse response
is considered reasonable and it is expected that a modern restraints system could be tuned to
manage the vehicle response." 462

   The IIHS Small Overlap Rating is based on dummy injury criteria as well as vehicle intrusion
in specified locations within the vehicle. Figure 5.74 illustrates how the light-weighted model
(T5-LW) compares to the baseline model (T3-BL) along with the results from the original crash
test (TC13-018). The light-weighted model achieves a Good rating in the intrusion part of the
evaluation.
                                         IIHS Structural Rating
                                                              TC13-O18 Test
                                                              T3-BL Model
                                                              T4-GA Model
                                                              TS-LW Model
                             Lower Occupant Compartment
                                                     Upper Occupant Compartment
                                                                                     462
  Figure 5.74 Results of the Project Models from Baseline to Light Weighted on the IIHS Small Overlap

   The overall mass reduction results for the LDT were 455kg mass reduction for $2115 and
included added mass to the light-weighted truck of 17kg mass and removal of the possible 89kg
mass reduction which remained to be considered when aluminum components are put into place
for the original steel suspension, wheel, etc. One of the peer reviewers for this report provided
comments to support a decision of the mass add for the aluminum suspension, wheel, etc.  The
decision was made that an additional 5kg mass would be needed to assure the aluminum
                                             5-199

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
components for the test requirement results. As a result, a mass credit of 22kg is assigned for a
light duty aluminum intensive pickup truck to meet the IIHS small overlap test.

5.2.8   State of Other Vehicle Technologies

5.2.8.1 Electrified Power Steering: State of Technology

   Compared to conventional hydraulic power steering, electrified power steering can reduce
fuel consumption and CCh emissions by reducing overall accessory loads. Specifically, it
reduces or eliminates the parasitic losses associated with belt-driven power steering pumps
which consistently draw load from the engine to pump hydraulic fluid through the steering
actuation systems even when the wheels are not being turned. Power steering may be electrified
on light duty vehicles with a standard 12V electrical system. Electrified power steering is also an
enabler for vehicle electrification since it provides power steering when the engine is off.

   Power steering systems can be electrified in two ways.  Manufacturers may choose to
completely eliminate the hydraulic portion of the steering system and provide electric-only
power steering (EPS) or they may choose to move the hydraulic pump from a belt driven
configuration to a stand-alone electrically driven hydraulic pump. The latter system is referred to
as electro-hydraulic power steering (EHPS).

5.2.8.1.1      Electrified Power Steering in the 2012 FRM

   In the  2012 FRM analysis, the agencies estimated a 1 to 2 percent effectiveness of EPS and
EHPS for light duty vehicles, based on the 2002 NAS report, Sierra Research Report and
confidential OEM data. The 2010 Ricardo study also confirmed this estimate.

   For the 2012 FRM, the agencies estimated the DMC at $88 (2007$).  Converting to 2010$,
this DMC becomes $92 for this Draft TAR. The agencies consider EPS technology to be on the
flat portion of the learning curve and have applied a low complexity ICM of 1.24 through 2018
then 1.19 thereafter.

5.2.8.1.2      Developments since the FRM

   Since the FRM, EPS has been successfully implemented on all light duty vehicle classes
(including trucks) with a standard 12V electrical system eliminating the need to consider EFtPS
on larger vehicles.

   For the cost and effectiveness assumptions the agencies are adopting for the GHG Assessment
and CAFE Assessment for this Draft TAR analysis, see Sections 5.3 and 5.4.

5.2.8.2 Improved Accessories: State of Technology

   The accessories on an engine, including the alternator, coolant and oil pumps are traditionally
mechanically-driven. A reduction in CCh emissions and fuel consumption can be realized by
driving them electrically, and only  when needed ("on-demand").

   Electric water pumps and electric fans can provide better control of engine cooling. For
example, coolant flow from an electric water pump can be reduced and the radiator fan can be
shut off during engine warm-up or cold ambient temperature conditions which will reduce warm-
up time, reduce warm-up fuel enrichment, and reduce parasitic losses.
                                             5-200

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Indirect benefit may be obtained by reducing the flow from the water pump electrically during
the engine warm-up period, allowing the engine to heat more rapidly and thereby reducing the
fuel enrichment needed during cold starting of the engine. Further benefit may be obtained when
electrification is combined with an improved, higher efficiency engine alternator. Intelligent
cooling can more easily be applied to vehicles that do not typically carry heavy payloads, so
larger vehicles with towing capacity present a challenge, as these vehicles have high cooling fan
loads. Both agencies also included a higher efficiency alternator in this category to improve the
cooling system.

   The agencies considered whether to include electric oil pump technology for the rulemaking.
Because it is necessary to operate the oil pump any time the engine is running, electric oil pump
technology has insignificant effect on efficiency. Therefore, the agencies decided to not include
electric oil pump technology for this final rule, consistent with the proposal.

   In MYs 2017-2025 final rule, the agencies used the effectiveness value in the range of 1 to 2
percent based on technologies discussed above.  NHTSA did not apply this technology to large
pickup truck due to the utility requirement concern for this vehicle class.

   In the 2017-2025 rule, the agencies estimated the DMC of IACC1 at $71 (2007$). Converting
to 2010$, this DMC becomes $75 for this analysis, applicable in the 2015MY, and consistent
with the heavy-duty GHG rule.  The agencies consider IACC1 technology to be on the flat
portion of the learning curve and have applied a low complexity ICM of 1.24 through 2018 then
1.19 thereafter.

   Cost is higher for IACC2 due to the inclusion of a higher efficiency alternator and a mild level
of regeneration. The agencies estimate the DMC of the higher efficiency alternator and the
regeneration strategy at $45 (2010$) incremental to IACC1, applicable in the 2015MY.
Including the costs for IACC1 results in a DMC for IACC2 of $120 (2010$) relative to the
baseline case and applicable in the 2015MY.  The agencies consider the IACC2 technology to be
on the flat portion of the learning curve.  The agencies have applied a low complexity ICM of
1.24 through 2018 then 1.19 thereafter.

   For the cost and effectiveness assumptions the agencies are adopting for the GHG Assessment
and CAFE Assessment for this Draft TAR analysis, see Sections 5.3 and 5.4.

5.2.8.3 Secondary Axle Disconnect: State of Technology

5.2.8.3.1      Background

   All-wheel drive (AWD) and four-wheel drive (4WD) vehicles provide improved traction by
delivering torque to both the front and rear axles, rather than just one axle. Driving two axles
rather than one tends to consumes more energy due to additional friction and rotational inertia.
Some of these  losses may be reduced by providing a secondary axle disconnect function that
disconnects one of the axles when driving  conditions do not call for torque to be delivered to
both axles.

   The terms AWD and 4WD are often used interchangeably. The term AWD has come to be
associated with light-duty passenger vehicles that provide variable operation of one or both axles
on ordinary roads.  The term 4WD is often associated with larger truck-based  vehicle platforms
                                             5-201

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
that provide for a locked driveline configuration and/or a low range gearing meant primarily for
off-road use.

   Many 4WD vehicles provide for a single-axle (or two-wheel) drive mode that may be
manually selected by the user.  In this mode, a primary axle (perhaps the rear) will be powered,
while the other axle (known as the secondary axle) is not. Even though the secondary axle is not
contributing torque, energy may still be consumed by rotation of its driveline components
because they are still connected to the non-driven wheels. This energy loss directly results in
increased fuel consumption and CCh emissions that could be avoided by disconnecting the
secondary axle components under these conditions.

   Further, many light-duty AWD systems are designed to variably divide torque between the
front and rear axles in normal driving, in order to optimize traction and handling in response to
driving conditions. Even when the secondary axle is not delivering torque, it typically remains
engaged with the driveline and continues to generate losses that could be avoided by a more
advanced disconnect feature. For example, Chrysler has estimated that the secondary axle
disconnect in the Jeep Cherokee reduces friction and drag attributable to parasitics of the
secondary axle by 80 percent when in disconnect mode.463 Some of the sources of secondary axle
parasitics include lubricant churning, seal friction, bearing friction, and gear train losses.464'465

   Many part-time 4WD systems, such as those seen in light trucks, use some type of secondary
axle disconnect to provide shift-on-the-fly capabilities. In many of these vehicles, particularly
light trucks, the rear axle is permanently driven and the front axle is secondary. The secondary
axle disconnect is therefore part of the front differential assembly in these vehicles. Light-duty
passenger cars that employ AWD may instead permanently power the front wheels while making
the rear axle secondary, as currently in production in the Jeep Cherokee 4WD system.

   As part of a shift-on-the-fly 4WD system, the secondary axle disconnect serves two basic
purposes. First, in two-wheel drive mode, it disengages the secondary axle from the driveline so
the wheels do not turn the secondary driveline at road speed, reducing wear and parasitic energy
losses. Second, when shifting from two- to four-wheel drive "on the fly" (while moving), the
secondary axle disconnect couples the secondary axle to its differential side gear only after the
synchronizing mechanism  of the transfer case has spun the secondary driveshaft up to the same
speed as  the primary driveshaft.

   4WD  systems that have a  disconnect typically do not have either manual- or automatic-
locking hubs. To isolate the  secondary wheels from the rest of the secondary driveline, axle
disconnects use a sliding sleeve to connect or disconnect an axle shaft from the differential side
gear.

5.2.8.3.2     Secondary Axle Disconnect in the FRM

   At the time of the FRM, the agencies were not aware of any manufacturer offering secondary
axle disconnect in the U.S. on AWD unibody-frame vehicles.  Secondary axle disconnect was
included in the FRM analysis with the expectation that this technology could be introduced by
manufacturers within the MYs 2017-2025 time period.

   The 2012 FRM analysis assigned an effectiveness of 1.2 to 1.4 percent for secondary axle
disconnect. The agencies estimated the DMC at $82 (2010$). The agencies considered
                                             5-202

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
secondary axle disconnect technology to be on the flat portion of the learning curve and applied a
low complexity ICM of 1.24 through 2018 then 1.19 thereafter.

5.2.8.3.3     Developments since the FRM

   Since the FRM, the agencies have continued to monitor developments in AWD secondary
axle disconnects and their adoption in the light-duty vehicle fleet. We gathered information by
monitoring press reports, holding meetings with suppliers and OEMs, and attending industry
technical conferences.

   EPA coordinated with Transport Canada and Environment and Climate Change Canada on a
project to characterize AWD systems present in the market today. The primary objectives of this
project are to gain an overview of AWD technology in general and to understand the potential
effect of advances in these systems on GHG performance in comparison to their 2WD variants.
A comprehensive technical characterization of 17 in-production AWD systems has been
completed465. It includes characterization of system architecture, operating modes, and current
usage in the fleet. It also estimated and compared the mass and rotational inertia of AWD
components and parts to those of 2WD variants in order to better understand the weight increase
associated with AWD. Additionally, the all-wheel-drive components of three AWD vehicles
(the 2015 Jeep Cherokee Limited 4x4, 2015 Ford Fusion AWD, and 2015 Volkswagen Tiguan
Trendline 4motion) underwent a teardown in order to accurately characterize their mass and
rotational inertia and estimate their approximate cost. One of the teardown vehicles, the Jeep
Cherokee, includes a secondary axle disconnect, indicating that this technology has begun to
appear in light-duty vehicles since the FRM. In 2014, Chrysler Group LLC presented a very
positive outlook on the advantages of this system for improving fuel efficiency while retaining a
highly competitive off-road capability.466 This suggests that the addition of secondary axle
disconnect systems need not be accompanied by loss of traction and handling capability.

   The study reinforced the  perception that AWD is rapidly increasing in popularity in the
vehicle fleet, with about one-third of all vehicles sold in North America in 2015  having AWD
capability. The prevalence of AWD varies significantly between vehicle segments and trim
levels. Sedans have the lowest AWD availability, while AWD versions outnumber 2WD
versions in the SUV and pickup segments, particularly among the higher trim levels in each
segment.

   The study identified several areas of potential efficiency improvement for AWD systems.
These included system level improvements such as: use of a single shaft Power Transfer Unit
(PTU), which can save up to 10kg in mass compared to a two-shaft unit; careful integration into
vehicle architecture; downsizing the driveline to further reduce mass while providing sufficient
traction in adverse conditions; and use of electric rear axle drive (eRAD). Component level
improvements were also identified, including: use of fuel-efficient bearings, low drag seals,
improved lubrication strategies, use of high-efficiency lubricants, advanced CV joints, and dry
clutch systems.  Design improvements such as hypoid offset optimization, bearing preload
optimization, use of single-shaft power transfer units (PTUs)  and an optimized propshaft gear
ratio were also suggested to have potential. Use of weight-reducing metals such  as magnesium,
and manufacturing improvements such as vacuum die casting and improved hypoid
manufacturing were also cited as opportunities. The authors'judgement of the relative potential
for AWD efficiency improvements offered by each opportunity are depicted in Figure 5.75.
                                             5-203

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                       cost / deterioration
                                                            savings/ improvements
            Weight
            Fuel Consumption
            Performance
            Packaging
             System Level
             Disconnect system FWD
             Disconnect system RWD
             downsizing
             Component Level
             FE bearings
            Actuator technology
             Lubrication strategies
             Advanced CV-joints
             Dry clutch systems
                 Figure 5.75 Summary of AWD Efficiency Improvement Potentials465

   Various sources cited in the study suggested that AWD disconnect systems have the ability to
lower fuel consumption of AWD vehicles between 2 percent and 7 percent, significantly higher
than the estimates of 1.2 percent to 1.4 percent used in the 2012 FRM. However, it should be
noted that a disconnect strategy must balance fuel efficiency with other concerns such as vehicle
dynamics, traction and safety requirements, which may act to reduce its actual GHG
effectiveness.

   The study also identified three primary technological trends taking place in AWD system
design, including: actively controlled multi-plate clutches (MFCs), active disconnect systems
(ADS), and electric rear axle drives (eRAD). While controlled MFCs appear to be the dominant
technology in on-demand  systems, ADS is a more recent trend and holds promise for reducing
real world fuel consumption. eRAD is the most recent emerging technology with potential for
even greater improvements (as seen in the Volvo XC90 Hybrid SUV).

   The teardown analysis  analyzed three power transfer units (PTUs) and rear drive modules
(ROMs) from the Ford Fusion, Jeep Cherokee and VW Tiguan. These were non-destructively
disassembled and analyzed with respect to mass, rotational inertia and the presence of specific
design features. Figure 5.76 shows the contribution of individual AWD driveline components to
the total additional mass of the AWD variant of each vehicle compared to the 2WD variant.
Further analysis of rotational inertias of these parts suggested that rotational inertias add very
little equivalent mass and  therefore probably do not carry a large impact on fuel consumption.
                                              5-204

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                           135
                                                             78
                                                        0.9
                                                        11,4
                     Fusion
Cherokee
Tiguan
                                                                     l Other
                                                                     Propshaft
                                                                     I Half shafts
                                                                     I RDM

                                                                     IPTL
   Figure 5.76 Contribution of Individual AWD Driveline Components to Total Additional Vehicle Mass

   The study included a high-level cost analysis for these parts, including the mechanical
disconnect device and modifications necessary to the torque transfer device (TTD). The total cost
of adding secondary axle  disconnect to a vehicle was estimated at approximately $90 to $100.
Although this cost estimate was informally derived based primarily on the experience and
expertise of the authors, it compares well to the total cost (TC) figure attributed to 2017 in the
FRM analysis, at $98. The authors noted that the cost for the Jeep Cherokee system would likely
be higher because this system was designed to accommodate a planetary low gear, which adds
mass and cost not related  to the AWD disconnect function.

   In addition to the in-production disconnect concepts described in the Transport Canada AWD
report, activity continues  in the development of innovative secondary axle disconnect concepts.
For example, in 2015, Schaeffler presented a novel design for a clutch mechanism for use in
AWD disconnect.467 Suppliers are also designing and marketing modular solutions for
integration into existing OEM products.464 Developments such as these suggest that multiple
potential paths will exist for disconnect technology to accompany the increasing growth and
popularity of AWD in light-duty vehicles.

   In conjunction with the AWD characterization project described above, Transport Canada is
also conducting a program of coastdown testing, chassis dynamometer testing, and on-road
testing of several Canada-specification AWD vehicles at Transport Canada facilities. This
portion of the effort is not yet completed but the results may become available to inform the
proposed determination.

   For the cost and effectiveness assumptions the agencies are adopting for the GHG Assessment
and CAFE Assessment for this Draft TAR analysis, see Sections 5.3 and 5.4.
                                             5-205

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.2.8.4 Low-Drag Brakes: State of Technology

   Low or zero drag brakes reduce or eliminate the sliding friction of disc brake pads on rotors
when the brakes are not engaged. By allowing the brake pads to pull or be pushed away from the
rotating disc either by mechanical or electric methods, the drag on the vehicle is reduced or
eliminated.

5.2.8.4.1       Background

   The reduction of brake drag is a technology that the vehicle manufacturers have focused on
for many years. The ability to allow the brake disc pads to move away  from the rotor and
thereby reduce friction is a known technology. This has been historically implemented by
designing a caliper and rotor system that allows the piston in the caliper to retract. However, if
the pads are allowed to move too far away from the rotor, the first pedal apply made by the
vehicle operator can feel spongy and have excessive travel.  This can lead to customer
dissatisfaction regarding braking performance and pedal feel. For this reason, in conventional
hydraulic-only brake systems, manufacturers are limited by how much they can allow the pads to
move away from the rotor.

5.2.8.4.2       Low DrasBrakes in the FRM

   The 2012-2016 final rule and Draft TAR estimated the effectiveness of low drag brakes to be
as much as 1 percent. NHTSA and EPA have slightly revised the effectiveness down to 0.8
percent based on the 2011 Ricardo study.

   In the  2012-2016 rule, the agencies estimated the DMC at $57 (2007$). This DMC becomes
$59 (2010$) for this analysis after adjusting to 2010 dollars. The agencies consider low drag
brake technology to be off the learning curve (i.e., the DMC does not change year-over-year) and
have applied a low complexity ICM of 1.24 through 2018 then 1.19 thereafter.

5.2.8.4.3       Developments since the FRM

   Recent developments in braking systems have allowed suppliers to provide brakes that have
the potential for zero drag. In this system the pad is allowed to move away from the rotor in
much the same way that is done in today's conventional brake systems,  but in a zero drag brake
system the pedal feel is separated from the hydraulics by a pedal simulator.  This system is very
similar to the brake systems that have been designed for hybrid and electric vehicles. In hybrid
and electric vehicles, some of the primary braking is done through the recuperation of kinetic
energy in the drive system. However, the pedal feel and the deceleration that the operator
experiences is  tuned to provide a braking experience that is equivalent to that of a conventional
hydraulic brake system. These "brake-by-wire" systems have highly tuned pedal simulators that
feel like typical hydraulic brakes and seamlessly transition to a conventional system as required
by conditions.  The application of a pedal simulator and brake-by-wire system is new to non-
electrified vehicle applications. By using this type of system vehicle manufacturers can allow
the brake pads to move farther away from the rotor and still  maintain the initial pedal feel and
deceleration associated with a conventional brake system.

   In addition, to reducing brake drag, the zero drag brake system also provides ancillary
benefits.  It allows for a faster brake apply and greater deceleration than is normally applied by
the average vehicle operator.  It also allows manufacturers to tune the braking for different
                                             5-206

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
customer preferences within the same vehicle.  This means that a manufacturer can provide a
"sport" mode which provides greater deceleration with less pedal displacement and a "normal"
mode which might be more appropriate for day-to-day driving. These electrically driven systems
also facilitate other brake features such as panic brake assist, automatic braking for crash
avoidance and could support future autonomous driving features.

   The zero drag brake system also eliminates the need for a brake booster. This saves both cost
and weight in the overall system.  Elimination of the conventional vacuum brake booster could
also improve the effectiveness of stop-start systems.  Typical stop-start systems need to restart
the engine if the brake pedal is cycled because the action drains the booster of stored vacuum.
Because the zero drag brake system provides braking assistance electrically, there is no need to
supplement lost vacuum  during an engine off event.

   Finally,  many of the engine technologies being considered to improve efficiency reduce
pumping losses through reduced throttle.  The reduction in throttle could result in supplemental
vacuum being required to operate a conventional brake system.  This is situation in many diesel
powered vehicles. Diesel engines run without a throttling  and often require supplemental
vacuum for brake boosting. By using a zero drag brake  system, manufacturers may realize the
elimination of brake drag as well as the ancillary benefits described above and avoid the need for
a supplemental vacuum pump.

   For the cost and effectiveness assumptions the agencies are adopting for the GHG Assessment
and CAFE  Assessment for  this Draft TAR analysis, see  Sections 5.3 and 5.4.

5.2.9   Air Conditioning Efficiency and Leakage Credits

   Air conditioning (A/C) is a virtually standard automotive  accessory, with over 95 percent of
new cars and light trucks sold in the United States being equipped with mobile air conditioning
(MAC) systems. This high penetration means that A/C systems have the potential to exert a
significant influence on the energy consumed by the  light duty vehicle fleet, as well as to GHG
emissions resulting from refrigerant leakage.

   The FRM allowed vehicle manufacturers to generate credits for improved A/C systems
toward complying with the CCh and fuel consumption fleetwide average standards.  In the EPA
program, manufacturers can generate credits for improved performance of both direct emissions
(refrigerant leakage) and indirect emissions (tailpipe emissions attributable to the energy
consumed by A/C). In both cases, a selection of "menu" credits in grams per mile are available
for qualifying technologies, with the magnitude of each  credit being estimated based on the
expected reduction in CCh  emissions resulting from the technology.  See 40 CFR 86.1868-12. In
the NHTSA program, manufacturers are allowed to generate fuel consumption improvement
values for purposes of CAFE compliance based on the use of A/C efficiency-improving
technologies. However,  manufacturers cannot  count reductions in A/C leakage toward their
CAFE calculations since these improvements do not affect fuel economy.

   Since the FRM, many manufacturers have generated  and banked credits through this program
and continue to do so today. In the FRM, the agencies estimated that significant penetration of
A/C technologies would  occur to gain these credits, and this  was reflected in the stringency of
the standards. See e.g. 77 FR at 62805/3.
                                             5-207

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   EPA and NHTSA projected that the 2017-2025 rule would result in significant improvements
in the efficiency of automotive air conditioning (A/C) systems.  Also, EPA projected that the
program would lead to significant reductions in GHGs from reduced A/C refrigerant leakage and
from industry adoption of lower global warming potential (GWP) refrigerants.  Additional
information that has become available, as well as changes in the overall regulatory environment
affecting the A/C technology developments in the light-duty vehicle industry, reinforces our
earlier conclusions that these technologies will continue to expand and play an increasing role in
overall vehicle GHG reductions and regulatory compliance.

5.2.9.1 A/C Efficiency Credits

5.2.9.1.1     Background on the A/C Efficiency Credit Program

   The 2012 FRM established two test procedures to determine eligibility for A/C efficiency
credits. The two test procedures are the idle test and the AC 17 test. These procedures were
assigned to different roles depending on the model year for which the test is conducted.

   For model years 2014 to 2016, there were three options for qualifying for A/C efficiency
credits:  1) running the A/C Idle Test, as described in the MYs 2012-2016 final rule, and
demonstrating compliance with the CCh and fuel consumption threshold requirements, 2)
running the A/C Idle Test and demonstrating compliance with engine displacement adjusted CCh
and fuel consumption threshold requirements, and 3) running the AC 17 test and reporting the test
results.

   For model years 2017-2019, the AC 17 test becomes the exclusive means manufacturers will
have to demonstrate eligibility for A/C efficiency credits, again by reporting the test results. By
reporting test results, manufacturers gain access to the credits on the menu based  on the design of
their AC system. In MYs 2020 and thereafter, however, the AC 17 test will be used not only to
demonstrate eligibility for efficiency credits, but also to partially quantify the amount of the
credit. AC 17 test results ("A" to "B" comparison) equal to or greater than the menu value will
allow manufacturers to claim the full menu value for the credit.  A test result less than the menu
value will limit the amount of credit to that demonstrated on the AC 17 test. In addition, for MYs
2017 and beyond, A/C fuel consumption improvement values will be available  for CAFE in
addition to efficiency credits being available for GHG compliance. These adjustments to the
utilization and design of the A/C test procedures were largely a  result of new data collected, as
well as the extensive technical comments submitted on the proposal.

5.2.9.1.2     Idle Test Procedure

   Starting in MY2014, manufacturers have been required to demonstrate the efficiency of a
vehicle's A/C system by running an A/C Idle  Test as a prerequisite to CCh credit  and fuel
consumption improvement value eligibility (the amount of credit determined separately by
means of the credit menu). If a vehicle met the emissions threshold of 14.9 grams per minute
(g/min) CCh or  lower on this test, a manufacturer was eligible to receive full credit (CAFE
improvement values) for efficiency-improving hardware or controls installed on that vehicle.
The vehicle would be able to receive A/C credits based on a menu of technologies specifying the
credit amount associated with each technology. For vehicles with a result between 14.9 g/min
and 21.3 g/min, a downward adjustment factor was applied to the eligible credit amount, with
vehicles testing higher than 21.3 g/min not being eligible to  receive credits.  The details of this
                                             5-208

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
idle test can be found in the MYs 2012-2016 final rule.  See 75 FR at 25426-27. This
methodology for accessing the credit menu based on the Idle Test results (and threshold
requirements) continued to apply for model years 2014-2016. The 2017-2025 FRM did not
make any fundamental changes to the previous rule. EPA did, however, provide an optional new
threshold requirement for MYs 2014-2016 reflecting the comments submitted on the idle test.

   Prior to the 2017-2025 FRM, manufacturers had the opportunity to run the Idle Test on a wide
variety of vehicles and discovered that even though there may be a small correlation between
engine displacement and the Idle Test result, the trend was important enough that small vehicles
had higher A/C idle emissions and were more inclined to fail to meet the threshold for the Idle
Test than were larger vehicles.  Specifically, vehicles with smaller displacement engines had a
higher Idle Test result than those with larger displacement engines, even within the same  vehicle
platform. This was causing some small vehicles with advanced A/C systems to fail the Idle Test.
The load placed on the engine by the A/C system did not seem to be consistent, and in certain
cases, larger vehicles perform better than smaller ones on the A/C idle test. These effects were
attributed in part to the fact that the  brake-specific fuel consumption (bsfc) of a smaller
displacement engine is generally lower at idle than that of a larger displacement engine, causing
larger engines to move from a less efficient region to a more efficient region when A/C is
operated at idle, while smaller engines enjoy less of this effect or may drop into a less efficient
region.  The 2017-2025 TSD presented additional analyses and adjustments to address these  and
similar difficulties with the A/C Idle Test.

   In the 2012 FRM, the agencies recognized the limitations of the Idle Test and provided for a
gradual phasing out of this test in favor of the AC 17 Test, to be described below. The primary
disadvantage of the Idle Test is that it does not capture the majority of the driving or ambient
conditions in the real world when the A/C is in operation, and thus may only  encourage the
technologies that improve idle performance under narrow temperature conditions. Another
limitation is that the narrow range of engine operating conditions present during the Idle Test
make it difficult to quantify the incremental improvement of a given technology to generate an
actual credit over non-idle operation (without a menu).

5.2.9.1.3     AC 17 Test Procedure

   In preparation for the 2017-2025 NPRM, the agencies sought to develop a more capable test
procedure that could more reliably generate an appropriate credit value based on an "A" to "B"
comparison, that is, a comparison of substantially similar vehicles in which one has the
technology and the other does not.  The result of this effort was the AC 17 Test Procedure, which
we believe is capable (in part) of detecting the effect of more efficient A/C components and
controls strategies during a transient drive cycle, rather than just idle.

   To develop this test, EPA initiated a study that engaged automotive manufacturers, USCAR,
component suppliers, SAE, and CARB.  This effort also explored the applicability and
appropriateness of a test method or procedure which combines the results of test-bench,
modeling/simulation, and chassis dynamometer testing into a quantitative metric for quantifying
A/C  system (fuel) efficiency. The goal of this exercise was the development of a reliable,
accurate, and verifiable assessment and testing method while also minimizing a manufacturer's
testing burden. For a complete description  of the AC 17 test, please refer to the 2017-2025 TSD.
                                             5-209

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   The agencies believe that the AC 17 test procedure more accurately reflects the impact that
A/C use (and in particular, efficiency-improving components and control strategies) has on
tailpipe CCh emissions and fuel consumption. In the FRM, EPA established this test to be
phased in starting in MY2014 as an option and in MYs 2017-2019 as the exclusive means of
determining eligibility for A/C efficiency credits (CAFE improvement values), and thereafter as
both an eligibility test and as a partial means of determining credit amount. That is, use of the
AC 17 test procedure to conduct A-to-B comparison tests becomes mandatory in 2020 as the
exclusive test means for accessing the A/C efficiency menu and quantifying the credits. If the
delta of the A-to-B test is greater than the value in the credit menu, the manufacturer receives the
menu value, otherwise the value is scaled. However, an engineering assessment can still be
conducted as an alternative to A-to-B testing to build the case for a specific credit value if, for
example, a baseline vehicle does not exist on which to base the A-to-B comparison. See 76 FR
74938, 74940.

5.2.9.1.4     Manufacturer Uptake of A/C Efficiency Credits since the 2012 FRM

   Many manufacturers have taken advantage of the A/C credit program to generate and bank
A/C efficiency credits, which have become an important contributor to industry compliance
plans.  As summarized in the EPA Manufacturer Performance Report for the 2014 model year468,
17 auto manufacturers included A/C efficiency credits as part of their compliance demonstration
in the 2014 model year. These amounted to more than 10 million Mg of credits, or about 25
percent of the total net credits reported. This is equivalent to about 3 grams per mile across the
2014 fleet. Including the 2012 and 2013 model years, A/C efficiency credits totaled over 24.4
million Mg.

   The A/C credit menu includes several A/C efficiency-improving technologies that were well
defined and had been quantified for effectiveness at the time of the FRM.  The vast majority of
A/C efficiency credits were claimed through this mechanism.

   The agencies expect that additional technologies for improving A/C efficiency that were not
anticipated at the time of the FRM may continue to emerge in the future. Although such
technologies will not be added to the design-based credit menu, these technologies will continue
to be eligible for credit under the off-cycle credit  program.

   An off-cycle credit application for this purpose should be supported by results of testing under
the AC 17 test protocol using an "A to B"  comparison, that is, a comparison of substantially
similar vehicles in which one has the technology and the other does not. Applications for A/C
efficiency credits made under the off-cycle credit program rather than the A/C credit program
will continue to be subject to the A/C efficiency credit cap.

   To date, the agencies have received one off-cycle credit application for an A/C efficiency
technology.  In December 2014, General Motors submitted an off-cycle credit application for the
Denso  SAS A/C compressor with variable crankcase suction valve technology, requesting an off-
cycle GHG credit of 1.1 grams CCh per mile. EPA evaluated the application and found that the
methodologies described therein were sound and  appropriate. Therefore, EPA approved the
credit application.
                                             5-210

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.2.9.1.5     Evaluation of the AC 17 Test Procedure

   Prior to the 2012 FRM, EPA collaborated with several OEMs to conduct independent testing
on a variety of vehicles and air conditioning technologies on the AC 17 test cycle. The purpose
of this test effort was to gain insight regarding the appropriateness of the AC 17 test for verifying
the reduction in CCh emissions which are expected from A/C technologies on the efficiency
credit menu.  Initially, six vehicles were tested, including three pairs of carlines with some
element of difference in their air conditioner systems.  The results of these tests were discussed
in the 2012 TSD, Section 5.1.3.7, beginning on page 5-44. This collaborative effort continued to
include a variety of additional vehicles tested by several OEMs at AC17-capable test facilities.469
This preliminary testing showed that the AC 17 test is capable of low test-to-test variability, and
is suitable for evaluating the relative efficiency improvement of A/C technologies, when
confounding factors are minimized.  In cases where comparison of the AC 17 results do not
directly demonstrate the effectiveness of a technology, the test results can still be useful within
an engineering analysis for justifying the test methodology to determine A/C CO2 credits.

   EPA also initiated a round-robin test program between facilities of several USCAR members
in an effort to determine the repeatability of the AC 17 test among various test facilities and to
identify potential sources of variability. A 2011 Ford Explorer was selected for these tests. Four
test sites were utilized, located at Ford, GM, Chrysler, and an EPA-contracted facility at
Daimler. Each facility had a full environmental chamber capable of fulfilling all requirements of
the test. Four tests were run at each facility, after which the vehicle was returned to Ford for
confirmation. Each test measured CO2 emissions with A/C  off and A/C on, to capture the
difference (delta) in CO2 emissions, which represents the GHG effect of A/C usage.

   Figure 5.77 through Figure 5.79  compares the results of each test at each test site. Although
some variability was observed between test sites, consistency within a given site  was good,
suggesting that the AC 17 test procedure is able to capture the difference in CO2 emissions
between A/C on and A/C off.

   Several sources of variation were identified by analysis of these results. Variations in solar
load may have resulted from variations in sensor location and soak start time. Temperature
control was also a potential issue. Although most labs could maintain temperature within the
required tolerance of the test procedure, humidity was more difficult to maintain  for the long
duration of the test. Overcorrecting may  occur, but can be improved by optimizing sensor
location to better represent ambient conditions.  The complexity and length of the test can lead to
an increased potential for voided tests, and may require more frequent calibration of the test cell
equipment. Although this test program was not fully described in materials accompanying the
FRM, many of the issues  observed during this testing were addressed in the final form of the
rule.
                                             5-211

-------
                       Technology Cost, Effectiveness, and Lead-Time Assessment
                             CO2 Summary, A/C ON
    o


450 -
400 -

300 -

OK,r\ -

I

1
*Bag1,SC03 "Bag 2, HWFET • Average |

* » » »447 * *
• ^ .374 ••• *378
• • • • 364 • *
• * • • 353

. • . .291 ' ' ' '29°
" • " "2S1 . • .-.276
05 V 10 « *, <0 0) 0 «-  fti fl? ^ ^ ^ •$ a? w v w ^ ^ J?
_^J?J?J? C?C?C?Cj' ^^^*
//fl ///«?

»
-* — 452
• "369




*r c
^


Figure 5.77 Variability of AC17 Round Robin Testing on 2011 Ford Explorer, A/C On
                             CO2 Summary, A/C OFF
        500 -,
        450 •
     1  400
     2
     o
        350 -
        300 -
                         »Bag3, SC03  "Bag 4, HWFET  • Average  |
                    *414
               * *  *  *402  *  « *  >397
                                               387
               •••347
                     279
                                >342
                                 281
                                         • .331
                                                                   «344
                                                              ._.«_. 276
                                                        I 262

////   ////'

                                                -
                                        S'sSr    ^•^'5'^r    OOO

                                         #   ////
                                             cfcfcfcf
Figure 5.78 Variability of AC17 Round Robin Testing on 2011 Ford Explorer, A/C Off
                                       5-212

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                             CO2 Summary, DELTA {A/C ON - A/C OFF)
60 •
50
40 •
30
20 •
10 •
n -
»SC03
I


<• 44
I ^


i 16
111

• HWFET • Average


T 51
I" J«
I f
f28 I»

• 15 J14

Figure 5.79 Variability of AC17 Round Robin Testing on 2011 Ford Explorer, Delta between A/C on and Off
   Although these tests demonstrated that the AC 17 test was able to resolve the difference
between A/C on and A/C off, they did not address its ability to resolve smaller differences, such
as the effect of an individual technology in an A to B test.  As the size of an effect diminishes,
the difficulty of resolving it against a much larger baseline value becomes more challenging.
With the baseline CCh g/mi value for most vehicles being in the hundreds, and the effect of a
single A/C technology possibly in the low single digits, test-to-test variation must be very small
to reliably detect the effect. As the AC 17 A-to-B test becomes a requirement beginning in
MY2020, this issue is being examined closely by the industry and EPA.

   Since the 2012 FRM, USCAR members have conducted an ongoing test program to assess the
ability of the AC 17 test to resolve the GHG impact of individual A/C efficiency technologies in
an A to B test, and thereby function in the role assigned to it in the FRM as a means for
quantifying and qualifying for A/C  credits. EPA has followed this effort by direct coordination
with member OEMs and by participating in meetings of the SAE Interior Climate Control
Committee.

   At this time, the USCAR test program is not yet complete, and results are not yet conclusive.
Preliminary results are encouraging, although uncertainties continue to exist. In general, OEMs
have expressed concern about several issues:

   a) Difficulty of obtaining or constructing old-technology vehicles, particularly those from
      earlier model years, on which to base A-to-B comparisons.
   b) Factors such as test-to-test variability and the small magnitude of the effect being
      measured result in the need  for multiple tests to be conducted to yield a statistically
      reliable result, possibly increasing the test burden beyond what the agencies anticipated.
   c) Members suggested that bench testing and engineering analysis may be preferable to A-
      to-B  AC 17 testing as a means of qualifying for menu credits, if these difficulties are not
      resolved in further testing.
                                             5-213

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Overall, members have expressed greater confidence in their ability to conduct A-to-B
comparisons of software-related technologies (for example, default to recirculated air) than for
hardware-based technologies (for example, compressor design changes) because the former can
be implemented by relatively simple changes to software. A-to-B comparisons of hardware
technologies would be more difficult because of the requirement to produce test specimens
configured with and without the technology.

   In January 2016, EPA received additional comment and analysis from several USCAR
members regarding their most recent experience with AC 17 testing. In this interaction, many of
the issues discussed above were further outlined. Manufacturers have continued to experience a
significant number of voided tests and are continuing to work to identify the sources of such
events, which are commonly associated with long tests that demand careful environmental
control. Test-to-test variation is sometimes seen to exceed the magnitude of the credit value that
is the subject of the test. Although averaging of the results of multiple tests has shown some
success at establishing a reliable outcome, concerns were expressed about the resulting test
burden, due to the length of each test, the control requirements, and the limited availability of the
required specialized test cells. The availability of base vehicles without the technology being
assessed in an A-to-B comparison was  also echoed as a concern. Manufacturers suggested that
the use of prior year models may be infeasible when several intervening model years are
involved, due to the confounding effect of other technologies introduced to the vehicle during
that time.  This was expressed as being particularly true for the problem of assessing hardware-
based technologies, which may require building of prototype  installations that may  require
additional engineering resources to develop. Within individual test efforts, consistency of results
was good in some tests but exhibited inconsistencies in others, of which the manufacturers had
not yet achieved a full understanding but continue to study. Issues such as the complexity of
modern climate control systems and the presence of confounding factors such as powertrain
differences were cited as possible factors.

   An application for off-cycle credits submitted by General Motors in December 2014 provides
an additional source of information on the results of AC 17 A-to-B testing, which was used to
support the application.  GM cited several issues relating to the use of the AC 17 test procedure to
identify the CCh benefit claimed in the application:

   a) GM pointed out that the AC 17 A-to-B test was enabled  by coincidental availability of a
valid baseline compressor (a variable compressor without the variable crankcase suction valve
technology) in the Holden Commodore and that this compressor coincidentally could be easily
bolted into the Cadillac ATS.  GM reiterated that this is an uncommon situation and not
representative of future expectations;

   b) GM stated that this hardware obstacle "prevents ready testing of the benefits of the SAS
compressor on other GM models on which it has been implemented;"

   c) There were some difficulties with torque and pressure measurement which was cited as
example of "control issues that may be expected to arise when attempting to do this type of
baseline technology testing for hardware on a vehicle that was never actually designed and
optimized to use that hardware."
                                             5-214

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Despite these difficulties, GM found that the AC 17 test procedure was able to resolve a 1.3
g/mile CCh improvement, which was in good agreement with the 1.1 g/mile suggested by bench
testing.  However, because test-to-test variability was greater for the AC 17 tests than for the
bench tests, GM chose to request the 1.1 g/mile shown by the bench tests, which GM regarded as
more precise.

   As an alternative to the A-to-B testing requirement, the 2012 FRM provided manufacturers
the option to qualify for A/C credits through bench testing supported by engineering analysis.
This option continues to be available after the 2020 AC 17 requirement goes into effect. EPA has
encouraged, and continues to encourage, the use of bench test results and engineering analysis to
support applications for A/C efficiency credits.

   In 2016, USCAR members initiated a Cooperative Research Program (CRP) through the
Society of Automotive Engineers (SAE) to develop bench testing standards for the four hardware
technologies in the credit menu (blower motor control, internal heat exchanger, improved
evaporators and condensers, and oil separator). The specific standards under development are
listed in Table 5.27. The intent of the program is to streamline the process of conducting bench
testing and engineering analysis in support of an application for A/C credits by creating uniform
standards for bench testing and for establishing the expected GHG impact of the technology in a
vehicle application. The AC 17 test may continue to have a supporting role in some of these
standards. EPA continues to monitor the development of these standards by coordinating with
the CRP as well as participating in the applicable SAE standards development committees.
 Table 5.27 Hardware Bench Testing Standards under Development by SAE Cooperative Research Program
Number
J2765
J3094
J3109
J3112
Title
Procedure for Measuring System COP of a Mobile Air Conditioning
System on a Test Bench
Internal Heat Exchanger (IHX) Measurement Standard
HVAC PWM Blower Controller Efficiency Measurement
A/C Compressor Oil Separator Effectiveness Test Standard
Status
Published
Work in Progress
Work in Progress
Work in Progress
5.2.9.1.6
Conclusions and Future Work
   The agencies have evaluated and considered the results of AC 17 testing presented by
stakeholders. This data suggests that the AC17 test is capable of measuring the difference in
CCh emissions between A/C on and A/C off, and in some cases, is also capable of resolving
differences in CCh emissions resulting from hardware and software differences (A-to-B).
However, in many of the A-to-B comparisons, test-to-test variability and the small magnitude of
the effect to be measured has led to the need for multiple repeated tests to identify the effect with
statistical significance, potentially adding to the test burden required to obtain A/C credits.

   At this time, the results of USCAR testing of the AC 17 test procedure is not yet complete, and
not yet conclusive. The agencies await the availability of additional data in order to more fully
evaluate the role of the AC 17 test procedure under the GHG program.  EPA also anticipates that
the ongoing test program by USCAR members will result in development of a guidance letter
recommending best practices for conducting AC 17 testing.
                                             5-215

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   EPA will continue coordination with USCAR to obtain any additional data regarding the
effectiveness of the AC17 test in discerning A/C efficiency differences in A-to-B comparisons.
Sources of this data may include additional A-to-B testing by USCAR, as well as any future
applications for A/C off-cycle credits that are supported by the results of AC 17 testing. EPA
will also continue to coordinate with manufacturers through meetings with industry stakeholders,
participation in the SAE interior climate control committees, coordination with the SAE CRP,
and any other applicable venues.

   The agencies invite additional comment regarding stakeholder experience with the AC 17 test
procedure and its ability to resolve GHG emissions differences by A-to-B testing.

   Although it is anticipated that new A/C technologies may have emerged since the 2012 FRM
that are not represented in the credit menu, the agencies do not have plans to add additional items
to the credit menu nor change the values assigned to those that are currently in the menu.
Manufacturers may continue to apply for credits for new technologies through the off cycle
credit program.

5.2.9.2 A/C Leakage Reduction and Alternative Refrigerant Substitution

5.2.9.2.1      Leakase

   As we observed in the rule, manufacturers have developed a number of technologies for
reducing the leakage of refrigerant to the atmosphere. These include fittings, seals, heat
exchanger/compressor designs, and hoses. Vehicle manufacturers consider low-leak
technologies to be among the most cost-effective approaches to improving overall vehicle GHG
emission performance.

   Table 5.28 shows two metrics of the continued industry-wide progress toward durable, low-
leak systems. One trend is the annual increase in the generation of leakage credits already
apparent in the early years of the program as manufacturers have taken advantage of leakage-
reduction incentives. More on this trend, as well as a breakdown of leakage credits by
manufacturer, are found in EPA's Manufacturer Performance Report for the 2014 Model Year470
Specifically, 15 manufacturers reported A/C leakage credits in the 2014 model year, amounting
to more than 16.5 million Megagrams (Mg) of credits, or more than 40 percent of the total net
credits reported for the model year. This equates to GHG reductions of about 5 grams per mile
across the 2014 vehicle fleet. The table also shows the trend toward more leak-proof A/C
systems in terms of refrigerant leakage scores across the industry,  as indicated by the average
industry-wide A/C system leakage scores that the State of Minnesota requires automakers to
report (using the SAE J-2727 method).471
   Table 5.28 Trends in Fleetwide Mobile Air Conditioner Leakage Credits and Average Leakage Rates

Credits: (Million Megagrams)
MN SAE J-2727 Leakage Rate (g/yr)
2009
6.2
15.1
2010
8.3
14.7
2011
8.9
14.6
2012
11.1
14.5
2013
13.2
13.9
2014
16.6
13.0
2015
Not Yet Reported
12.1
5.2.9.2.2
Low-GWP Refrigerants
   In support of the LD GHG rules, EPA projected that the industry would fully transition to
lower-GWP refrigerants between Model Year (MY) 2017 and MY2021, beginning with 20
                                             5-216

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
percent transition in MY2017, to be followed by a 20 percent increase in substitution in each
subsequent model year, completing the transition by MY2021  (77 FR 62779, 62778, 62805). Put
another way, the stringency of the MY2021 and later light duty GHG standards is predicated on
100 percent substitution of refrigerants with lower GWPs than HFC-134a. On July 20, 2015,
EPA published a final rule under the Significant New Alternatives Policy (SNAP) program that
changes the listing status of HFC-134a to unacceptable for use in A/C systems of newly-
manufactured LD motor vehicles beginning in MY2021, except where permitted for some export
vehicles through MY2025 (80 FR 42870)MM EPA's decision to take this action was based on
the availability of other substitutes that pose less overall risk to human health and the
environment, when used in accordance with required use conditions. Thus all new LD vehicles
sold in the United States will have transitioned to an alternative, lower-GWP refrigerant by
MY2021.

   The July 20, 2015 SNAP final rule has no effect on how manufacturers may choose to
generate  and use air conditioning leakage credits under the LD GHG standards. As stated in that
final rule, " [n]othing in this final rule changes the regulations establishing the availability of air
conditioning refrigerant credits under the GHG standards for MY2017-2025, found at 40 CFR
86.1865-12 and 1867-12.  The stringency of the standards remains unchanged....
[Manufacturers may still generate and utilize credits for substitution of HFC-134a through the
2025 model year." EPA also there noted that the SNAP rule was not in conflict with the
Supplemental Notice of Intent (76 FR 48758, August 9, 2011) that described plans for EPA and
NHTSA's joint proposal for model years 2017-2025,  since EPA's GHG program continues to
provide the level of air conditioning credits available to manufacturers as specified in that
Notice:  "[T]he Supplemental Notice of Intent states that '(rn)anufacturers will be able to earn
credits for improvements in air conditioning . . . systems, both for efficiency improvements  . .  .
and for leakage or alternative, lower-GWP refrigerants used (reduces [HFC] emissions).' 76 FR
at 48761. These credits remain available under the light-duty program at the level specified in
the Supplemental Notice of Intent, and using the same demonstration mechanisms set forth in
that Notice." 80 FR 42896-97.

   EPA has listed three lower-GWP refrigerants as acceptable, subject to use conditions (listed at
40 CFR Part 82, Subpart G), for use in newly-manufactured LD vehicles: HFO-1234yf, HFC-
152a, and carbon dioxide (CCh or R-744). Manufacturers are currently manufacturing LD
vehicles using HFO-1234yf, and they are actively developing LD vehicles using CCh472 and
considering the use of HFC-152a in a secondary loop A/C systems.473

   EPA expects that vehicle manufacturers will use HFO-1234yf for the vast majority of
vehicles. As discussed in the EPA Manufacturer Performance Report referenced above, the use
of HFO-1234yf expanded considerably in the 2014 model year, from two manufacturers and
42,384 vehicles in the 2013 model year, to five manufacturers and 628,347 vehicles in the 2014
model year. Although this is a large increase, it is still a relatively small fraction (less than 5
percent) of the total 2014 model year production. This trend reinforces EPA's projection that the
industry will have transitioned 20 percent of the fleet by MY2017, as discussed above. Fiat
MM HFC-134a will remain listed as acceptable subject to narrowed use limits through MY2025 for use in newly
  manufactured LD vehicles destined for export, where reasonable efforts have been made to ascertain that other
  alternatives are not technically feasible because of lack of infrastructure for servicing with alternative refrigerants
  in the destination country. (40 CFR Part 82, Subpart G, Appendix B.


                                             5-217

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Chrysler accounted for 86 percent of these vehicles, introducing HFO-1234yf across a number of
models, including the 300, Challenger, Charger, Cherokee, Dart, and Ram 1500 trucks. Jaguar
Land Rover achieved the greatest penetration within their fleet, using HFO-1234yf in
approximately 80 percent of Jaguar Land Rover vehicles produced in the 2014 model year.

   Finally, regarding supply of alternative refrigerants, the July 2015 SNAP final rule stated that
EPA "considered the supply of the alternative refrigerants in determining when alternatives
would be available.  At the time the light-duty GHG rule was promulgated, there was a concern
about the potential supply of HFO-1234yf.  Some commenters indicated that supply is still a
concern, while others, including two producers of HFO-1234yf, commented that there will be
sufficient supply. Moreover, some automotive manufacturers are developing systems that can
safely use other substitutes, including CCh, for which there is not a supply concern for the
refrigerant. If some global light-duty motor vehicle manufacturers use CCh or another
acceptable alternative, additional volumes of HFO-1234yf that would have been used by those
manufacturers will then become available.  Based on all of the information before the agency,
EPA believes production plans for the refrigerants are in place to make available sufficient
supply no later than MY2021 to meet current and projected demand  domestically as well  as
abroad, including, but not limited to, the EU." (80 FR 42891; July 20, 2015)

5.2.9.2.3      Conclusions

   As described in this section, there is strong evidence that auto manufacturers are continuing to
improve the leak-tightness of their A/C systems. In addition, many manufacturers are
transit!oning to the use of low-GWP alternative refrigerants in a number of vehicle models.  We
believe that the current trends among automakers toward the use of alternative refrigerants to
comply with the LD vehicle GHG standards, EPA's change in listing status of HFC-134a  to
"unacceptable" by MY2021, and the parallel increase in the supply of the leading alternative
refrigerant ensure that our earlier projections that a complete transition to alternative refrigerants
by MY2021 will in fact become reality.

   The MY2017-2025 LD GHG rule also encourages manufacturers to continue to use low-
leakage technologies even when using alternative refrigerants.  Although some leakage may still
occasionally occur, the low GWPs of the new refrigerants, as compared to that of HFC-134a,
considerably reduce concerns about refrigerant leakage from a climate perspective.

5.2.10  Off-cycle Technology Credits

5.2.10.1       Off-cycle Credits Program

5.2.10.1.1     Off-cycle Credits Program Overview

   EPA and NHTSA provide an opportunity for credits for off-cycle technologies.  EPA initially
included off-cycle technology credits in the MY2012-2016 rule and revised the program in the
MY2017-2025 rule.474 NHTSA adopted equivalent off-cycle credits for MYs 2017 and later in
the MY2017-2025 rule.475 "Off-cycle" emission reductions and fuel consumption improvements
can be achieved by employing off-cycle technologies that result in real-world benefits, but where
that benefit is not adequately captured on the test procedures used by manufacturers to
demonstrate compliance with and fuel economy emission standards.
                                             5-218

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The intent of the off-cycle provisions is to provide an incentive for CCh and fuel consumption
reducing off-cycle technologies that would otherwise not be developed because they do not offer
a significant 2-cycle benefit. EPA and NHTSA limited the eligibility to technologies whose
benefits are not adequately captured on the 2-cycle test and NHTSA added further limitations on
technologies that might otherwise be incentivized through its safety regulations.476  The
preamble to the final rule provided a detailed discussion of eligibility for off-cycle credits.477
Technologies that are integral or inherent to the basic vehicle design including engine,
transmission, mass reduction, passive aerodynamics, and base tires are not eligible. Any
technology that was included in the agencies' standard-setting analysis also may not generate
off-cycle credits (with the exception of active aerodynamics and engine stop-start systems).  EPA
established this approach believing that the use of 2-cycle technologies would be driven by the
standards and no additional credits would be necessary or appropriate. This approach also limits
the program to off-cycle technologies that could be clearly identified as add-on technologies
more conducive to A/B testing that would be able to demonstrate the benefits of the technology.

   There are three pathways by which a manufacturer may generate off-cycle CCh credits. The
first is a predetermined list of credit values for specific off-cycle technologies that may be used
beginning in MY2014.478 This pathway allows manufacturers to use conservative credit values
established in the MY2017-2025  final rule for a wide range of technologies, with minimal data
submittal or testing requirements.  In cases where additional laboratory testing can demonstrate
emission benefits, a second pathway allows manufacturers to use a broader array of emission
tests (known as "5-cycle" testing because the methodology uses five different testing procedures)
to demonstrate and justify off-cycle CCh credits.479  The additional emission tests allow emission
benefits to be demonstrated over some elements of real-world driving not captured by the GHG
compliance tests, including high speeds, rapid accelerations, and cold temperatures.  Credits
determined according to this methodology do not undergo additional public review.  The third
and last pathway allows manufacturers to seek EPA approval to use an alternative methodology
for determining the off-cycle CCh credits.480 This option is only available if the benefit of the
technology cannot be adequately  demonstrated using the 5-cycle methodology. Manufacturers
may also use this option for model years prior to 2014 to demonstrate off-cycle CCh reductions
for technologies that are on the predetermined list, or to demonstrate reductions that exceed those
available via use of the predetermined list.  The manufacturer must also demonstrate that the off-
cycle technology is effective for the full useful life of the vehicle. Unless the manufacturer
demonstrates that the technology is not subject to in-use deterioration, the manufacturer must
account for the deterioration in their analysis.

   The pre-defined list of technologies and associated car and light truck credits is shown in the
tables below.481 The regulations include a definition of each technology that the technology
must meet in order to be eligible for the menu credit.482  Manufacturers are not  required to
submit any other emissions data or information beyond meeting the definition and useful life
requirements to use the pre-defined credit value. Credits based on the pre-defined list are subject
to an annual manufacturer fleet-wide cap of 10 g/mile.
                                              5-219

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                    Table 5.29 Off-cycle Technologies for Cars and Light Trucks
Technology
High Efficiency Exterior Lighting (at 100W)
Waste Heat Recovery (at 100W; scalable)
Solar Roof Panels (for 75 W, battery charging only)
Solar Roof Panels (for 75 W, active cabin ventilation
plus battery charging)
Active Aerodynamic Improvements (scalable)
Engine Idle Start-Stop w/ heater circulation system
Engine Idle Start-Stop without/ heater circulation
system
Active Transmission Warm-Up
Active Engine Warm-Up
Solar/Thermal Control
Credit for Cars
g/mi (gallons/mi)
1.0 (0.000113)
0.7 (0.000079)
3.3 (0.000372)
2.5 (0.000282)
0.6 (0.000068)
2.5 (0.000282)
1.5 (0.000169)
1.5 (0.000169)
1.5 (0.000169)
Up to 3.0 (0.000338)
Credit for Light Trucks
g/mi (gallons/mi)
1.0 (0.000113)
0.7 (0.000079)
3.3 (0.000372)
2.5 (0.000282)
1.0 (0.000113)
4.4 (0.000496)
2.9 (0.000327)
3.2 (0.000361)
3.2 (0.000361)
Up to 4.3 (0.000484)
 Table 5.30  Off-cycle Technologies and Credits for Solar/Thermal Control Technologies for Cars and Light
                                          Trucks
Thermal Control
Technology
Glass or Glazing
Active Seat Ventilation
Solar Reflective Paint
Passive Cabin Ventilation
Active Cabin Ventilation
Credit (gCOVmi)
Car
Up to 2.9 (0.000326)
1.0 (0.000113)
0.4 (0.00005)
1.7 (0.000191)
2.1 (0.000236)
Truck
Up to 3.9 (0.000439)
1.3 (0.000146)
0.5 (0.00006)
2.3 (0.000259)
2.8 (0.000315)
   The two other pathways available to generate off-cycle credits require additional data. The 5-
cycle testing pathway requires 5-cycle testing with and without the off-cycle technology to
determine the off-cycle benefit of the technology. The final pathway, often referred to as the
public process includes a public comment period and is available for technologies that cannot be
demonstrated on the 5-cycle test. Manufacturers must develop a methodology for demonstrating
the benefit of the off-cycle technology and the methodology is made available for public
comment prior to an EPA determination whether or not to allow the use of the methodology to
generate credits. The data needed for this demonstration may be extensive, especially in cases
where the effectiveness of the technology is dependent on driver response or interaction with the
technology.  As discussed below, all three methods have been used successfully by
manufacturers to generate off-cycle credits.
5.2.10.2
Use of Off-cycle Technologies to Date
A wide array of off-cycle technologies were used by manufacturers in MY2014 to generate off-
cycle GHG credits using the pre-defined menu.483  Table 5.31 below shows the percent of each
manufacturers' production volume using each of the menu technologies reported to EPA for
MY2014 by the manufacturer.  Table 5.32 shows the g/mile benefit that each manufacturer
reported across its fleet from each off-cycle technology. Like the preceding table, Table 5.32
                                             5-220

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
provides the mix of technologies used in MY2014 across the manufacturers and the extent to
which each technology benefits each manufacturer's fleet. Fuel consumption improvement off-
cycle credits are not available in the CAFE program until MY2017 and therefore only GHG off-
cycle credits have been generated by manufacturers thus far.
    Table 5.31 Percent of 2014 Model Year Vehicle Production Volume with Credits from the Menu, by
                               Manufacturer & Technology (%)
Manufacturer

BMW
Fiat Chrysler
Ford
GM
Honda
Hyundai
Jaguar Land
Rover
Kia
Mercedes
Nissan
Subaru
Toyota
Fleet Total
Active
Aerodynamics
0)
D
_C
5
0.0
16.4
38.4
6.7
0.0
2.1
0.0
1.8
0.0
4.6
2.0
0.0
9.8
Ride height
adjustment
0.0
3.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.2
0.0
Thermal Control Technologies
Passive cabin
ventilation
0.0
99.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
11.4
15.0
Active cabin
ventilation
85.1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.1
Active seat
ventilation
3.9
6.1
2.6
1.2
1.3
0.8
5.0
0.9
2.2
1.8
0.0
2.5
2.3
Glass or
glazing
2.9
99.3
97.2
52.3
0.0
84.4
98.1
76.1
3.9
0.0
0.0
52.9
50.7
Engine & Transmission
Warmup
Solar reflective
surface
coating
0.0
1.3
12.5
15.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
25.5
8.7
Active engine
warmup
78.5
58.0
9.6
0.0
0.0
0.0
0.0
0.0
0.0
19.5
0.0
9.2
14.2
Active
transmission
warmup
0.0
11.7
16.2
0.0
58.5
16.7
0.0
22.7
0.0
55.7
0.0
53.8
23.2
Other
Engine idle
stop-start
0.0
0.0
3.4
6.7
0.0
0.0
93.0
0.6
65.3
0.9
0.0
12.5
5.5
High efficiency
exterior lights
98.1
73.3
52.9
28.2
28.2
36.2
100.0
59.5
35.7
50.1
0.0
44.5
43.0
Solar panel(s)
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.2
0.0
0.0
0.0
                                             5-221

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
    Table 5.32 Off-Cycle Technology Credits from the Menu, by Manufacturer and Technology (g/mi)
Manufacturer

BMW
Fiat Chrysler
Ford
GM
Honda
Hyundai
Jaguar Land
Rover
Kia
Mercedes
Nissan
Subaru
Toyota
Fleet Total
Active
Aerodynamics
0)
D
_C
5
-
0.1
0.3
0.0
-
0.0
-
0.0
-
0.0
0.0
-
0.1
Ride height
adjustment
-
0.0
-
-
-
-
-
-
-
-
-
0.0
0.0
Thermal Control Technologies
Passive cabin
ventilation
-
2.0
-
-
-
-
-
-
-
-
-
0.2
0.3
Active cabin
ventilation
2.0


-


-

-
-

-
0.0
Active seat
ventilation
0.0
0.0
0.2
0.2
0.0
0.1
0.8
0.2
0.1
0.1
-
0.2
0.1
Glass or glazing
0.0
1.8
1.3
0.7
-
0.3
1.2
0.3
0.1
-
-
0.6
0.7
Engine & Transmission
Warmup
Solar reflective
surface coating
-
0.0
0.1
0.1


-

-
-

0.1
0.0
Active engine
warmup
1.6
1.7
0.2
-


-

-
0.3

0.1
0.3
Active
transmission
warmup
-
0.4
0.4
-
1.3
0.3
-
0.3
-
1.2

1.1
0.5
Other
Engine idle
stop-start
-
0.0
0.1
0.1


2.5
0.0
1.7
0.0

0.2
0.1
High efficiency
exterior lights
0.3
0.2
0.1
0.1
0.1
0.0
0.5
0.1
0.4
0.2
-
0.1
0.1
Solar panel(s)
-
-
-
-
-
-
-
-
-
0.0
-
-
0.0
0.0" indicates that the manufacturer did implement that technology, but that the overall penetration rate was not high enough to round to 0.1
grams/mile, whereas a dash indicates no use of a given technology by a manufacturer.
   The credits shown above are based on the pre-defined credit list.  Thus far, GM is the only
manufacturer to have been granted off-cycle credits based on 5-cycle testing.  These credits are
for an off-cycle technology used on certain GM gasoline-electric hybrid vehicles.  The
technology is an auxiliary electric pump, which keeps engine coolant circulating in cold weather
while the vehicle is stopped and the engine is off, thus allowing the engine stop-start system to
be active more frequently in cold weather.

   The third pathway allows manufacturers to seek approval to use an alternative methodology
for determining the off-cycle technology CCh credits. Several  manufacturers have petitioned for
and been granted use of an alternative methodology for generating credits. In the fall of 2013,
Mercedes requested off-cycle credits for the following off-cycle technologies in use or planned
for implementation in the 2012-2016 model years: stop-start systems, high-efficiency lighting,
infrared glass glazing, and active seat ventilation. EPA approved methodologies for Mercedes to
determine these off-cycle credits in September of 2014.484 Subsequently, FCA, Ford, and GM
requested off-cycle credits under this pathway.  FCA and Ford submitted applications for off-
cycle credits from high efficiency  exterior lighting, solar reflective glass/glazing,  solar reflective
paint, and active seat ventilation.  Ford's application also demonstrated off-cycle benefits from
active aerodynamic improvements (grill shutters), active transmission warm-up, active engine
warm-up technologies, and engine idle stop-start. GM's application described the real-world
benefits of an air conditioning compressor with variable crankcase suction valve technology.
EPA approved the  credits for FCA, Ford, and GM in September of 2015.485 Although EPA has
                                              5-222

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
granted the use of alternative methodologies, manufacturers have yet to report credits to EPA
based on those alternative methodologies.

   As discussed above, the vast majority of credits in MY2014 were generated using the pre-
defined menu. Even though the program has been in place for only a few model years,  and
MY2014 is the first year the pre-defined list may be used, the level of credits achieved has
already been significant for some manufacturers.  FCA and Jaguar Land Rover generated the
most off-cycle credits on a fleet-wide basis, reporting credits equivalent to about 6 g/mile and 5
g/mile, respectively.NN  Several other manufacturers report fleet-wide credits in the range of
about 1 to 4 g/mile. The fleet total across all manufacturers was equivalent to about 2 g/mile for
MY2014. The agencies expect that as manufacturers continue to expand their use of off-cycle
technologies, the fleet-wide impacts will continue to grow with some manufacturers potentially
approaching the 10 g/mile fleet-wide cap applicable to credits that are based on the pre-defined
list.

5.3    GHG Technology Assessment

5.3.1   Fundamental Assumptions

5.3.1.1 Technology Time Frame and Measurement Scale for Effectiveness and Cost

   The effectiveness and cost associated with applying a technology will depend on the starting
technologies from which improvements are measured. For example, two vehicles that start with
different technologies will likely have different cost and effectiveness associated with adopting
the same combination of technologies.  The importance of clearly specifying the point of
comparison for cost and effectiveness estimates was highlighted in the 2015 NAS committee's
finding "that understanding the base or null vehicle, the order of technology application, and the
interactions among technologies is critical for assessing the costs and effectiveness for meeting
the standards."

   As long as the point of comparison is maintained consistently throughout the analysis for both
the baseline and future fleets, the decision of where to place an origin along the scale of cost and
effectiveness is inconsequential. For EPA's technology assessment, the origin is defined to
coincide with a "null technology package," which represents a technology floor such that all
technology packages considered in this assessment will have equal or greater effectiveness,
consistent with the FRM approach.  While other choices would have been  equally valid, this
definition of a "null package" has the practical benefit of avoiding technology packages with
negative effectiveness values, while also allowing for a direct comparison  of effectiveness
assumptions with the FRM.
  ' The credits are reported to EPA by manufacturers in Megagrams. EPA has estimated a g/mile equivalent.
                                             5-223

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
      Effectiveness
      Cost
             "Null" Technology
                   Package
                 gC02/minu
f
                                   _A_
                                          Technology
Package 1 ^
          Eff2.1(%)=  l-[l-Eff2H
                   null^°
Technology 1
 Package 2 1
 gC02/mi1    gC02/mi
                                                                    decreasing emissions
                                           Eff2-nlll(%)
"Null" Technology
Parkaoo ^ ^
Technology
v Package 1
* 1 1 ° 1 I
S° increasing cost COSti

Technology
NPackage2
1
Cost2

'
      Figure 5.80 The "Null Technology Package" and Measurement Scale for Cost and Effectiveness

   When technologies can be specifically identified for individual vehicle models, it is possible
to estimate cost and effectiveness values specifically for those models. To the extent possible
with the available information, EPA has attempted to consider this. This is the case, for
example, with mass reduction and improvements in aerodynamics and tire rolling resistance,
where for this assessment EPA has uniquely characterized the various levels of those
technologies for individual models based on available road load data. For other technologies, the
information that is broadly available across the entire fleet is not detailed enough to distinguish
differences that arise to different implementations of the technologies.

5.3.1.2 Performance Assumptions

   When determining cost and effectiveness values for specific technologies, it is important to
compare the technologies on a consistent basis, so that the relative cost-effectiveness of the
technologies can be fairly compared. The National Academy of Science states in their 2011
report: "Estimating the cost of decreasing fuel consumption requires one to carefully specify a
basis for comparison. The committee considers that to the extent possible, fuel consumption cost
comparisons should be made at equivalent acceleration performance and equivalent vehicle
size."486 This is because "objective comparisons of the cost-effectiveness of different
technologies for reducing [fuel consumption] can be made only when vehicle performance
remains equivalent."487 The National Academy of Science engaged the University of Michigan
for their 2015 report to perform a set full vehicle simulations. As a ground rule, "Each engine
configuration was modeled to maintain, as closely as possible, the torque curve of the baseline
naturally aspirated engine so that equal performance, as measured by 0-60 mph acceleration
time, would be maintained"488 The agencies agree that it is appropriate to  objectively compare
technology costs and effectiveness, that maintaining constant vehicle performance is the
appropriate way to achieve that goal, and that the NAS's recommendation of "equivalent
acceleration performance" is appropriate. Thus, the costs and effectiveness presented in this
document are based on the application of technology packages while holding the underlying
acceleration performance constant.

   In most  cases, equivalent acceleration performance is achieved by "engine downsizing":
reducing the size (and thus the output power/torque) of the engine in advanced vehicle  packages
until a series of performance metrics are maintained within a reasonable range of the target value
                                              5-224

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
similar to the methodology used in the FRM. A smaller engine will typically be more efficient at
the same speed and torque than a larger engine (as pumping losses are reduced), so this
methodology properly accounts for effectiveness that could be used for acceleration performance
as fuel consumption reduction, thus allowing an objective and fair comparison of technologies.
Our process maintains performance neutrality. As recommended by the NAS (2011), EPA is
working under the premise that technology cost assessments should be made under the
assumption of equivalent performance. As such, the ALPHA modeling runs generate
effectiveness values which maintain a set of acceleration metrics within a reasonably small
window.

   EPA recognizes that manufacturers have many vehicle attribute and manufacturing
constraints. Manufacturers will make many  product planning decisions and the final products
will have engine displacement which represent the OE's decision in its product plans. As a
modeling convenience, when calculating effectiveness, EPA assumes the appropriate component
sizing to maintain performance. Even if our model produces a greater variation in technology
packages than exists today (for example, by producing two levels of tire rolling resistance on a
vehicle platform compared to just one today), this does not require that manufactures actually
produce a greater variety of component sizes than exist currently in order for our overall results
to be valid. In actual vehicle design, manufacturers will design discretely sized components, and
for each vehicle choose the available size closest to the optimal for the given load and
performance requirements. For example, in  some cases, the chosen engine will be slightly
smaller than optimal (and thus lower fuel consumption), and in some cases the chosen engine
will be slightly larger than optimal (and thus higher fuel consumption). The same assumption is
applied to drivetrain, suspension, chassis components, etc. For example, brake rotors may be
sized in 15mm diameter increments, and manufacturers will apply the size that most closely
matches the performance and load requirements of that application. Just as the manufacturers are
doing today, EPA expects that they will average these product decisions across their entire fleet.
In our analysis, on average, the actual fleet of vehicles will use the appropriate component size,
and CCh emissions and performance of the fleet will average out, with no significant net change
compared to the original analysis with unconstrained component sizes.

   In gathering information on technology effectiveness, the agencies relied on a wide variety of
sources. These sources provided information on the costs and effectiveness of various
technologies, but not all comparisons were done on a rigorously performance-neutral basis.
Thus, it was often necessary to recalculate the effectiveness of a particular technology when the
original comparison was done without the assumption of equivalent performance. For example,
the 2011 NAS report, in discussing continuously variable valve lift (CVVL)489 cites Energy and
Environmental Analysis, Inc.,490 which "estimates  a 6.5 to 8.3 percent reduction in fuel
consumption at constant engine size and 8.1 to 10.1 percent with an engine downsize to maintain
constant performance."

   When EPA modeled effectiveness of specific technologies of their combinations, it was
careful to maintain a minimum deviation of acceleration performance from the baseline vehicle.
As the NAS notes, "truly equal performance involves nearly equal values for a large number of
measures such as acceleration (e.g., 0-60 mph, 30-45 mph, 40-70 mph, etc.), launch (e.g., 0-30
mph), grade-ability (steepness of slopes that can be climbed without transmission downshifting),
maximum towing capability, and others."491 However, they furthermore state that "in the usage
                                             5-225

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
herein, equal performance means 0-60 mph times within 5 percent. This measure was chosen
because it is generally available for all vehicles."

   In vehicle simulations conducted in support of the FRM using MSC EASY5 or using the
Response Surface Model (RSM) data analysis tool, EPA defined overall equivalent performance
such that 0-30 mph and 0-60 mph acceleration times were kept within a performance window
defined as no more than 5 percent slower or 10 percent faster than a baseline vehicle. Additional
performance criteria were then cross-checked to ensure no significant degradation in vehicle
utility. For example, simulation of grade-ability  at 60 mph with a 5,000 Ib. trailer (both in top
gear and in any transmission gear) was used to cross check maintaining the utility of full size
trucks. Within the FRM analysis, the 0-30 mph and 0-60 mph performance window criteria were
found to be sufficient to maintain equivalence with other indicators of vehicle performance and
utility, including trailer grade-ability.

   In vehicle simulation modeling in ALPHA performed since the FRM, EPA investigated using
additional performance criteria to define an overall performance metric. Four acceleration
performance metrics were chosen: 0-60 time, Vi  mile time, 30-50 passing time, and 50-70
passing time. These metrics were chosen to give a reasonably broad set of acceleration metrics
that would be sensitive  enough to represent true  acceleration performance, but not so sensitive
that minor changes in vehicle parameters would  significantly change the final metric. For each
vehicle class, a baseline configuration was chosen, the vehicle package was run over the
performance cycle, and the times for each performance metric were extracted. These four metrics
were summed for the baseline vehicle. For each  vehicle technology package based on the same
vehicle class, a nominal engine size was determined based on the estimated performance effect
of the technologies included in the package. The same performance cycle  was run and the sum of
the four metrics compared to the baseline sum. If the sum was not within three percent (tighter
than the 5 percent band suggested by NAS), the  size of the engine was adjusted and the
performance cycle rerun until an equivalent acceleration performance was attained. When the
sum was within three percent, the CCh emissions modeling over the standard drive cycles was
performed using the engine  size determined.

   In general, the criteria used to define equivalent performance for the FRM analysis and for
analyses using the ALPHA model since the FRM have resulted in comparable changes in engine
displacement when comparable levels of vehicle technology are applied within the EPA
"standard car" class for effectiveness analyses. For the Draft TAR, EPA has continued to rely on
the performance criteria from the FRM analysis  within its  analyses of technology effectiveness,
however, the addition of Vi mile time, 30-50 passing time,  and 50-70 passing time performance
metrics are still under consideration for the Proposed and Final Determinations.

   For the purpose of specification and costing of plug-in vehicles (BEVs  and PHEVs, or
collectively, PEVs), acceleration performance was maintained by a different method to account
for differences in the way power is developed by electric motors and conventional engines.
Originally, in the 2012 FRM analysis, PEVs of a given vehicle class (small car, large car, etc.)
were assigned an electric motor power rating (kW) that would preserve the same engine-power-
to-weight ratio that was observed in conventional vehicles of that class. This method assumed
that the all-electric acceleration of an electrified  vehicle relates to the power rating of the electric
motor in the same way that the engine-powered  acceleration of a conventional vehicle relates to
the power rating of the engine. However, electric motors differ from combustion engines in that
                                             5-226

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
they deliver maximum torque at the lowest end of their speed range, while combustion engines
must develop significant speed to deliver a comparable torque. This can allow an electric motor
to deliver higher acceleration at low speeds than a comparable engine of the same nominal power
rating, and potentially higher acceleration overall. An analysis of 2012 FRM motor power
assumptions suggested that the modeled PEV motors may have been significantly more powerful
than necessary for the intended acceleration performance. For this Draft TAR analysis, EPA
derived an empirical equation relating PEV power-to-weight ratio to reported 0-60 acceleration
time based on an informal study of MY2012-2016 BEVs and PHEVs. A target 0-60 time was
selected for each PEV configuration comparable to that of conventional vehicles, and the motor
power assigned based on this equation. The PEV motor sizing methodology is described in more
detail in Section 5.3.4.3.7.1.

5.3.1.3 Fuels

  Fuel specifications for the gasoline and diesel fuels used for demonstration  of compliance
with light-duty vehicle GHG and CAFE standards are contained within the Title 40, Part 86 of
the U.S. Code of Federal Regulations. Tabulated values are reproduced here for reference
purposes in Table 5.33 and Table 5.34 for gasoline and diesel, respectively. Analyses of the
effectiveness  of powertrain technologies over the regulatory drive cycles used fuel properties
conforming to these specifications.
       Table  5.33 Test Fuel Specifications for Gasoline without Ethanol (from 40 CFR §86.113-04)
Item
Research octane, Minimum2
Octane sensitivity2
Distillation Range ( °F):
Evaporated initial boiling point3
10% evaporated
50% evaporated
90% evaporated
Evaporated final boiling point
Hydrocarbon composition (vol %):
Olefins
Aromatics
Saturates
Lead, g/gallon (g/liter), Maximum
Phosphorous, g/gallon (g/liter), Maximum
Total sulfur, wt. %4
Dry Vapor Pressure Equivalent (DVPE), psi (kPa)5
Regular
93
7.5

75-95
120-135
200-230
300-325
415 Maximum

10% Maximum
35% Maximum
Remainder
0.050 (0.013)
0.005 (0.0013)
0.0015-0.008
8.7-9.2 (60.0-63.4)
Reference Procedure1
ASTM D2699; ASTM D2700
ASTM D2699; ASTM D2700

ASTM D86





ASTM D1319


ASTM D3237
ASTM D3231
ASTM D2622
ASTM D5191

                  Table 5.34 Petroleum Diesel Test Fuel (from 40 CFR §86.113-94)
Property
(i) Cetane Number
(ii) Cetane Index
(iii) Distillation range:
(A) IBP
(B) 10 pet. Point
Unit





Type 2-D
40-50
40-50

340-400 (171.1-204.4)
400-460 (204.4-237.8)
Reference
Procedure1
ASTM D613
ASTM D976



                                             5-227

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
(C) 50 pet. Point
(D) 90 pet. Point
(E)EP
(iv) Gravity
(v) Total sulfur
(vi) Hydrocarbon composition: Aromatics,
minimum (Remainder shall be paraffins,
naphthenes, and olefins)
(vii) Flashpoint, min
(viii) Viscosity
°F ( °C)


"API
ppm
pet
°F ( °C)
centistokes
470-540 (243.3-282.2)
560-630 (293.3-332.2)
610-690 (321.1-365.6)
32-37
7-15
27
130 (54.4)
2.0-3.2
STM D86


ASTM D4052
ASTM D2622
ASTM D5186
ASTM D93
ASTM D445
1 ASTM procedures are incorporated by reference in §86.1
   EPA's analysis of effectiveness with gasoline fueled engines did not include analysis of
effectiveness using Tier 3 certification gasoline (E10, 87 AKI) although protection for operation
in-use on 87 AKI E10 gasoline was included in the analysis of engine technologies considered
both within the original FRM and within the Draft TAR. A correction factor (or R-factor) for
application to future vehicles certified using Tier 3 gasoline that will allow correction of CCh
emissions in a manner that accounts for differences between Tier 2 and Tier 3 certification fuels
is currently under regulatory development.

5.3.1.4 Vehicle Classification

   The vehicle classes for which EPA has estimated effectiveness are consistent with the FRM
and six vehicle classes developed for the lumped parameter model. Table 5.35 presents the
mapping of lumped parameter model vehicle classes into model-specific vehicles to help the
reader understand how the vehicle classes are used for modeling.
                               Table 5.35 EPA Vehicle Classes
EPA Vehicle Class
Subcompact/ Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Lump Parameter
Classification
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Example
Fiesta
Focus
Yaris
Fusion
Taurus
Camry
300
Mustang
Escape
Rav4
Tacoma
Explorer
4Runner
Caravan
F150
Tundra
OMEGA Model
Vehicle Type
1
2,3,4
5,6
7,13
8, 9, 10, 14, 15, 18
11, 12, 16, 17, 19
                                              5-228

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.2   Approach for Determining Technology Costs

   Section 5.3.2.1 presents sources and approaches to estimating direct manufacturing costs.
Section 5.3.2.1.4 presents the methods used to address indirect costs in this analysis. Section
5.3.2.1.4 presents the learning effects applied throughout this analysis. In Section 5.3.2.1 the
individual technology costs are presented including: the direct manufacturing costs (DMC), their
indirect costs (1C) and their total costs (TC, TC=DMC+IC).

5.3.2.1 Direct Manufacturing Costs

   Estimates of direct manufacturing costs (DMC) come from many sources: detailed paper
studies and analyses; published reports; supplier and OEM provided data which would generally
be confidential business information (CBI); etc. The agencies consider the best source of DMC
estimates to be those from tear-down studies. The 2015 NAS report503 agreed with this
assessment and encouraged the agencies to make use of tear-down studies where available
stating, "the use of teardown studies has improved the  agencies' estimates of costs" (NAS pp. S-
3) and "Updated cost estimates using teardown cost studies of recently introduced spark-ignition
engine technologies, including all vehicle integration costs, should be developed to support the
mid-term review," (NAS pp. S-4) and "EPA and NHTSA should conduct a teardown cost study
of a modern diesel engine with the latest technologies to provide an up-to-date estimate of diesel
engine costs." (NAS pp. S-5) The summary below provides our sources for many of the
technologies considered in this analysis.

5.3.2.1.1      Costs from Tear-down Studies

   As in the 2017-2025 FRM, there are a number of technologies in this analysis that have been
costed using the rigorous tear-down method described  in this section. As a general matter, the
agencies believe, and the NAS agrees,492 that the most rigorous method to derive technology cost
estimates is to conduct studies involving tear-down and analysis of actual vehicle components. A
"tear-down" involves breaking down a technology into its fundamental parts and manufacturing
processes by completely disassembling actual vehicles and vehicle subsystems and precisely
determining what is required for its production.  The result of the tear-down is a "bill of
materials" for each and every part of the vehicle or vehicle subsystem. This tear-down method of
costing technologies is often used by manufacturers to benchmark their products against
competitive products. Historically, vehicle and vehicle component tear-down has not been done
on a large scale by researchers and regulators due to the expense required for such studies. Many
technology cost studies in the literature are based on information collected from OEMs,
suppliers, or "experts" in the industry and are thus non-reproducible and non-transparent. In
contrast, EPA sponsored teardown studies are completely transparent and include a tremendous
amount of data and analyses to improve accuracy. While tear-down studies are highly accurate at
costing technologies for the year in which the study is intended, their accuracy, like that of all
cost projections, may diminish over time as costs are extrapolated further into the future because
of uncertainties in predicting commodities (and raw material) prices,  labor rates, and
manufacturing practices. The projected costs may be higher or lower than predicted.

   Since the early development of the 2012-2016 rule,  EPA has contracted with FEV, Inc. to
conduct tear-down cost studies for a number of key technologies evaluated by the agencies in
assessing the feasibility of future GHG and CAFE standards. The analysis methodology included
procedures to scale the tear-down results to smaller and larger vehicles, and also to different
                                             5-229

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
technology configurations. FEV's methodology was documented in a report published as part of
the MY2012-2016 rulemaking process.493

   Additional cost studies were completed and used in support of the 2017-2025 FRM. These
include vehicle tear downs of a Ford Fusion power-split hybrid and a conventional Ford Fusion
(the latter served as a baseline vehicle for comparison). In addition to providing power-split HEV
costs, the results for individual components in these vehicles were subsequently used to develop
costs for the P2 hybrid used in the following MY2017-2025 FRM.00 This approach to costing P2
hybrids was undertaken because P2 HEVs were not yet in volume production at the time of
hardware procurement for tear-down. Finally, an automotive lithium-polymer battery was torn
down to provide supplemental battery costing information to that associated with the NiMH
battery in the Fusion, because automakers were moving to Li-ion battery technologies due to the
higher energy and power density of these batteries. As noted, this HEV cost work, including the
extension of results to P2 HEVs,  has been documented in a report prepared by FEV and was used
in support of the 2017-2025 FRM. Because of the  complexity and comprehensive scope of this
HEV analysis, EPA commissioned a separate peer review focused exclusively on the new tear
down costs developed for the HEV analysis. Reviewer comments generally supported FEV's
methodology and results, while including a number of suggestions for improvement, many of
which were subsequently incorporated into FEV's analysis and EPA final report.  The peer
review comments and responses were made available in the rulemaking docket.

   Additional cost studies were completed and used in support of the Draft TAR. These include
an 14 mild hybrid system (2013 Malibu with eAssist) replacing a conventional 14 engine, an 14
diesel engine replacing a conventional V6 gasoline engine, and a turbocharged 14 engine
replacing a V6 gasoline engine. This latest turbocharged study replaces the original study as this
technology has evolved significantly over the past few years. Peer reviews have been completed
for the mild hybrid and diesel cost studies.

   Over the course of this contract between EPA and FEV, FEV performed teardown-based
studies on the technologies listed below. These completed studies provide a thorough evaluation
of the new technologies'  costs relative to their baseline (or replaced) technologies.

      1)  Stoichiometric gasoline direct injection (SGDI) and turbocharging with engine
          downsizing (T-DS) on a DOHC (dual overhead cam) 14 engine, replacing a
          conventional DOHC 14 engine
      2)  SGDI and T-DS on a  SOHC (single overhead cam) on a V6 engine, replacing a
          conventional 3-valve/cylinder SOHC V8 engine
      3)  SGDI and T-DS on a DOHC 14 engine, replacing a DOHC V6 engine
      4)  6-speed automatic transmission (AT), replacing a 5-speed AT
      5)  6-speed wet dual clutch transmission (DCT) replacing a 6-speed AT.
      6)  8-speed AT replacing a 6-speed AT
      7)  8-speed DCT  replacing a 6-speed DCT
      8)  Power-split hybrid (Ford Fusion with 14 engine) compared to a conventional vehicle
          (Ford Fusion with V6). The results from this tear-down were extended to address P2
          hybrids. In addition, costs from individual components in this tear-down study were
          used by the agencies in developing cost estimates for PHEVs and EVs.
 ' Describe what P2 hybrid means.
                                            5-230

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
       9)  Fiat Multi-Air engine technology. (Although results from this cost study are included
          in the rulemaking docket, they were not used in the 2017-2025 rulemaking's technical
          analyses because the technology is under patent and therefore not considered in the
          2017-2025 timeframe).

   In addition, FEV and EPA extrapolated the engine downsizing costs for the following
scenarios that were based on the above study cases:

       •  Downsizing a SOHC 2 valve/cylinder V8 engine to a DOHC V6
       •  Downsizing a DOHC V8 to a DOHC V6
       •  Downsizing a SOHC V6 engine to a DOHC 4 cylinder engine
       •  Downsizing a DOHC 4 cylinder engine to a DOHC 3 cylinder engine

   Since the 2017-2025 FRM, the following teardown studies have been completed:

       1)  Mild hybrid with stop-start technology (Chevrolet Malibu 14 engine with eAssist),
          replacing a conventional 14 engine.
       2)  14 diesel engine, replacing a conventional V6 gasoline engine.
       3)  New iteration of SGDI and T-DS on a DOHC 14 engine, replacing a DOHC V6
          engine.
   FEV has also updated the cost estimates for all of the teardown studies.

   Additional teardown work has been done in the area of mass reduction technologies. This
work is highlighted in greater detail in Section 5.2 of this report.

   The agencies have relied on the findings of FEV for estimating the cost of the technologies
covered by the tear-down studies. However, note that FEV based their costs on the assumption
that these technologies would be mature when produced in large volumes (450,000 units or more
for each component or subsystem). If manufacturers are not able to employ the technology at the
volumes assumed in the FEV analysis with fully learned costs, then the costs for each of these
technologies would be expected to be higher. There is also the potential for stranded capital if
technologies are introduced too rapidly for some indirect costs to be fully recovered. While the
agencies consider the FEV tear-down analysis results to be generally valid for the 2022-2025
timeframe for fully mature, high sales volumes,  FEV performed supplemental analysis to
consider potential stranded capital costs,  and we have included these in our primary analyses of
program costs.

5.3.2.1.2     Electrified Vehicle Battery Costs

   As in the 2012 FRM, EPA has used the BatPaC model494 to estimate battery costs for
electrified vehicles. Developed by Argonne National Laboratory (ANL) for the Vehicle
Technologies Program of the U.S. Department of Energy (DOE) Office of Energy Efficiency and
Renewable Energy, the BatPaC model allows users to estimate the manufacturing cost of battery
packs for various types of electrified powertrains given battery power and energy requirements
as well as other design parameters.

   In the 2015 NAS report (p. 4-25), the NAS committee endorsed the importance of the use of a
bottom-up battery cost model such as BatPaC, further finding that "the battery cost estimates
used by the agencies are broadly accurate" (Finding 4.4, p. 4-43). Since the publication of the
                                            5-231

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
FRM, BatPaC has been further refined and updated with new costs for some cathode chemistries
and cell components, improved thermal management calculations, and improved accounting for
plant overhead costs. Further changes were released in late 2015 and include additional
chemistries, updated material costs, improved calculation of electrode thickness limits, and
improved estimation of cost and energy requirements of certain manufacturing steps and material
production processes.495 EPA has used the most recent version of BatPaC to revise the battery
cost projections used in the GHG assessment of this Draft TAR analysis.

   In the 2012 FRM, the agencies developed costs and effectiveness values for the mild and P2
FIEV configurations, two different all-electric mileage  ranges for PHEVs (20 and 40 in-use
miles) and three different mileage ranges for BEVs (75, 100 and 150 in-use miles). In this Draft
TAR analysis, EPA has developed cost and effectiveness values for a new 48-Volt mild hybrid,
and has changed the 150-mile BEV configuration to a 200-mile configuration. Additional
updates to the inputs and methodology applied to electrified vehicles are described in Section
5.3.4.

5.3.2.7.3     Specific PMC Changes since the 2012 FRM

   EPA looked at all the latest public data and information, carefully reviewed all the NAS
estimates, the latest teardown studies, and in the end determined that teardown studies remain the
most robust source of cost estimates. This analysis uses updated technology costs from teardown
studies conducted since the FRM including mild hybrid (high voltage) and mild hybrid (48V)
which is based in large part on the mild hybrid high voltage teardown. EPA has updated costs
from prior teardowns (largely the transmission teardowns) based on updated studies conducted
by FEV to those prior teardowns. Remaining costs for technologies such as valve timing and lift,
friction reduction, etc., have been updated to 2013 dollars since all costs in this analysis are in
2013 dollars.  Lastly, EPA has updated battery and non-battery costs for electrified vehicles based
on a newer version of the ANL BatPaC model. Key battery pack design parameters such as
usable capacity and cell sizes have been reviewed and revised where appropriate to reflect trends
in industry practice that have been observed since the FRM. Additionally, EPA has added new
technologies not used in the FRM, specifically a 48-Volt mild hybrid, a more capable naturally
aspirated Atkinson cycle engine with a high compression ratio, a Miller cycle engine and
electrified vehicles with different ranges. For the more capable Atkinson cycle engine, costs
reported by NAS have been used as technology cost inputs.

5.3.2.1.4     Approach to Cost Reduction through Manufacturer Learning

   For some of the technologies considered in this analysis, manufacturer learning effects would
be expected to play a role in the  actual end costs. The "learning curve" or "experience curve"
describes the  reduction in unit production costs as a function of accumulated production volume.
In theory, the cost behavior it describes applies to cumulative production volume measured at the
level of an individual manufacturer, although it is often assumed—as both agencies have done in
past regulatory analyses—to apply at the industry-wide level, particularly in industries that
utilize many common technologies and component supply sources. The agencies believe there
are indeed many factors that cause costs to decrease over time. Research in the costs of
manufacturing has consistently shown that, as manufacturers gain experience in production, they
are able to apply innovations to simplify machining and assembly operations, use lower cost
materials, and reduce the number or complexity of component parts. All of these factors allow
manufacturers to lower the per-unit cost of production  (i.e., the manufacturing learning curve).
                                             5-232

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   NAS recommended that the agencies "continue to conduct and review empirical evidence for
the cost reductions that occur in the automobile industry with volume, especially for large-
volume technologies that will be relied on to meet the CAFE/GHG standards." (NAS pp. 7-23)
EPA has conducted such a review under contract to ICF looking at learning in mobile source
industries. The goal of the effort was to provide an updated assessment on learning and its
existence in manufacturing industries. An extensive literature review was conducted and the
most applicable and appropriate studies were chosen with the help of a subject matter expert
(SME) that is one of the leading experts in this area.pp EPA hoped that the study would provide
clear learning rates that could be applied in various mobile source manufacturing industries
rather than the more general learning rates used in the past.  That study was completed in
September of 2015. A peer review was initiated and completed, but the subsequent final report,
which would include responses to the peer review, was not completed in time for inclusion in the
docket supporting this Draft TAR.

   In the contracted  study, ICF performed this literature review and analysis of learning in the
mobile source sector with the assistance of a Subject Matter Expert (Dr. Linda Argote of
Carnegie Mellon University).  The draft report, Cost Reduction through Learning in
Manufacturing Industries and in the Manufacture of Mobile Sources, was subsequently peer-
reviewed by three well-known experts in the field  of learning (Marvin Lieberman, Ph.D.,
University of California, Los Angeles (UCLA) Anderson School of Management; Natarajan
Balasubramanian, Ph.D., Whitman School of Management, Syracuse University; and Chad
Syverson,  Ph.D., University of Chicago Booth School of Business). The peer review was carried
out for EPA by RTI International based on EPA Science Policy Council Peer Review Handbook,
4th Edition, and was completed in May 2016.

   The study consists of two parts: a literature review and an estimate of a mobile source
progress ratio. A total of 53 studies on learning were examined, with 20 of these selected for
detailed review (the  other 33 received a more cursory review and are not discussed in detail in
the report). Five of these studies were used as the  basis to estimate the progress ratio for the
mobile source sector. On the basis  of these studies, the SME noted: "The mean learning rate is
estimated to be -0.245, with a standard error of 0.0039.  Thus, the lower bound for a 95 percent
confidence interval for the learning rate is -0.253; the upper bound is -0.238.  These estimates
translate into a mean progress ratio of 84.3 percent. The confidence interval around this number
ranges from 83.9 percent to 84.8 percent, suggesting that one can be reasonably confident that
the progress ratio falls in this interval. Thus, the best estimate of the progress ratio in mobile
source industries is 84 percent." This is the value that EPA has used in this Draft TAR.

   As a result, the learning curve recommended for use by the report has slightly lower learning
rates than those EPA has used in the past. Past EPA studies have used a learning rate based  on a
curve that resulted in a 20 percent cost reduction for each doubling of volume; the recommended
rate results in cost reductions of 15 percent. As such, EPA has updated learning rates to be
consistent with the recommendation of the report.  The curve used in this analysis is:
                                          — axt+i
                                   y^ _ ™Z>


Where:
 ' The SME was Dr. Linda Argote of Carnegie Mellon University.
                                             5-233

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                    yt+l = Costs required to produce a unit at time t+1
                    a = Costs required to produce the first unit
                    xt+i = Cumulative number of units produced through period t+1
                    b = A parameter measuring the rate at which unit costs change as
       cumulative output increases; i.e., the learning rate
   For this analysis, EPA has used this equation to estimate the learning effects and have
generated the learning curves shown below. How these learning curves were actually generated
using the above curve is described in a memorandum contained in the docket.496 In general, the
new learning factors were generated in a way to provide similar results to past analyses.
However, because the new rate is lower, there are subtle differences especially in years further
from the "base" year (i.e., the year where the learning factor is 1.0). The docket memorandum
makes this clearer by providing the new factors alongside the factors used in the 2012 FRM for
comparison.

   Learning effects are applied to most but not all technologies because some of the expected
technologies are already used rather widely in the industry and, presumably, learning impacts
have already occurred. Learning effects on the steep-portion of the learning curve was applied
for only a handful of technologies that are considered to be new or emerging technologies. Most
technologies have been considered to be more established given their current use in the fleet and,
hence, learning effects on the flat portion of the learning curve have been applied. The learning
factor curve applied to each technology are summarized in Table 5.36 with the actual year-by-
year factors  for each corresponding curve shown in Table 5.37.
                                             5-234

-------
                     Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.36 Learning Effect Algorithms Applied to Technologies Used in this Analysis
Technology
Aero, active
Aero, passive
Atkinson, level 1
Atkinson, level 2
Cam configuration changes
V6OHVtoV6DOHC
V6SOHCtoV6DOHC
VSOHVtoVSDOHC
VSSOHCtoVSDOHC
V8SOHC3VtoV8DOHC
Charger, in-home, EV
Charger, in-home, PHEV20
Charger, in-home, PHEV40
Charger, in-home, labor
Cylinder deactivation
Direct injection, stoichiometric, gasoline
Diesel, advanced (TierS)
Diesel, lean NOx trap
Diesel, selective catalytic reduction
Downsizing, associated with turbocharging
!4DOHCtol3DOHC
!4DOHCtol4DOHC
V6OHVtol4DOHC
V6SOHCtol4DOHC
V6DOHCtol4DOHC
V8OHVtoV6DOHC
V8SOHCtoV6DOHC
V8SOHC3VtoV6DOHC
Engine friction reduction, level 1
Engine friction reduction, level 2
EGR, cooled
Electric power steering
EV75, battery pack
EV100, battery pack
EV200, battery pack
EV75, non-battery items
EV100, non-battery items
EV200, non-battery items
HEV, Mild, battery pack
HEV, Mild, non-battery items
HEV, Strong, battery pack
HEV, Strong, non-battery items
HEV, Plug-in, battery pack
HEV, Plug-in, non-battery items
Improved accessories, level 1
Improved accessories, level 2
Low drag brakes
Lower rolling resistance tires, level 1
Learning Factor "Curve"3
24
24
24
24

28
23
28
23
23
26
26
26
1
24
23
23
23
23

23
23
28
23
23
28
23
23
1
1
23
24
26
26
26
28
28
28
31
23
31
23
26
23
24
24
1
1
                                    5-235

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
Lower rolling resistance tires, level 2
Lube, engine changes to accommodate low friction lubes
Mass reduction <15%
Mass reduction >=15%
Secondary axle disconnect
Stop-start
Turbo, 18-21 bar
Turbo, 24 bar
Turbo, Miller-cycle
TRX11/12
TRX21/22
32
1
30
30
24
25
23
23
23
23
23
              Note:
              a See table below.
   The actual year-by-year factors for the numbered curves shown in Table 5.37.

      Table 5.37 Year-by-year Learning Curve Factors for the Learning Curves Used in this Analysis
Curve
1
22
23
24
25
26
27
28
29
30
31
32
2012
1.000
1.37
1.00
1.09
2.03
3.05
1.00
1.13
1.17
1.29
3.18
1.74
2013
1.000
1.33
0.98
1.06
1.62
2.44
0.91
1.09
1.13
1.24
2.54
1.61
2014
1.000
1.29
0.96
1.03
1.28
2.11
0.84
1.06
1.09
1.20
2.03
1.51
2015
1.000
1.25
0.94
1.00
1.00
1.89
0.80
1.03
1.06
1.17
1.62
1.43
2016
1.000
1.21
0.92
0.98
0.91
1.74
0.76
1.00
1.03
1.13
1.28
1.36
2017
1.000
1.18
0.91
0.96
0.84
1.61
0.74
0.98
1.00
1.09
1.00
1.30
2018
1.000
1.15
0.89
0.94
0.80
1.51
0.71
0.96
0.98
1.06
0.91
1.25
2019
1.000
1.13
0.88
0.92
0.76
1.43
0.69
0.94
0.96
1.03
0.84
1.20
2020
1.000
1.11
0.87
0.91
0.74
1.36
0.67
0.92
0.94
1.00
0.80
1.16
2021
1.000
1.08
0.85
0.89
0.71
1.30
0.66
0.91
0.92
0.98
0.76
1.12
2022
1.000
1.06
0.84
0.88
0.69
1.25
0.64
0.89
0.91
0.96
0.74
1.09
2023
1.000
1.04
0.83
0.87
0.67
1.20
0.63
0.88
0.89
0.94
0.71
1.06
2024
1.000
1.02
0.82
0.85
0.66
1.16
0.62
0.87
0.88
0.92
0.69
1.04
2025
1.000
1.00
0.82
0.84
0.64
1.12
0.61
0.85
0.87
0.91
0.67
1.01
   Importantly, where the factors shown in Table 5.37 equal "1.00" represents the year for which
any particular technology's cost is based. Thus, if curve 1 is applied to a technology - such as in
the case of low friction lubes - it assumes no additional learning takes place over time. In the
case of stop-start technology,  curve 25 is applied. In this case, the cost estimate used for stop-
start is considered a MY2015  cost. Therefore, its learning factor equals 1.00 in 2015 and then
decreases going forward to represent lower costs due to learning effects. Its learning factors are
greater than 1.00 in years before 2015 to represent "reverse" learning, i.e., higher costs than our
2015 estimate since production volumes have, presumably, not yet reached the point where our
cost estimate can be considered  valid. Not all of the learning curve factors follow this rule using
the updated curve approach used in this Draft TAR. Also of interest is that only curves 25 (stop-
start), 26 (EV & PHEV batteries) and 31 (mild and strong HEV batteries) show any steeper
learning beyond the 2017-2020 timeframe, and even those curves show less than 5 percent year-
over-year cost reductions beyond 2020. In other words, most curves are well into the flatter
portion of the learning curve,  and even those that are not are well beyond the steep learning that
occurs at the early stages of learning, by the timeframe considered in this Draft TAR.
                                              5-236

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Because of the nature of full electric and plug-in electric vehicle battery pack development,
the industry is arguably early in the learning-by-doing phase for the types of batteries considered.
Our approach, consistent with that used in the FRM, has been to develop a direct manufacturing
cost based on sales of 450,000 units. EPA has considered that to be a valid MY2025 cost (i.e.,
the cost is based in 2025). With that as the MY2025 cost, the costs are considered as understood
today and a best fit learning curve is projected between the costs in those near-term and long-
term years. This is described in more detail in the docket memorandum mentioned earlier.497
Note that the 450,000 unit sales is considered a valid MY2025 volume for batteries because that
volume is meant to represent volumes at a given production line (a battery supplier production
line,  not an OEM vehicle production line) and takes into consideration worldwide demand for
automotive and other mobile source battery packs not just U.S. directed automotive battery
packs.

   Note that the effects of learning on individual technology costs can be seen in the cost tables
presented in Section  5.3.4, below. For each technology, the direct manufacturing costs for the
years 2017 through 2025 are shown. The changes shown in the direct manufacturing costs from
year-to-year reflect the cost changes due to learning effects.

5.3.2.2 Indirect Costs

5.3.2.2.1     Methodologies for Determining Indirect Costs

   To produce a unit of output, vehicle manufacturers incur direct and indirect costs. Direct costs
include cost of materials and labor costs. Indirect costs are all the costs associated with
producing the unit of output that are not direct costs - for example, they may be related to
production (such as research and development [R&D]), corporate operations (such as salaries,
pensions, and health  care costs for corporate staff), or selling (such as transportation, dealer
support, and marketing). Indirect costs are generally recovered by allocating a share of the costs
to each unit of good sold. Although it is possible to account for direct costs allocated to each unit
of good sold, it  is more challenging to account for indirect costs allocated to a unit of goods sold.
To make a cost  analysis process more feasible, markup factors, which relate total indirect costs
to total direct costs, have been developed. These factors are often referred to as retail price
equivalent (RPE) multipliers.

   Cost analysts and  regulatory agencies (including both EPA and NHTSA) have frequently used
these multipliers to predict the resultant impact on costs associated with manufacturers'
responses to regulatory requirements. The best approach, if it were possible, to determining the
impact of changes in direct manufacturing costs on a manufacturer's indirect costs would be to
actually estimate the  cost impact on each indirect cost element. However, doing this within the
constraints of an agency's time or budget is not always feasible, or the technical, financial, and
accounting information to carry out such an analysis may simply be unavailable.

   RPE multipliers provide, at an aggregate level, the relative shares of revenues (Revenue =
Direct Costs + Indirect Costs + Net Income) to direct manufacturing costs. Using RPE
multipliers implicitly assumes that incremental changes in direct manufacturing costs produce
common incremental changes in all indirect cost contributors as well as net income. However, a
concern in using the RPE multiplier in cost analysis for new technologies added in response to
regulatory requirements is that the indirect costs of vehicle modifications are not likely to be the
same for different technologies. For example, less complex technologies could require fewer
                                              5-237

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
R&D efforts or less warranty coverage than more complex technologies. In addition, some
simple technological adjustments may, for example, have no effect on the number of corporate
personnel and the indirect costs attributable to those personnel. The use of RPEs, with their
assumption that all technologies have the same proportion of indirect costs, is likely to
overestimate the costs of less complex technologies and underestimate the costs of more
complex technologies.

   To address this concern, modified multipliers have been developed by EPA, working with a
contractor, for use in rulemakings.498 These multipliers are referred to as indirect cost multipliers
(or ICMs). In contrast to RPE multipliers, ICMs assign unique incremental changes to each
indirect cost contributor as well as net income.

       ICM = (direct cost + adjusted indirect cost)/(direct cost)

   Developing the ICMs from the RPE multipliers requires developing adjustment factors based
on the complexity of the technology and the time frame under consideration: the less complex a
technology, the lower its ICM, and the longer  the time frame for applying the technology, the
lower the ICM. This methodology was used in the cost estimation for the recent light-duty MYs
2012-2016 and MYs 2017-2025 rulemaking and for the heavy-duty MYs 2014-2018 rulemaking.
There was no serious  disagreement with this approach in the public comments to any of these
rulemakings. The ICMs for the light-duty context were developed in a peer-reviewed report from
RTI International and were subsequently discussed in a peer-reviewed journal article.499
Importantly, since publication of that peer-reviewed journal article, the agencies have revised the
methodology to include a return on capital (i.e., profits) based on the assumption implicit in
ICMs (and RPEs) that capital costs are proportional to direct costs, and businesses need to be
able to earn returns on their investments.

   There is some level of uncertainty surrounding both the ICM and RPE markup factors. The
ICM estimates used in this Draft TAR, consistent with the FRM,  group all technologies into
three broad categories and treat them as if individual technologies within each of the three
categories (low, medium, and high complexity) will have exactly the same ratio of indirect costs
to direct costs. This simplification means it is likely that the direct cost for some technologies
within a category will be higher and some lower than the estimate for the category in general.
Additionally, the ICM estimates were developed using adjustment factors developed in  two
separate occasions: the first, a consensus process, was reported in the RTI report; the second, a
modified Delphi method, was conducted separately and reported  in an EPA memorandum. Both
these panels were composed of EPA staff members with previous background in the automobile
industry; the memberships of the two panels overlapped but were not the same. The panels
evaluated each element  of the industry's RPE  estimates and estimated the degree to which those
elements would be expected to change in proportion to changes in direct manufacturing costs.
The method and the estimates in the RTI report were peer reviewed by three industry experts and
subsequently by reviewers for the International Journal of Production Economics. However, the
ICM estimates have not yet been validated through a direct accounting of actual indirect costs for
individual technologies. RPEs themselves are  also inherently difficult to estimate because the
accounting statements of manufacturers do not neatly categorize all cost elements as either direct
or indirect costs. Hence, each researcher developing an RPE estimate must apply a certain
amount of judgment to the allocation of the costs. Since empirical estimates of ICMs are
ultimately derived from the same data  used to  measure RPEs, this affects both measures.
                                             5-238

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
However, the value of RPE has not been measured for specific technologies, or for groups of
specific technologies. Thus applying a single average RPE to any given technology by definition
overstates costs for very simple technologies, or understates them for advanced technologies.
5.3.2.2.2
Indirect Cost Estimates Used in this Analysis
   Since their original development in February 2009, the agencies have made some changes to
both the ICM factors and to the method of applying those factors relative to the factors
developed by RTI and presented in their reports. These changes have been described and
explained in several rulemakings over the years, most notably the 2017-2025 FRM and the more
recent Heavy-duty GHG Phase 2 NPRM (80 FR 40137). In the 2015 NAS study, the committee
stated:  "The committee conceptually agrees with the Agencies' method of using an indirect cost
multiplier instead of a retail price equivalent to estimate the costs of each technology since ICM
takes into account design challenges and the activities required to implement each technology. In
the absence of empirical data, however, the committee was unable to determine the accuracy of
the Agencies' ICMs." (NAS Finding 7.1) EPA  continues to study the issues surrounding ICMs
but has not yet pursued further efforts given resource constraints and priorities in areas such as
technology benchmarking and cost teardowns. For this Draft TAR analysis, recognizing there are
uncertainties in the use of either ICM or RPE as indicators of indirect costs, as discussed above,
EPA chose to assess indirect costs using both the ICM and RPE approaches. NHTSA is
employing a similar approach of assessing costs based on both ICM and RPE  factors for the
CAFE analysis, as described in Section 5.4.  For the ICM case, EPA has applied the ICMs as
shown in Table 5.38. Near term values account for differences in the levels of R&D, tooling,
and other indirect costs that will be incurred. Once the program has been fully implemented,
some of the indirect costs will no longer be attributable to the standards and, as such, a lower
ICM factor is applied to direct costs. For the RPE case, EPA has applied an RPE factor of 1.5x
direct costs.  (EPA has also applied an RPE factor of 2.Ox direct costs for mass reduction costs,
as discussed below).
                    Table 5.38 Indirect Cost Multipliers Used in this Analysis1
                                                                  ,500

Complexity
Low
Medium
Highl
High2
2017-2025 FRM and this Draft TAR
Near term
1.24
1.39
1.56
1.77
Long term
1.19
1.29
1.35
1.50
   Here are two important aspects to the ICM method employed by EPA. First, the ICM consists
of two portions: a small warranty-related term and a second, larger term to cover all other
indirect costs elements. The breakout of warranty versus non-warranty portions to the ICMs are
presented in Table 5.39. The latter of these terms does not decrease with learning and, instead,
remains constant year-over-year despite learning effects which serve to decrease direct
manufacturing costs. Learning effects are described in the next section. The second important
note is that all indirect costs are forced to be positive, even for those technologies estimated to
have negative direct manufacturing costs.
                                             5-239

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                    Table 5.39 Warranty and Non-Warranty Portions of ICMs

Complexity
Low
Medium
Highl
High2
Near term
Warranty
0.012
0.045
0.065
0.074
Non-warranty
0.230
0.343
0.499
0.696
Long term
Warranty
0.005
0.031
0.032
0.049
Non-warranty
0.187
0.259
0.314
0.448
   The complexity levels and subsequent ICMs applied throughout this analysis for each
technology are shown in Table 5.40.
   Table 5.40 Indirect Cost Markups (ICMs) and Near Term/Long Term Cutoffs Used in EPA's Analysis
Technology
Aero, active
Aero, passive
Atkinson, level 1
Atkinson, level 2
Cam configuration changes
V6OHVtoV6DOHC
V6SOHCtoV6DOHC
VSOHVtoVSDOHC
VSSOHCtoVSDOHC
V8SOHC3VtoV8DOHC
Charger, in-home, EV
Charger, in-home, PHEV20
Charger, in-home, PHEV40
Charger, in-home, labor
Cylinder deactivation
Direct injection, stoichiometric, gasoline
Diesel, advanced (TierS)
Diesel, lean NOxtrap
Diesel, selective catalytic reduction
Downsizing, associated with turbocharging
!4DOHCtol3DOHC
!4DOHCtol4DOHC
V6OHVtol4DOHC
V6SOHCtol4DOHC
V6DOHCtol4DOHC
V8OHVtoV6DOHC
V8SOHCtoV6DOHC
V8SOHC3VtoV6DOHC
Engine friction reduction, level 1
Engine friction reduction, level 2
EGR, cooled
Electric power steering
EV75, battery pack
EV100, battery pack
EV200, battery pack
EV75, non-battery items
ICM Complexity
Low2
Med2
Med2
Med2

Med2
Med2
Med2
Med2
Med2
Highl
Highl
Highl
None
Med2
Med2
Med2
Med2
Med2

Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Low2
Low2
Med2
Low2
High2
High2
High2
High2
Short term thru
2018
2024
2018
2024

2018
2018
2018
2018
2018
2024
2024
2024
2024
2018
2018
2018
2018
2018

2018
2018
2018
2018
2018
2018
2018
2018
2018
2024
2024
2018
2024
2024
2024
2024
                                             5-240

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
EV100, non-battery items
EV200, non-battery items
HEV, Mild, battery pack
HEV, Mild, non-battery items
HEV, Strong, battery pack
HEV, Strong, non-battery items
HEV, Plug-in, battery pack
HEV, Plug-in, non-battery items
Improved accessories, level 1
Improved accessories, level 2
Low drag brakes
Lower rolling resistance tires, level 1
Lower rolling resistance tires, level 2
Lube, engine changes to accommodate low friction lubes
Mass reduction <15%
Mass reduction >=15%
Secondary axle disconnect
Stop-start
Turbo, 18-21 bar
Turbo, 24 bar
Turbo, Miller-cycle
TRX11/12
TRX21/22
High2
High2
Highl
Med2
Highl
Highl
High2
Highl
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Med2
Low2
Med2
Med2
Med2
Med2
Low2
Low2
2024
2024
2024
2018
2024
2018
2024
2018
2018
2018
2018
2018
2018
2018
2024
2024
2018
2018
2018
2024
2024
2018
2024
   For mass reduction costs, EPA has developed a new approach to calculating indirect costs due
to the unique nature of the direct manufacturing costs that EPA has developed (see Section
5.3.4.6.1). Mass reduction strategies, unlike other efficiency technologies, often involve multiple
systems and components on a vehicle. A portion of the indirect costs for parts that have design
and production outsourced to suppliers are incorporated into the direct manufacturing cost
estimates. Components that are designed in-house and possibly produced in-house by the
manufacturer, such as the body and frame structures, have higher indirect costs applied. This
distinction between supplier and in-house parts is consistent with the recommendations of a
study done by Argonne National Laboratory.501  In that study,  the authors suggested retail price
equivalent markups of 1.5x direct costs for parts sourced from a supplier, and 2x direct costs for
parts sourced internally. The end result, presumably, is an equal total cost, but the markups
account for differences in where the indirect costs are incurred. Using that as a basis EPA
adjusted the supplied technology ICMs (shown in Table 5.38) by the ratio 2/1.5 to determine in-
house ICMs at the "engineered solution" mass reduction point (described in Sections 5.3.4.6.1.1
and 5.3.4.6.1.2) which happened to be approximately 20 percent mass reduction level for the car
teardown study and the truck teardown study. Since those mass reduction levels were deemed
"medium" complexity levels in the FRM, and because EPA still believes that to be a good
assessment of the complexity level, EPA has worked with only the medium complexity ICMs in
the context of mass reduction. As a result, the ICMs used for mass reduction are as shown in
Table 5.41. For RPE based indirect costs, EPA simply used the 1.5x and 2x multipliers applied to
the same DMCs used in the ICM case.
                                             5-241

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
              Table 5.41 Mass Reduction Markup Factors used by EPA in this Draft TAR

Markup & Complexity
ICM - Medium complexity
RPE - complexity not applicable
Supplier Provided Mass Reduction
Near term
1.39
1.5
Long term
1.29
1.5
In-house Provided Mass Reduction
Near term
1.85
2.0
Long term
1.72
2.0
   The final element of the unique nature of the indirect cost calculations developed by EPA for
mass reduction in this analysis, is to calculate the indirect costs using the above ICMs or RPEs
only at the engineered solution point. Notably, EPA applied the markups to the sum of the
absolute values of all mass reduction ideas throughout the entire direct manufacturing cost curve.
In that way, negative direct costs that are projected at the lower mass reduction levels still have a
positive impact on calculated indirect costs. Once the indirect costs were determined via this
methodology at the engineered solution, EPA generated an indirect cost curve extending through
$0/kg at 0 percent mass reduction and $8.75/kg/% at the engineered solution for cars and
$13.23/kg/% for trucks (see Table 5.42 and Table 5.43 for the values of X). The indirect costs at
all mass reduction levels between those points lie on that generated cost curve. Inherent in this
approach is the assumption that the proportion of mass reduction from supplier and in-house
components remains constant at all levels of mass reduction, based on the proportion at the
engineered solution.  Those curves are shown in Table 5.42 for cars and in Table 5.43 for trucks.
          Table 5.42 Mass Reduction Indirect Cost Curves used by EPA for Cars Using ICMs

Near term
Long term

Supplied tech
DMC
In-house tech
DMC
Supplied tech
DMC
In-house tech
DMC
$/kg DMC*
$1.75
$1.16
$1.75
$1.16
ICM
0.39
0.85
0.29
0.72
$/kg 1C at
Engineered
Solution
$0.678
$0.986
$0.507
$0.835
$/kg 1C at Engineered
Solution
$0.678+0.986=1.66
$0.507+0.835=1.34
$/kg/%
1C
curve**
$8.75x
$7.06x
Notes:
* Calculated as the absolute value of all direct manufacturing costs needed to achieve the engineered solution.
** Where x is the percent mass reduction.
          Table 5.43 Mass Reduction Indirect Cost Curves used by EPA for Trucks Using ICMs

Near term
Long term

Supplied tech
DMC
In-house tech
DMC
Supplied tech
DMC
In-house tech
DMC
$/kg DMC*
$2.59
$2.09
$2.59
$2.09
ICM
0.39
0.85
0.29
0.72
$/kg 1C at
Engineered
Solution
$1.00
$1.78
$0.75
$1.50
$/kg 1C at Engineered
Solution
$1.00+1.78=2.78
$0.75+1.50=2.25
$/kg/%
1C
curve**
$13.23x
$10.73x
Notes:
                                              5-242

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
* Calculated as the absolute value of all direct manufacturing costs needed to achieve the engineered solution.
** Where x is the percent mass reduction.
           Table 5.44 Mass Reduction Indirect Cost Curves used by EPA for Cars Using RPEs

Near & Long
term

Supplied tech
DMC
In-house tech
DMC
$/kg DMC*
$1.75
$1.16
RPE
0.5
1.0
$/kg 1C at
Engineered
Solution
$0.875
$1.16
$/kg 1C at Engineered
Solution
$0.875+1.16=2.04
S/kg/%
1C
curve**
$10.71x
Notes:
* Calculated as the absolute value of all direct manufacturing costs needed to achieve the engineered solution.
** Where x is the percent mass reduction.
          Table 5.45 Mass Reduction Indirect Cost Curves used by EPA for Trucks Using RPEs

Near & Long
term

Supplied tech
DMC
In-house tech
DMC
$/kg DMC*
$2.59
$2.09
ICM
0.5
1.0
$/kg 1C at
Engineered
Solution
$1.30
$2.09
$/kg 1C at Engineered
Solution
$1.30+2.09=3.39
S/kg/%
1C
curve**
$16.12x
Notes:
* Calculated as the absolute value of all direct manufacturing costs needed to achieve the engineered solution.
** Where x is the percent mass reduction.
5.3.2.3 Maintenance and Repair Costs

5.3.2.3.1     Maintenance Costs

   To estimate maintenance costs that could reasonably be attributed to the 2017-2025 standards,
the agencies looked—in the 2017-2025 FRM—at vehicle models for which there exists a version
with a fuel efficiency and GHG emissions improving technology and a version with the
corresponding baseline technology. The difference between maintenance costs for the two
models represent a cost which the agencies  attributed to the standards.  For example, the Ford
Escape Hybrid versus the Ford Escape V6 was considered when estimating the types of
maintenance cost differences that might be present for a hybrid vehicle versus a non-hybrid, and
a Ford F150 with EcoBoost versus the Ford F150 5.0L was considered when estimating the types
of maintenance cost differences that might be present for a turbocharged and downsized versus a
naturally aspirated engine. In the case of low rolling resistance tires, specific parts were
considered rather than specific vehicle models.

   By comparing the manufacturer recommended maintenance schedule of the items compared,
the differences in maintenance intervals for the two was estimated. With estimates of the costs
per maintenance event, a picture of the maintenance cost differences associated with the "new"
technology was developed.

   EPA continues to believe that the maintenance estimates used in the FRM are still reasonable
and have therefore used them again in this analysis. EPA distinguished maintenance from repair
costs as follows: maintenance costs are those costs that are required to  keep a vehicle properly
maintained and, as such,  are usually recommended by auto makers to be conducted on a regular,
                                              5-243

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
periodic schedule. Examples of maintenance costs are oil and air filter changes, tire
replacements, etc. Repair costs are those costs that are unexpected and, as such, occur randomly
and uniquely for every driver, if at all. Examples of repair costs would be parts replacement
following an accident or a mechanical failure, etc.

   In Chapter 3.6 of the final joint TSD supporting the 2012 FRM, the agencies presented a
lengthy discussion of maintenance costs and the impacts projected as part of that rule.502 Table
5.46 shows the results of that analysis, the maintenance impacts used in the 2012 FRM and again
in this analysis, although the costs here have been updated to 2013$. Note that the technologies
shown in Table 5.46 are  those for which EPA believes that maintenance costs would change; it is
clearly not a complete list of technologies expected to meet the MY2025 standards.
                     Table 5.46 Maintenance Event Costs & Intervals (2013$)
New Technology
Low rolling resistance tires level 1
Low rolling resistance tires level 2
Diesel fuel filter replacement
EVoil change
EV air filter replacement
EV engine coolant replacement
EV spark plug replacement
EV/PHEV battery coolant
replacement
EV/PHEV battery health check
Reference
Technology
Standard tires
Standard tires
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Cost per Maintenance
Event
$6.71
$51.55
$51.93
-$40.78
-$30.16
-$62.21
-$87.52
$123.37
$40.78
Maintenance Interval
(miles)
40,000
40,000
20,000
7,500
30,000
100,000
105,000
150,000
15,000
   Note that many of the maintenance event costs for EVs are negative. The negative values
represent savings since EVs do not incur these costs while their gasoline counterparts do. Note
also that the 2010 FRM is expected to result in widespread use of low rolling resistance tires
level 1 (LRRT1) on the order of 85 percent penetration. Therefore, as 2012 FRM results in
increasing use of low rolling resistance tire level 2 (LRRT2), there is a corresponding decrease in
the use of LRRT1. As  such, as LRRT2 maintenance costs increase with increasing market
penetration, LRRT1 maintenance costs decrease. Importantly, the maintenance costs associated
with lower rolling resistance tires is the incremental cost of the tires at replacement; it is not
associated in any way with a decrease in durability  of these tires.
5.3.2.3.2
Repair Costs
   Both EPA's and NHTSA's FRM central analyses accounted for the costs of repairs covered by
manufacturers' warranties, and a sensitivity analysis estimated costs for post-warranty repairs.
The indirect cost multipliers (ICMs) applied in the agencies' analyses include a component
representing manufacturers' warranty costs. For the cost of repairs not covered by OEMs'
warranties, the agencies evaluated the potential to apply an approach similar to that described
above for maintenance costs. As for specific scheduled maintenance items, the ALLDATA
subscription database applied above provides estimates of labor and part costs for specific repairs
to specific vehicle models. However, although ALLDATA also provides service intervals for
scheduled maintenance items, it does not provide estimates of the frequency at which specific
failures may be expected to occur over a vehicle's useful life. The agencies have not yet been
                                             5-244

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
able to develop an alternative method to estimate the frequencies of different types of repairs,
and are therefore unable to apply these ALLDATA estimates in order to quantify the cost of
repairs throughout vehicles' useful lives. Moreover, the frequency of repair of technologies that
do not yet exist in the fleet, or are only emerging today provides insufficient representation of
what they will be in the future with wider penetration of those technologies. As a result, the
agencies assume per-vehicle repair costs during the post-warranty period are the same as the
OEM warranty period. To ensure repair costs for newer technologies are considered, those costs
are proportional to incremental direct costs. The frequency of repair is scaled by vehicle survival
rates.

5.3.2.4 Costs Updated to 2013 Dollars

   EPA is using technology costs from many different sources. These sources, having been
published in different years, present costs in different year dollars (i.e., 2009 dollars or 2012
dollars). For this analysis, the agencies sought to have all costs in terms of 2013  dollars to be
consistent with the dollars used by EIA in its Annual Energy Outlook 2015. While the factors
used to convert from 2009 dollars (or other) to 2013 dollars are small, the agencies prefer to be
overly diligent in this regard to ensure consistency across our analyses. The agencies have used
the GDP Implicit Price Deflator for Gross Domestic Product as the converter, with the actual
factors used as shown in Table 5.47.
          Table 5.47 Implicit Price Deflators and Conversion Factors for Conversion to 2013$
Calendar Year ->
Implicit Price Deflators for Gross
Domestic Product
Factor applied to convert to 2013$
2006
94.814
1.126
2007
97.337
1.097
2008
99.246
1.075
2009
100
1.067
2010
101.221
1.054
2011
103.311
1.033
2012
105.166
1.015
2013
106.733
1.000
Source: Bureau of Economic Analysis, Table 1.1.9 Implicit Price Deflators for Gross Domestic Product; last revised
on June 24, 2015; accessed on 7/8/2015 at !3iaJbea.£QY.

5.3.3   Approach for Determining Technology Effectiveness

   EPA reevaluated the effectiveness values for all technologies discussed in 2017-2025 LD
final rule for this Draft TAR, as well as prominent technologies that have emerged since then.
The process used to determine the effectiveness of each technology for this Draft TAR is similar
to the one used for the FRM. Along with the vehicle benchmarking and full vehicle simulation
process, EPA reviewed available data including the 2015 LD National Academy of Sciences
report503, confidential manufacturer estimates, OE and supplier meetings, technical conferences,
literature reviews, and press announcements regarding technology effectiveness. In most cases,
multiple sources of information were considered in the process of determining the effectiveness
values used in this assessment.

   Full vehicle simulation modeling has been used in both of the previous light-duty greenhouse
gas rules to establish the effectiveness of technologies, and is regularly applied by vehicle
manufacturers, suppliers, and academia to evaluate and choose alternative technologies to
improve vehicle efficiency. In the 2015 NAS report,503 the committee recognized the important
contribution of full vehicle simulation and lumped parameter modeling in these previous
rulemakings, and recommended continued use of these methods as the best way of assessing
technologies and the combination of technologies. While the full vehicle simulation modeling
results from Ricardo Engineering used in the 2017-2025MY FRM have been found to be robust
                                              5-245

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
and accurate, some of the underlying analyses performed by Ricardo were proprietary and could
not be fully disclosed to the public.

   For this Draft TAR, EPA is employing its own full vehicle simulation model; Advanced
Light-duty Powertrain and Hybrid Analysis tool (ALPHA). The ALPHA model has been
developed and refined over several years and used in multiple rulemakings to evaluate the
effectiveness of vehicle technology packages. Using ALPHA improves the transparency of the
process and provides additional flexibility to allow consideration of the most recent
technological developments and vehicle implementations of technologies. Input data for the
ALPHA model has been created largely through benchmarking activities. Benchmarking is a
commonly used technique that is intended to create  a detailed characterization of a vehicle's
operation and performance. For the purposes of developing ALPHA, and for establishing overall
technology effectiveness, EPA performed many benchmarking activities including measuring
vehicle performance over the standard emission cycles and measuring system and component
performance on various test stands.

5.3.3.1 Vehicle Benchmarking

   As part of its mandated evaluation of the appropriateness of the MY2022-2025 standards,
EPA is re-assessing any potential changes to the cost and the effectiveness of advanced
technologies available to manufacturers. See section 86.1818-12 (h) (l)(i) and (ii).
Benchmarking is a process by which detailed vehicle, system, and component performance is
characterized. Benchmarking is commonly used by  vehicle manufacturers, automotive suppliers,
national laboratories, and universities in order to gain a better understanding of how vehicles are
engineered and to create large datasets that can be applied in modeling and other analyses. In its
effort to assess light-duty vehicles in preparation for the MTE, EPA has benchmarked over
twenty vehicles, with the results summarized in 15 peer-reviewed SAE papers.504 505 As the
result of these activities,  EPA has calibrated the ALPHA full vehicle simulation model and
applied the results of this model to establish and confirm technology effectiveness. In addition,
EPA has also been able to capture the performance of current vehicles, which is an important
goal of the MTE. Over the coming years, the agency intends to continue to benchmark additional
vehicles to inform the Proposed and Final Determination.

   The ALPHA model has been used to confirm and update, where necessary, efficiency data
from the previous Ricardo study, such as from advanced downsized turbo and naturally  aspirated
engines. It is also being used to quantify effectiveness from advanced technologies which the
agencies did not project to be part of a compliance pathway during the FRM,  such as
continuously variable transmissions (CVTs), multi-mode normally aspirated engines, and clean
diesel engines. The ALPHA model accounts for synergistic effects between technologies and has
been used by EPA to calibrate the Lumped Parameter Model to incorporate the latest technology
package effectiveness data into the OMEGA compliance model.

   To simulate drive cycle performance, the ALPHA model requires various vehicle input
parameters, including vehicle inertia and road loads, and component efficiencies and operations.
Vehicle benchmarking is the detailed process for obtaining these parameters.

5.3.3.1.1      Detailed Vehicle Benchmarking Process
                                             5-246

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   The following discussion describes the vehicle benchmarking elements used as required for
the vehicles tested by EPA for this Draft TAR. The vehicle benchmarked in this example is a
2013 Chevy Malibu 1LS as detailed in Table 5.48. This vehicle was chosen as representative of a
midsize car with a typical conventional powertrain with a naturally aspirated engine and a 6
speed automatic transmission. The first task of the vehicle benchmarking process involved
collecting data from on-road and dynamometer testing (Figure 5.81) before removing the engine
and transmission for separate component testing. Major components such as the engine and
transmission of a vehicle must be isolated and evaluated separately to create accurate
performance maps to be included in the ALPHA model.
                          Table 5.48 Benchmark Vehicle Description
Model
Engine
Powertrain
Gear Ratios
Tire Size
EPA Label Fuel Economy
Emissions Equivalent Test Weight (ETW)
Emissions Target Road Load A
Emissions Target Road Load B
Emissions Target Road Load C
Fuel Economy ETW
Fuel Economy Target Road Load A
Fuel Economy Target Road Load B
Fuel Economy Target Road Load C
20 13 Chevy Malibu 1LS
2.5L inline-4, GDI, naturally aspirated
Conventional FWD 6-speed automatic
transmission
4.584,2.965, 1.912, 1.446, 1.000, 0.746 with 2
, GM6T40
89 final drive
215/60/R16
22 City, 34 Highway, 26 Combined MPG
4,000 Ibs (1814 kg)
38.08 Ibs (169.4 N)
0.2259 Ibs/mph (2.248 N/m/s)
0.01944 lbs/mphA2 (0.4327 N/(m/s)A2)
3,625 Ibs (1644 kg)
28.62 Ibs (127.3 N)
0.1872 Ibs/mph (1.863 N/m/s)
0.01828 lbs/mphA2 (0.4069 N/(m/s)A2)
                   Figure 5.81 Chevy Malibu Undergoing Dynamometer Testing
5.3.3.1.1.1    Engine Testing

   The engine was removed from the vehicle and installed in an engine dynamometer test cell, as
shown in Figure 5.82. The complete vehicle exhaust and emission control systems were included
                                            5-247

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
in the test setup. All necessary signals including the transmission input and output shaft speed
signals were supplied by the test stand to prevent engine controller fault codes. The engine was
fully instrumented to collect detailed performance information (e.g., exhaust/coolant
temperatures, cam angles, throttle position, mass airflow).
                               Figure 5.82 Engine Test Cell Setup


   The engine fuel consumption was measured at the steady state torque and speed operating
points as shown in Figure 5.83.
                                        Engine Map Points
                   300


                   250


                   200


                   150

                 g^
                 "sT 100
                 o
                    50


                     0


                   -50


                   -100
                           1000    2000
                                         3000    4000    5000
                                          Speed (RPM)
                                                              6000
                                                                     7000
5.3.3.1.1.2
                  Figure 5.83 Engine Map Points

Transmission Testing
   The 6-speed automatic transmission was removed from the vehicle and installed on a test
stand as shown in Figure 5.84. The transmission control solenoid commands were reverse
                                              5-248

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
engineered and the transmission was manually controlled during testing. Transmission line
pressure was externally regulated to match the pressures measured during chassis dynamometer
testing.  Torque and speed were measured at the input of the transmission and both outputs. The
input to the transmission was driven by an electric motor.
                        Figure 5.84 GM6T40 Transmission during Testing

   The transmission losses were measured at input torques ranging from 25 to 250 Nm and input
speeds ranging from 500 to 5000 RPM. For efficiency testing the torque converter clutch was
fully locked by manually overriding the clutch control solenoid. Tests were performed at two
transmission oil temperatures, 37 C and 93 C. Total efficiency for each gear during operation at
93 C, including pump and spin losses, is shown in Figure 5.85.

                          Transmission Total Efficiency -- All Gears, 93C 10bar
                     2000
                              4000
                                                  50
100    150

Input Torque (Nm)
                                                                     200
                                                                           250
                                      6000   0

                   Input Speed (RPM)

              Figure 5.85 Transmission Efficiency Data at 93 C and 10 Bar Line Pressure
                                              5-249

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The torque converter was tested unlocked in 6th gear to determine speed ratio (SR), K
factorQQ and torque ratio curves. The input speed to the transmission was held at 2000 RPM
while decreasing the output speed to traverse the SR curve from 1.0 to 0.35 (limited due to line
pressure and transmission slip). The data below  SR 0.35 was extrapolated using the higher SR
data. The torque converter data is  shown in Figure 5.86, with the K factor curve normalized by
dividing by the K factor at SR 0 (torque converter stall). Normalizing the K factor curve allows
for scaling the curve up or down by multiplying by a new stall K value.

                                Torque Ratio and Normalized K Factor
1.8'
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1
n Q
1.











X"\^










1









1
^1




1 	 •- 	 1




1




"^—B—







X^
N




c c
-• — Torque Ratio
-• — K Factor (norm)





\.
1









i










\ /
V

1

II



/
/
1

^A •
                  0    0.1   0.2    0.3   0.4    0.5   0.6   0.7   0.8   0.9    1
                                         Speed Ratio

        Figure 5.86 Torque Converter Torque Ratio and Normalized K Factor Versus Speed Ratio

   Transmission spin losses were measured in each gear with a locked torque converter and no
load applied to the output shaft while varying the input speed from 500 RPM to 3000 to 5000
RPM depending on the chosen gear. Spin loss testing was performed at 5 bar and 10 bar line
pressures and 37 C (cold) and 93 C (operating) oil temperatures. Figure 5.87 shows the spin loss
data at 93 C for all gears and both line pressures.
QQ
  K-factor is approximately equal to rpm/sqrt(torque).
                                              5-250

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                             Transmission Spin Loss Data - All Gears, 93C
5.3.3.7.2
                  500   1000   1500   2000   2500   3000  3500  4000   4500   5000
                                       Input Speed (RPM)

                          Figure 5.87 Transmission Spin Losses at 93C
Development of Model Inputs from Benchmarking Data
   After compiling the raw data, it was necessary to adapt the data to a form suitable for use by
the ALPHA model, including filling any data gaps and interpolating or extrapolating as required.

5. 3. 3. 7. 2. 7    Engine Data

   For use with the ALPHA model, the engine's fuel consumption map was created by
converting the set of points to a rectangular surface. In addition, an estimate of the engine inertia
was required since it plays a significant role in the calculation of vehicle performance and fuel
economy.506 The resulting engine data was reviewed with manufacturers prior to use in the
ALPHA model.
5.3.3.7.2.2   Engine Map

   Figure 5.88 shows one of the engine maps generated from the test stand data in terms of
brake-specific fuel consumption (BSFC).
                                             5-251

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                    Chevy Malibu 2.5L BSFC Map
              250-
              200-
              150 -
              100 -
                                                                         170KW
                                                                         150KW
                                                                         130KW
                                                                         110kW
                                                                         90 kW
              50 gP'
                    1000   1500   2000  2500   3000   3500   4000   4500  5000  5500   6000
                                         Speed ( RPM )


                           Figure 5.88 Chevy Malibu 2.5L BSFC Map
5.3.3.1.2.3    Inertia

   Engine inertia plays a significant role in vehicle performance and fuel economy, particularly
in the lower gears due to the high effective inertia (proportional to the square of the gear ratio)
and higher acceleration rates.

   To estimate the combined inertia of the engine, its attached components, and the torque
converter impeller, a simple test was performed in-vehicle: the engine was accelerated with the
transmission in park to the engine's maximum governed speed, then the ignition was keyed off,
and the engine speed and torque were observed until the engine stopped. Engine speed and
reported engine torque data (shown as negative during ignition off) were collected. The data was
then run through a simple simulation and the inertia varied until the model deceleration rate
reasonably matched the observed deceleration rate down to 500 RPM. Figure 5.89 shows the
model result using a 0.2 kg-mA2 total inertia with the engine drag torque.
                                              5-252

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                 Malibu 2.5L Spindown Inertia Test
                                                              CAN Speed
                                                              Model Speed
                                                              CAN Torque
                                                                          -20
                            0.5
1        1.5
  Time (Seconds)
2.5
                           Figure 5.89 Engine Spin down Inertia Test

   A wet torque converter from the 2013 Malibu was weighed and measured to estimate the
inertia. The weight of 12.568 kg and total diameter of 0.273 m gives an estimated 0.0585 kg-mA2
total inertia. For the purposes of modeling this inertia was then proportioned 2/3 for the impeller
side and 1/3 for the turbine side based on the inertia split from other known torque converters.

   Subtracting the estimated torque converter inertia results in an engine (including all attached
components) inertia of approximately 0.161 kg-mA2 (0.2 - 2/3*0.0585).

   The exact proportioning of the inertia makes no difference to the outcome of the model (since
the total inertia is always the same) but can guide future work or estimates of component inertias.

5.3.3.1.2.4    Transmission Data

   For use with the model, the total transmission efficiency data needed to be separated into gear
efficiency and pump/spin torque losses. Torque converter back-drive torque ratio and K factor
also needed to be calculated.

5.3.3.1.2.5    Gear Efficiency and Spin Losses

   To separate the gear efficiency from the total efficiency (which includes the pump/spin
losses), the total efficiency data for each gear was converted to torque loss data and the spin loss
torques were subtracted. The resulting gear torque loss data was then converted to an efficiency
lookup tables.  Some data points had to be extrapolated to cover the full speed and/or torque
range. For example, first gear was only tested to 150 Nm but the full table required data up to
250 Nm. Figure 5.90 shows the estimated gear efficiencies for all gears. This process was
followed for both the 37 C and 93 C data.
                                             5-253

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Transmission pump losses were factored out of the spin losses (as a rough approximation,
since no pump loss data was available), using the lowest common spin loss to represent the pump
loss.

                          Transmission Gear Efficiency - All Gears, 93C 10 bar
              0.86
              300
                      200
                              100
                    Torque (Nm)
                                       0   0
                                                 1000
                                                       2000
                                                              3000
                                                                    4000
                                                                           5000
                                                        Speed (RPM)
                 Figure 5.90 Gear Efficiency Data At 93 C and 10 Bar Line Pressure
5.3.3.1.2.6    Torque Converter

   To complete the model inputs for the torque converter, the torque ratio and K factor need to
be calculated for the full range of speed ratios.

   The torque converter back-drive torque ratio is assumed to be 0.98 for all speed ratios. The
back-drive K factor is calculated from the drive K factor mirrored relative to speed ratio (SR)  1
and shifted upwards by 70 percent. The K factor at SR 1 is calculated, for modeling purposes, as
7.5 times the highest drive K factor. In practice the K factor at SR 1 is either poorly defined or
near infinite so the model requires a large value but not so large as to make the solver unstable.
Figure 5.91 shows the given (SR < 0.95) and calculated torque converter data.

   These additional data points have little effect on the modeled fuel economy but are required
for model operation and smooth transitions from positive to negative torques.
                                              5-254

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                 20
                 18
                 16
                 14
                 12
                 10
                                   Torque Ratio and Normalized K Factor
                                                               - Torque Ratio
                                                               - K Factor (norm)
                  0     0.2    0.4    0.6     0.8     1     1.2    1.4
                                          Speed Ratio
 Figure 5.91 Torque Converter Drive and Back-Drive Torque Ratio and Normalized K Factor versus Speed
                                          Ratio
5.3.3.1.3
Vehicle Benchmarking Summary
   Section 5.3.3.1 outlined the vehicle benchmarking process for a typical vehicle. While
complex, this process yields the necessary input parameters for physics based full vehicle
simulation models such as ALPHA. The following list represents the main model input
parameters generated from the benchmarking process:

       •  Engine Maps:
          0   Fuel Consumption
          0   BSFC
          0   Friction/Inertia
          0   Performance
       •  Transmission Maps
          0   Efficiency
          0   Torque Converter
          0   Shifting Strategy
       •  Vehicle:
          0   Road Loads
          0   Mechanical Loads
          0   Electrical Loads
   This information plus the remaining known vehicle characteristics (mass, etc.) provide the
model with all of the necessary information needed for simulation. During the initial
                                              5-255

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
development of the ALPHA model, this complete data set from several vehicles was used to
validate all of the internal calculations of the model. Once the model was validated, a wide
variety of engines, transmissions, and other vehicle components were introduced to model
current and future vehicles. This process is described in Section 5.3.3.2.

5.3.3.2 ALPHA Vehicle Simulation Model

   The Advanced Light-Duty Powertrain and Hybrid Analysis (ALPHA) tool was created by
EPA to evaluate the Greenhouse Gas (GHG) emissions of Light-Duty (LD) vehicles. In the two
prior rules, EPA relied on Ricardo to conduct full vehicle simulations. In order to have additional
flexibilities and transparency, EPA developed an in-house full vehicle simulation model that
could freely be released to the public. Model development, along with the data collection and
benchmarking that comes along with model calibration, is an extremely effective means of
developing expertise and deeper understanding on technologies. Better understanding of
technologies makes for more robust regulatory analysis. Having a model available in-house also
allows EPA to make rapid modifications as new data is collected, which cannot be done easily
with contractors.

   For the Draft TAR, EPA has achieved significantly higher levels of transparency for its
modeling than was anticipated when beginning the work several years ago.  Throughout this
section of the Draft TAR, EPA has provided details on the major technology assumptions built
into ALPHA. EPA has also provided extensive technical details in the  docket for the Draft TAR
describing the process used to build the fuel consumption maps for six of the engines mentioned
in the Draft TAR, as well as data maps for two transmissions.507 In the time leading up to the
publication of the Draft TAR, EPA has published over 15 peer-reviewed papers describing
results of key testing, validation and analyses.

   In-house development of the models continues to be more accurate,  efficient, transparent, and
cost-effective than relying on contractors. EPA began developing both light-and heavy-duty
vehicle simulations simultaneously as these vehicles share many of the same basic components.
The light-duty vehicle model (ALPHA), and the heavy-duty model (GEM), share the same basic
architecture.

   EPA has validated the ALPHA model using several sources including vehicle
benchmarking,508 stakeholder data, and industry literature.  While the ALPHA model is
continuing to be refined and calibrated, the version in use as of April 26, 2016 was externally
peer reviewed.509 To further enhance transparency, EPA has included the results  of this external
peer review on its website along with a copy of this specific version of the ALPHA model that
was reviewed (peer review input data and run-able MatLab Simulink source code).

5.3.3.2.1     General ALPHA Description.

   ALPHA is a physics-based, forward-looking, full vehicle computer simulation capable of
analyzing various vehicle types with different powertrain technologies, showing realistic vehicle
behavior. The software tool is a MATLAB/Simulink based desktop application.

   Within ALPHA, an individual vehicle is defined by specifying the appropriate vehicle road
loading (inertia weight and coast-down coefficients) and specifications of the powertrain
components. Powertrain components (such as engines or transmissions) are individually
parameterized and can be exchanged within the model.
                                            5-256

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Vehicle control strategies are also modeled, including engine accessory loading, decel fuel
shutoff, hybrid behavior, torque converter lockup, and transmission shift strategy. Transmission
shifting is parameterized and controlled by ALPHAshift,510 a shifting strategy algorithm that
ensures an appropriate shifting strategy when engine size or vehicle loading changes. The control
strategies used in ALPHA are modeled after strategies recorded during actual vehicle testing.

   Vehicle packages defined within ALPHA can be run over any pre-determined vehicle cycle.
To determine fuel consumption values used to calculate LD GHG rule CCh values, an FTP and
HWFET cycle are simulated, separated by a HWFET prep cycle as normally run during
certification testing. ALPHA does not include a temperature model, so the FTP is simulated
within the model assuming warm component efficiencies for all bags. Additional fuel
consumption due to the FTP cold start is calculated in post-processing by applying a fuel
consumption penalty to bags 1 and 2, depending on the assumed warmup strategy. Any vehicle
drive cycle can be defined and fuel economy simulated in ALPHA. For example, the results from
the US06, NEDC, and WLTP cycles (among others) are used to tune vehicle control strategy
parameters to match simulation results to measured vehicle test results across a variety of
conditions. In addition, performance cycles  have been defined, which are used to determine
acceleration performance metrics.

5.3.3.2.2      Detailed ALPHA Model Description

   The ALPHA model architecture is comprised of four systems: Ambient, Driver, Powertrain,
and Vehicle as seen in Figure 5.92. With the exception of Ambient and Driver, each system
consists of one or more subcomponents. The function of each system  and its respective
component models are discussed in this chapter. The structure and operation described in this
section incorporate numerous constructive comments from both public comments and peer
reviews. The model has been upgraded to integrate new technologies, improve the fidelity of the
simulation results and better match the operation of the benchmarked  vehicles. This all supports
our primary goal of accurately reflecting changes in technology for both the current and future
light duty  fleet. As part of this effort, substantial effort has been put forth to accurately track and
audit power flows through the model to ensure conservation of energy, and provide better data
on technology effectiveness.
                                             5-257

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                   ALPHA  Vehicle Model
    [^    I 	^ I	1
       I
                   Scope
[power! rain]

 [driver]

 [ambient]

 [vehicle]
                                     ^] -»<^[system_bus]|
                                    Memory
                                                             system_bus
                                                             mass_in_kg
                                                                      veh_spd_mps
                                                                    vehicle
                                                             system_bus      bus_out
                                                                    driver
                                                             sy stem_bus      bus_out
                                                                   ambient
                           Figure 5.92  ALPHA Model Top Level View
   One of the novel features of ALPHA is the inclusion of dynamic lookup tables. These tables
allow additional customization of models for specific vehicles. This is enabled by a table
description within the parameters for a component. This allows tables in the model such as
transmission losses to be parameterized in a way that best matches the available data for that
particular component.

5.3.3.2.2.1    Ambient System

   This system defines ambient conditions such as pressure, temperature, and road gradient,
where vehicle operations are simulated. ALPHA has been calibrated to generate fuel economy
results corresponding to chassis dynamometer certification tests; therefore conditions within the
simulation have been maintained to align with current test procedures.

5.3.3.2.2.2    Driver System

   The driver model in ALPHA is a purely proportional-integral control driver that features a
small look ahead to anticipate upcoming accelerations in the drive cycle. This is especially useful
at launch where the vehicle response may be delayed due to the large effective inertia in lower
gears. The driver in ALPHA is designed to follow a vehicle speed versus time driving cycle such
as the UDDS or HWFET. The driver is tuned to mimic activities of a real driver during a chassis
test,  including starting the engine, putting the transmission into gear and then operating both the
accelerator and brake pedals.
                                               5-258

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.3.2.2.3   Powertrain System

   The engine, transmission, electrical systems and accessories discussed in the following
section are combined to form vehicle powertrain systems. The conventional powertrain system
shown in Figure 5.13 contains sub-models representing each of the components. Additional
powertrains were constructed to simulate power split and P2 hybrid as well as full electric
drivetrains.

             ALPHA CVM Conventional Vehicle Model
                                                                                  *QD
                                                                                      force_out_N
                 Figure 5.93 ALPHA Conventional Vehicle Powertrain Components
5.3.3.2.2.3.1  Engine Subsystem

   The engine model is built around a steady-state fuel map covering all engine speed and torque
conditions with torque curves restricting operation between wide open throttle (full load) and
closed throttle (no load). The engine fuel maps for various engines are provided by benchmark
data, generated via tools like GT-Power, or adapted from other data sources. The engine fuel
map contains fuel mass flow rates vs engine crankshaft speed and brake torque. In-cylinder
combustion processes are not modelled.

   The steady-state fuel map used in ALPHA is adapted from the available test data or model
output by creating an interpolant grid covering the area between idle speed and redline speed,
and between the wide open throttle  and closed throttle curves. In some circumstances, portions of
the map (for example, those near redline speed or near the closed throttle curve) are extrapolated
from the original data. In general, these area represent engine operation which is either outside of
that used in two-cycle operation (near redline speed) or which uses little fuel in general (near the
closed throttle curve).

   During the simulation, the engine speed at a given point in the drive cycle is calculated from
the physics of the downstream speeds.  The quantity of torque required is  calculated from the
                                             5-259

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
driver model accelerator demand, an idle speed controller, and requests from the transmission
during shifts. The torque request is then limited by a torque response model which has been
tuned to match the torque response of naturally aspirated and turbocharged gasoline and diesel
engines. The resulting engine torque and speed are used to interpolate a fuel rate from the fuel
map.

   Additional sources of fuel consumption documented in benchmarking activities have been
included in the model as well. On gasoline engines, the torque management that occurs during
shifting is implemented such that the reduction in torque does not cause a corresponding
reduction in the fuel rate. This approximates the effect of the observed spark retard to lessen the
lurch associated with decelerating engine inertia during upshifts. Another source of additional
fueling occurs after engines transition out of decel fuel cutoff. Additional fuel is applied for a
few seconds for emissions control.  Finally, there are additional fuel penalties applied within the
simulation associated with rapid changes in engine power.

5.3.3.2.2.3.2   Electric Subsystem

   The electric subsystem consists  of 3 major components, battery, starter, and alternator.

   The battery model for ALPHA was created after a literature review of battery models,
particularly for hybrid vehicle applications. The same battery model structure511'512 is used for
both conventional and hybrid vehicles, with different calibrations used to simulate different
chemistries such as lead-acid or lithium ion. The model features an open circuit voltage that
varies with state of charge, a series resistance, and dual RC time constant filters to provide
realistic voltage response. Calibrations were generated from published literature or benchmark
testing for the open circuit voltage  and transient behavior. The simulated  battery also features a
thermal model, with the output current limited at extremes in temperature or state  of charge.

   The engine starter is modeled as a simplified electric motor. It has a fixed efficiency and is
commanded via a Boolean activation signal. The operation of the starter is characterized by a
desired cranking speed and a torque capacity. These values are generally  calculated to match the
engine specifications. When an engine start is requested a proportional integral controller is used
to determine the torque applied to accelerate the engine to the desired cranking speed, limited by
the torque capacity. The mechanical power required and efficiency then determine the resulting
electrical power consumed.

   The engine alternator is modeled as a simplified  electric generator with fixed efficiency. The
electrical output current is determined by a charging controller. The efficiency and electrical
power output can then be used to compute the mechanical load applied to the engine. The
charging controller can operate in two different modes. In a basic mode it always tries to charge
the battery to a fixed voltage target. It also features  an adaptive charging / alternator regen mode
that varies the voltage target and thus current output to driving conditions. Lower  electrical
output is provided during cruising,  enough to maintain a minimal state of charge. During
decelerations and transmission upshifts electrical output and thus mechanical load are increased
to capture energy that would otherwise be dissipated via the brakes or transmission. The adaptive
charging / alternator regen strategy exhibits increased variability of battery state of charge over
various driving cycles. Therefore it is necessary to precondition the model with  a prep cycle just
as would be done on  a test such as the HWFET to get accurate results.
                                              5-260

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.3.2.2.3.3  Accessories Subsystem

   The accessories subsystem in ALPHA is responsible for applying electrical and mechanical
loads to mimic those observed during testing. The system is capable of applying 4 different
loads: power steering, air conditioning, fan and a generic load to cover the remaining losses
observed. Each load can apply mechanical loads to the engine crankshaft and/or electrical loads
to the battery. Each load can be independently correlated to model signals via dynamic lookup
tables, and is calibrated to match test data. Baseline vehicles with mechanical power steering
often have mechanical losses that vary with engine speed, while future vehicles featuring electric
power steering have electrical losses that vary with vehicle speed.

5.3.3.2.2.3.4   Transmission Subsystem

   The transmission subsystem features different variants representing the major types of
transmissions that are currently in use in LD vehicles. The different transmission models are
built from similar components, but each features a unique control algorithm matching behaviors
observed during vehicle benchmarking.

   One of the features in ALPHA, which is required for the model to conserve energy, is
multiple speed integrators. One is located at each of the points in the driveline where rotational
inertias may become decoupled such as the transmission gearbox. These integrators use the
torque and upstream inertia to compute the resulting acceleration and thus speed for the upstream
components. For couplings that may become locked up, such as completing a transmission shift,
the torques and rotational inertia are then passed down toward the next integrator in the model.

5.3.3.2.2.3.4.1 Transmission Gear Selection

   All of the gear transmission models use a dynamic shift algorithm, ALPHAshift,513 to
determine the operating gear over the cycle. This employs a rule based approach utilizing the
engine torque curve and fuel map to select gears that optimize efficient engine operation and
provide a torque reserve as a traditional transmission calibration would. The ALPHAshift
algorithm attempts to select the minimum fuel consumption gear after applying constraints on
engine speed and torque reserve. It also allows downshifts due to high driver demand.RR

   The ALPHAshift algorithm contains  calibration parameters that can be tuned to match
benchmarked shift behavior data from a particular engine and transmission. A generic calibration
tuning strategy has been developed from these specific benchmarked calibrations,  and is useful
for simulating the shifting behavior of engine and transmission  combinations that are from
different vehicles or represent future technologies.

   The CVT transmission model uses a similar ALPHAshiftCVT514 algorithm for determining
gear ratio selection. It attempts to maintain operation on an engine speed vs requested power line
that minimizes fuel consumed. This method also has constraints for minimum engine speed and
the rate at which the gear ratio can be changed.
  Also known as a power downshift or kickdown.
                                             5-261

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.3.2.2.3.4.2 Clutch Model

   The clutch model in ALPHA can be modulated during launch and requires a fixed time to
engage. Torque is conserved across the clutch during engagement and the inertial effects of
accelerating and decelerating the upstream inertias are captured. This additional fidelity
necessitates a more complicated control algorithm to manage clutch slip during launch which is
included in the control strategy for the appropriate transmissions.

   Two clutches are bundled together to create the dual clutch module for the dual clutch
transmission. The dual clutch features a single integrator for calculating engine speed during
shifts.

5.3.3.2.2.3.4.3 Gearbox Model

   The gearbox model for ALPHA has been developed with the goal of simulating realistic
operation during shifts for all types of transmissions. The gearbox contains gear ratios and
properly scales torque and rotational inertia through the ratio change. Power loss within the
gearbox are applied via dynamic lookup tables which determine torque loss and/or gearbox
efficiency. These loss tables are typically constructed using signals such as input torque, input
speed, commanded gear and/or line pressure.

   Realistic shifting behavior is achieved with appropriate delays provided by a synchronizer
clutch model. The layout of the gearbox model is most similar to a manual transmission, but the
application for a planetary gearbox is a reasonable approximation once the neutral delay between
gears is omitted.

   The gearbox rotational inertias are split between a common input inertia, common output
inertia and a gear specific inertia. The common inertias represent rotational inertia always
coupled to the input or output shafts. The gear specific inertias, which are only used for planetary
automatic transmissions, are added or removed as gears are engaged or disengaged and incur
additional losses as the rotational inertia is spun up.

5.3.3.2.2.3.4.4 Torque Converter Model

   The torque converter model in ALPHA simulates a lockup-type torque converter. The torque
multiplication and resulting engine load are calculated via torque ratio and K-factor curves that
vary as a function of speed ratio across the torque converter. Base torque ratio and the K-factor
curves are often scaled in situations where detailed torque converter information is unavailable.

   The lockup behavior of the torque converter is accomplished by integrating a clutch model
similar to the one discussed above. The torque converter model also contains a pump loss torque
that is implemented via a dynamic lookup table to simulate the power required to operate the
pump on an automatic transmission.

5.3.3.2.2.3.4.5 Automatic Transmission & Controls

   The automatic transmission (AT) is composed of the torque converter and gearbox systems
discussed above. The AT is allowed to shift under load. During upshifts and torque converter
lockup the engine output torque is slightly reduced to minimize the resultant torque pulse
encountered by decelerating the engine inertia.
                                             5-262

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The torque converter lockup clutch command is determined based on transmission gear and
gearbox input speed. The thresholds that trigger lock and unlock of the torque converter are
calibrated to match benchmark data.

5.3.3.2.2.3.4.6 DCT Transmission & Control

   The ALPHA DCT model is constructed from two separate gearbox components and a dual
clutch module as described above. The dual clutch module features a dynamic lookup torque loss
table that can be used to represent all the gearbox losses in one location if loss information for
the separate gearboxes is not available. After a gear change to a new preselected gear is
requested, the dual clutch module will transition and begin applying  torque through the new gear.

   The DCT transmission controller also includes a low speed clutch engagement routine to
feather the clutch for low speed operation  or launch. Similar to the automatic transmission
engine output torque is reduced during upshifts to minimize the torque pulse at the wheels.

5.3.3.2.2.3.4.7CVT Transmission & Control

   The CVT transmission in ALPHA consists of the torque converter and gearbox modules.
When operating as a CVT the gearbox maintains a state of partial engagement allowing the gear
ratio to be constantly changed.

5.3.3.2.2.3.4.8 Driveline

   The driveline system contains all of the components that convert the torque at the
transmission output to force at the wheels. This includes drive shafts as well as driven axles,
consisting of a differential, brakes and tires. ALPHA is capable of simulating multiple axles, but
it is often simpler to convert a driveline to a single axle equivalent.

   The driveshaft is a simple component for transferring torque while adding additional
rotational inertia. It is only used for rear wheel drive vehicles.

   The final drive is modeled as a gear ratio  change with an associated torque loss and/or
efficiency. These losses are applied via a  dynamic lookup table. For front wheel drive
transmissions, the final drive losses are often difficult to separate. In these situations all losses
are applied in the gearbox.

   The brake system on each axle applies  a torque to the axle proportional to the brake pedal
position from the driver model. The brake torque capacity is scaled to match the stopping
requirements of the vehicle.

   The tire component model transfers the torques and rotational inertias from upstream
components to a force and equivalent mass that is passed to the vehicle model. This conversion
uses the loaded tire radius and adds the tire's rotational inertia. A force associated with the tire
rolling resistance is not simulated because these losses are included in the road load ABC
coefficients applied within the vehicle subsystem.

5.3.3.2.2.3.5   Vehicle System

   The vehicle system consists of the chassis, its mass  and forces associated with aerodynamic
drag, rolling resistance, and changes in road  grade. The vehicle system also contains the vehicle
speed integrator that computes acceleration from the input force and equivalent mass which is
                                             5-263

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
integrated to generate vehicle speed and distance traveled. The road load force is calculated from

the ABC coefficients determined through coast down testing, or modified to simulate future

improvements.


5.3.3.2.3     Energy Auditing


   One of the quality control components within the ALPHA model is an auditing of all the

energy flows. This auditing enables verification that the physics represented in the model is

done correctly, generally resulting in a simulation energy error less than a few hundredths of a

percent. The audit data can also be compared between simulations to verify that individual

component losses are reasonable when compared to baseline packages or products that may

feature similar technologies. An example energy audit report for a package similar to a current

production sedan is shown in the figure below. It should be noted that the lack of final drive

losses in this case is attributed to the vehicle being front wheel drive, and the thus the final drive

losses are included in the gearbox.

                          	 Energy Audit Report  	
                 Total Energy consumed           = 205975.66 ku
                      Fuel Energy                 = 205971.83 kJ
                      Stored Energy               =     3.83  kJ
                          Battery Internal  Losses =     0.91  kn      0.00%
                      Kinetic Energy              =     0.00  kJ
                      potential Energy            =     0.00  kj

                 Usable System Energy Provided    = 63307.72  kJ
                      Engine Energy               = 63304.80  kJ
                          Engine Efficiency      =    30.73  %
                      stored Energy               =     2.92  ki
                      Kinetic Energy              =     0.00  kJ
                      Potential Energy            =     0.00  ku

                 Energy consumed by ABC roadload = 37465.54  kj     59.17%
                 Energy Consumed by gradient     =     0.00  kj      0.00%
                 Energy Consumed by brakes        = 11429.43  kJ     18.05%
                 Energy Consumed by Accessories  =  3505.29  kJ      5.54%
                      Starter                     =     0.45  kl      0.00%
                      Alternator                  =  1225.99  kl      1.94%
                      Battery Stored Charge        =     0.00  kJ      0.00%
                      Engine Fan                  =     0.00  kJ      0.00%
                          Electrical              =     0.00  kJ      0.00%
                          Mechanical              =     0.00  ki      0.00%
                      power steering              =     0.00  ki      0.00%
                          Electrical              =     0.00  k3      0.00%
                          Mechanical              =     0.00  kJ      0.00%
                      Air Conditioning            =     0.00  kJ      0.00%
                          Electrical              =     0.00  ki      0.00%
                          Mechanical              =     0.00  kJ      0.00%
                      Generic Loss                =  2278.85  kJ      3.60%
                          Electrical              =  2278.85  kJ      3.60%
                          Mechanical              =     0.00  kn      0.00%
                      Total Electrical Accessories =  2278.85  ki      3.60%
                      Total Mechanical Accessories =     0.00  kJ      0.00%
                 Energy Consumed by Driveline    = 10913.96  kJ     17.24%
                      Launch Device              =  1796.61  k3      2.84%
                      Gearbox                    =  8150.72  kJ     12.87%
                          Pump Loss              =  3243.10  kJ      5.12%
                          Spin Loss              =  2837.18  kJ      4.48%
                          Gear/inertia Loss      =  2070.44  kJ      3.27%
                      Final  Drive                =     0.00  kJ      0.00%
                      Tire Slip                  =   966.63  kJ      1.53%
                 Net  System Kinetic Energy Change =     0.44  kJ      0.00%

                 Total Loss Energy               = 63314.66  kJ
                 Simulation Error                =    -6.94  kl
                 Energy conservation             =  100.011  %

                        Figure 5.94 Sample ALPHA Energy Audit Report
                                              5-264

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.3.2.4     ALPHA Simulation Runs

   ALPHA was used to perform a series of simulation runs, where various technology packages
were compared to a baseline vehicle. The baseline vehicle was chosen to have component
efficiencies and vehicle loads consistent with the baseline vehicles used in the modeling runs in
the FRM. Four acceleration performance metrics were calculated for the baseline vehicle: 0-60
time, Vi mile time, 30-50 passing time, and 50-70 passing time. These metrics were chosen to
give a reasonably broad set of acceleration metrics that would be sensitive enough to represent
true acceleration performance, but not so sensitive that minor changes in vehicle parameters
would significantly change the final metric.

   For each subsequent comparative run, a vehicle package was defined within ALPHA by
specifying powertrain components and road load specifications. ALPHA'S road load force at a
specific vehicle velocity (v) is determined by using the following formula: F = Cv2 + Bv + A
where the coastdown coefficients (A, B, and C) are derived from a least squares fit of data from
track coast-down tests.

   In ALPHA modeling, it is assumed that the A coefficient is a factor for the road load force
that is mostly associated with tire rolling resistance, the B coefficient is a small factor, which
represents higher order rolling resistance and gearing loss factors, and the C coefficient is a
factor which mostly represents aerodynamic air drag. Thus, changes in aerodynamic losses are
modeled by changing the C coefficient, and changes in rolling resistance losses are modeled by
changing the A coefficient. Changes in mass reduction are modeled by reducing the  test weight,
and by reducing the A coefficient (as rolling resistance is a function of vehicle weight).

   The nominal engine size for the package was determined based on the estimated performance
effect of the technologies included in the package. The same performance metrics calculated for
the baseline vehicle were calculated for each package, and the sum compared to the baseline
sum. If the sum was not within three percent, the torque output (and thus size) of the engine was
adjusted and the performance cycle rerun until an equivalent acceleration performance was
attained.

   Once the appropriate engine size was determined, the base engine map was adjusted by first
scaling the torque output of the original map by the appropriate factor, and then adjusting the
BSFC so as not to overestimate the efficiency gain from using a smaller engine. As engine size is
reduced, the cylinder surface area to volume ratio increases, which increases the relative heat
losses and decreases efficiency. An adjustment factor corresponding to approximately 1 percent
increase in BFSC for every 10 percent decrease in engine displacement was used to adjust the
engine maps.  This factor is consistent with the well-known rule of thumb governing efficiency
losses due to wall heat losses515, and with the process used by Ricardo, Inc. in the FRM, to scale
the BSFC maps.

   Once the engine was appropriately scaled, the final vehicle package was run through an FTP
and HWFET cycle simulation as described above to determine fuel consumption values.

5.3.3.2.5     Post-processing

   ALPHA simulation runs are performed assuming warm component efficiencies. Additional
fuel consumption due to the FTP cold start is calculated in post-processing by applying a fuel
consumption penalty to bags 1 and 2. These fuel consumption penalty factors represent
                                             5-265

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
additional fuel used to heat the catalyst, and additional energy lost to higher viscosity lubricating
oil in the engine and transmission. The fuel consumption penalties for "present" and "past"
vehicles are set at 15 percent (present) to 17 percent (past) for bag 1 and 2.5 percent for bag 2.
The penalty factors are applied during post-processing so that the fuel consumption for the
appropriate bag is increase by the indicated amount. These factors were determined by
comparing the "cold" FTP bags 1 and 2 to the "warm" bags 3 and 4 for a range of vehicles.

   Since the three-bag FTP is a standard test, the difference in fuel consumption between bags 1
and 3 of the FTP could be calculated for the entire fleet (available in the Test Car List data
files516), as seen in the graph below. However, the data sources for bag 4 are more limited. EPA
based the 2.5 percent penalty factor on test data available from conventional vehicle testing from
Argonne National Labs517 and from internal testing, where differences between bags 2 and 4
averaged about 2.5 percent.
0%
1

•
•

0 15 20
                                          25     30
                                            FTP mpg
                                                        35
                                                              40
                                                                     45
    Figure 5.95 Example: Difference in 2016, Between Bags 1 and 3 of the FTP, from the Test Car List.

   For simulation of advanced vehicle packages which included thermal management of the
engine or transmission, the penalty factors were reduced (to a minimum of 11 percent for bag 1
and 0 percent for bag 2) to account for the reduction in losses associated with faster component
warmup.
5.3.3.2.6
Vehicle Component Vintage
   Vehicle components (engines and transmissions) are assigned a vintage of "past," "present,"
or "future." The vintage of the component determines the assumed technology package
associated with the component, and thus the default value of some associated parameters.

   One parameter affected by vintage is electric accessory loading. The "past" value for electrical
loads includes a base electrical load of 154 W, additional power draw based on engine speed
(approximately 700 W at 2500 rpm and 1050 W at 6000 rpm), and an alternator efficiency of 55
percent. These values are based on the modeling Ricardo did for the FRM, and assumes
mechanical power steering. The "present" value for electrical load includes a base electrical  load
of 490 W, no additional variable accessory power draw, and an alternator efficiency of 65
percent. This is based on loads measured in various tested vehicles, in particular the Chevrolet
Malibu.518 The "future" electrical load maintains the same 490 W base electrical load, but with a
                                             5-266

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
high-efficiency (70 percent efficient) alternator. EPA is reviewing the values used for accessory
loading, and may update them based on the results of the review.

   Another parameter is the cold start penalty applied during post-processing. It is assumed that a
bag 1 cold start penalty of 17 percent is associated with past engines, and a bag 1 cold start
penalty of 15 percent is associated with present engines, as described in the section above. Future
engines receive a bag 1 cold start penalty of 11 percent, representing the effect of thermal
management of the engine included in the engine friction reduction package. Likewise, for past
and present transmissions, a bag 2  cold start penalty of 2.5 percent, while for future
transmissions the high-efficiency gearbox fast warmup technology is assumed, and a bag 2 cold
start penalty of 0 percent is applied.

   Future vintage transmissions are also assumed to be associated with early torque converter
lockup.

   Although the assigned vintage determines default values for accessory loads and cold start
penalty, these defaults can be overridden in the model to examine the effects of specific
technologies separately.

5.3.3.2.7      Additional Verification

   As an additional verification of ALPHA model simulations, technology package combinations
are further compiled and executed  using a hardware-in-the-loop (HIL) system. This process
enables powertrain, vehicle, and driver behavior to be observed in real time for both on-cycle
and off-cycle situations. Any undesirable behavior is analyzed and used to fine tune the
modeling process. These compiled HIL models are  also utilized in the vehicle benchmarking
process when testing vehicle subsystems such as engines, transmissions, battery modules, and
other components.  Figure 5.96 shows an example ALPHA model simulation  observation
display.
                                             5-267

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
             Figure 5.96 Example ALPHA Model UDDS Simulation Observation Display

   As part of EPA's on-going quality process, several comparative analyses were completed as
part of the ongoing MTE work.  ALPHA results have been compared Ricardo EASY5 results
from the original MY2017-2025 Light-Duty FRM, as well as with results from
Autonomie.519 When viewing the models as a calculators, then providing the same inputs to the
calculators should provide the same outputs. Results of both comparisons showed only minor
differences between simulation results due to specific model behaviors or implementations,
convincing EPA that these models are very close in terms of computational results when run
using the same input data and assumptions.

5.3.3.3 Determining Technology Effectiveness for MY2022-202 5

   EPA collected information on the effectiveness of current CCh emission reducing
technologies from a wide range  of sources.  The primary sources of information were the 2017-
2025 FRM, EPA's ALPHA model, EPA's vehicle benchmarking studies, the 2015 NAS Report,
OEM and Supplier meetings, and industry literature. In addition, EPA considered confidential
data submitted by vehicle manufacturers, along with confidential information shared by
automotive industry component suppliers in meetings with EPA, CARB, and NHTSA staff.
These confidential data sources were used primarily as a validation of the estimates  since EPA
prefers to rely on public data rather than confidential data wherever possible.

   EPA recognizes that technologies will be further developed and introduced for MY2022-2025
and that innovation by automobile manufacturers and suppliers will continue to occur. While it is
impossible for the agency to predict all of the technologies that will come to fruition, likely
trends can be identified in the development of automotive systems that impact GHG emissions
over the next decade. EPA uses  methods similar to those used by industry to identify and
                                            5-268

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
evaluate emerging automotive technology trends. The use of computer aided engineering (CAE)
tools for technology evaluation has been a key source of technology effectiveness data for
MY2022-2025 vehicle technology packages. A number of other sources of data are also used to
either validate CAE results or as independent sources of effectiveness data. Sources of data
include:

       1)  Engineering analysis of logical developments based on current or near-term
          technology
       2)  Review of peer-reviewed journal papers, U.S. Department of Energy Reports, and
          other public sources of peer-reviewed data
       3)  Purchase and review of proprietary reports by major automotive industry analytical
          firms (e.g., R.L. Polk, IHS Automotive)
       4)  Meetings with automobile manufacturers
       5)  Meetings with Tier 1 automotive suppliers
       6)  Contracts with major automotive engineering design, analysis, and services firms
          (e.g., FEV, Munro and Associates, Southwest Research Institute, Ricardo PLC) to
          purchase data or engineering services
       7)  "Proof of concept" research either conducted directly by EPA at EPA-NVFEL or
          under contract with engineering  services firms
       8)  CAE tools, including:
          a) Engine modeling (e.g., Ricardo WAVE, Gamma Technologies GT-POWER)
          b) Vehicle modeling (e.g., EPA LPM, EPA ALPHA, Ricardo RSM, MSC EASY5)
          c) HIL simulation of drive cycles
          d) Computational fluid dynamics (CFD) for initial component development
       9)  Chassis dynamometer testing
       10) Engine dynamometer testing
       11) Transmission dynamometer testing


   Data from all sources listed above is used to develop and validate vehicle effectiveness within
the EPA ALPHA model and EPA LPM. Modeling of technology package effectiveness within
the ALPHA model and LPM is the source of all technology package effectiveness data contained
within the OMEGA cost-effectiveness analyses. With respect to engine and powertrain
technologies, the general progression of data into the OMEGA analyses for this Draft TAR has
been:

       1)  Develop physics-based models of the technology with extensive validation of a base
          configuration to actual hardware (e.g., validation of an engine model to actual engine
          performance, combustion measurements and knock characteristics)
       2)  Use the validated physics-based model to evaluate hardware changes and to develop
          calibrations necessary to account for such hardware changes
       3)  Use the ALPHA model to determine the CO2 effectiveness of the powertrain package
          for different vehicle configurations
       4)  Compare the energy balance of ALPHA model results with vehicle benchmark results
          as an additional plausibility analysis.
       5)  Use ALPHA modeling results to provide a calibration for technology package
          effectiveness within the LPM
                                            5-269

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
       6)  Validate ALPHA modeling results using a variety of data sources (chassis
          dynamometer testing of production or developmental vehicles, HIL testing of
          developmental engine configurations, comparison with automobile manufacturer and
          Tier 1 supplier data, comparison with peer-reviewed/published data sources)
       7)  Update LPM calibration with validated ALPHA model technology package
          effectiveness
       8)  Use technology package effectiveness from the LPM within the OMEGA cost-
          effectiveness analysis for this Draft TAR


   The EPA analysis of naturally aspirated Atkinson cycle engines provides an example of an
analytical framework that integrates CAE together with other methods used by EPA to evaluate
future vehicle technologies. The 2.0L Mazda SKYACTIV-G engine was introduced in 2012 in
the U.S. This engine represents state-of-the art brake thermal efficiency and is the first non-HEV
application of an Atkinson cycle engine in a U.S. light-duty vehicle application. EPA conducted
chassis dynamometer testing of Mazda vehicles with the SKYACTIV-G engine and also
purchased versions of this engine marketed in the U.S. (13:1 geometric compression ratio) and
EU (14:1 geometric compression ratio) for detailed engine dynamometer mapping and HIL
testing. After both chassis dynamometer testing and initial engine dynamometer testing, an
engineering analysis was conducted to prioritize near-term technologies that could potentially
yield further brake thermal efficiency improvements, broaden areas of high thermal efficiency
and/or better align high brake thermal efficiency operation with both the regulatory drive cycles
and with urban driving with the goal of meeting the 2022-2025 GHG standards in a "standard
car" configuration (approximately D-segment size-class).

   The technologies chosen for further analysis included:

       •  Improving alignment of high brake thermal efficiency  operation with urban driving
          via road load reduction, switching to an advanced 8-speed automatic transmission,
          and using fixed 4/2 cylinder deactivation
       •  Improving brake thermal efficiency by increasing expansion the ratio from 13:1 to
          14:1 along with the addition of low-pressure-loop EGR for additional knock
          mitigation on standard pump fuel and additional pumping loss improvements


   An initial proof of concept evaluation of increased expansion ratio, low-pressure-loop cooled
EGR and cylinder deactivation was conducted using GT-POWER engine modeling.520 Engine
dynamometer testing with HIL simulation of regulatory drive cycles was used for initial proof of
concept evaluation of switching to use of an advanced 8-speed automatic transmission and using
road-load reduction and application of the 2.0L SKYACTIV-G to larger D-segment vehicles.521
Combinations of these technologies were also compared to similar vehicle configurations using
turbocharged, downsized GDI engines using the ALPHA vehicle model.522 An important part of
EPA's use of CAE has been to validate CAE results using other data sources. For example,
ALPHA modeling and HIL testing were validated using chassis dynamometer test data and GT-
POWER modeling was validated using engine dynamometer test data.
                                            5-270

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.3.4 Lumped Parameter Model

   It is widely acknowledged that full-scale physics-based vehicle simulation modeling is the
most thorough approach for estimating future benefits of a package of new technologies. This is
especially important for quantifying the efficiency of technologies and groupings (or packages)
of technologies that do not currently exist in the fleet or as prototypes. However, developing and
executing every possible combinations of technologies directly in a fleet compliance model using
full scale vehicle simulation would not be practical to implement.

   As part of rulemakings, EPA analyzes a wide array of potential technology options rather than
attempt to pre-select the "best" solutions. For example, analysis for the MYs 2017-2025 Light
Duty Vehicle GHG rule, EPA built over 800,000 packages for use in its OMEGA compliance
model, which spanned 19 vehicle classes and over 1,200 baseline vehicle models. The Draft
TAR analysis has expanded the number of baseline vehicle models to approximately 2,200. The
lumped parameter approach was again chosen as the most practical surrogate to estimate the
effectiveness of the technology package combinations for the Draft TAR analysis.

   As in the FRM,  the basis for calibrating and validating the lumped parameter model for this
assessment is the effectiveness data generated by the benchmarking and full vehicle simulation
modeling activities described earlier in this  section.  The lumped parameter model also allows
benchmarked and/or simulated vehicle packages to be separated into individual components to
properly account for the technologies already in the vehicle fleet to avoid any double counting of
these technologies. General Motors (Patton  et al)523 presented a vehicle energy balance analysis
to  highlight the synergies  that arise with the combination of multiple vehicle technologies. This
report demonstrated an  alternative methodology (to vehicle simulation) to estimate these
synergies, by means of a "lumped parameter" approach. This approach served as the basis for
EPA's lumped parameter  model.  The Lumped Parameter approach has recently been endorsed
by the National Academy of Science:  "In particular, the committee notes that the use of full
vehicle simulation modeling in combination with lumped parameter modeling has improved the
agencies'  estimation of fuel economy impacts."524

   As described in  Section 5.3.3.2.3, the ALPHA simulation results used to calibrate the lumped
parameter model are checked against conservation of energy requirements as part of the quality
assurance process.  Similarly, the basis for EPA's lumped parameter analysis is a first-principles
energy balance that estimates the manner in which the chemical energy of the fuel is converted
into various forms of thermal and mechanical energy on the vehicle. The analysis accounts for
the dissipation of energy into the different categories of energy losses, including each of the
following:

       •    Second  law losses (thermodynamic losses inherent in the combustion of fuel)
       •   Heat lost from the combustion process to the exhaust and coolant
       •   Pumping losses, i.e., work  performed by the engine during the intake and exhaust
           strokes
       •   Friction losses in the engine
       •    Transmission losses, associated with friction and other parasitic losses of the gearbox,
          torque converter (when applicable) and driveline
       •   Accessory losses, related directly to the parasitics associated with the engine
           accessories
                                             5-271

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
       •  Vehicle road load (tire and aerodynamic) losses
       •  Inertial losses (energy dissipated as heat in the brakes)


   It is assumed that each baseline vehicle has a fixed percentage of fuel lost to each category.
Each technology is grouped into the major types of engine loss categories it reduces. In this way,
interactions between multiple technologies that are applied to the vehicle may be determined.
When a technology is applied, the lumped parameter model estimates its effects by modifying
the appropriate loss categories by a given percentage. Then, each subsequent technology that
reduces the losses in an already improved category has less of a potential impact than it would if
applied on its own.

   Using a lumped parameter approach for calculating package effectiveness provides necessary
grounding to physical principles. Due to the mathematical structure of the model, it naturally
limits the maximum effectiveness achievable for a family of similar technologies. This can prove
useful when computer-simulated packages are compared to a "theoretical limit" as a plausibility
check. Additionally, the reduction of certain energy loss categories directly impacts the effects
on others. For example, as mass is reduced the benefits of brake energy recovery decreases
because there is not as much inertia energy to recapture.

   The LP model has been updated from the MYs 2017-2025 final rule for this Draft TAR.
Changes were made to include new technologies for 2017 and beyond and to improve fidelity for
baseline attributes and technologies. In addition, the LP model has been calibrated to follow the
results of the ALPHA full vehicle simulation model to facilitate the vehicle package  building
process used in the OMEGA model.

5.3.3.4.1     Lumped Parameter Model Usage in OMEGA

   The Lumped Parameter Model (LPM) is used in the OMEGA model to incrementally improve
the effectiveness of vehicle models in the baseline fleet. As a first step, approximately fifty
technology packages are created with increasing effectiveness for each vehicle type.  Several
example packages are shown in Table 5.49.
        Table 5.49 Example OMEGA Vehicle Technology Packages (values are for example only)
Package #
0
1
2
10
20
Technology Package
4-Speed Auto
6-Speed Auto
8-Speed Auto + DCP
8-Speed + DCP + TURB24
8-Speed + DCP + Aero2 + TURB24 + 10%MR
Technology
Package
Effectiveness
0%
4%
10%
20%
28%
   Step two selects the next vehicle in the baseline fleet and applies all fifty technology packages
in sequence using the LPM to calculate a new effectiveness value at each step. As the
technologies in the baseline vehicles have been tabulated based on publically available data, the
incremental effectiveness improvement will not include these baseline vehicle technologies to
avoid double counting. Table 5.50 contains an example baseline vehicle. Table 5.51 illustrates
the package application process.
                                             5-272

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                 Table 5.50 Example Baseline Vehicle (values are for example only)
Baseline Vehicle Technologies
6-Speed Auto + DCP
Baseline Vehicle
Effectiveness
6%
            Table 5.51 Example Package Application Process (values are for example only)
Package #



0
1
2
10
20
Technology Package



4-Speed Auto
6-Speed Auto
8-Speed Auto + DCP
8-Speed + DCP + TURB24
8-Speed + DCP + Aero2 + TURB24 + 10%MR
Technology
Package
Effectiveness

0%
4%
10%
20%
28%
Resulting
Vehicle
Incremental
Effectiveness
0%
0%
3%
11%
17%
   As shown, the incremental effectiveness is not simply additive as the LPM takes into account
synergies and dis-synergies between the existing and applied technologies. This process also
enables the OMEGA model to assign baseline vehicles a cost to represent their existing
technologies and calculate an incremental cost to match with the incremental effectiveness as
each technology package is applied. The completed technology package effectiveness values
from the LPM are compared to the corresponding ALPHA model results as shown in Table 5.52
as a final check before they are used in the OMEGA model.  This calibration process is an
important step to ensure that full vehicle simulation results from the ALPHA model are used as
the primary effectiveness inputs to the OMEGA model.
                                             5-273

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                         Table 5.52 Example LPM Calibration Check
Technology Package
Standard Car+LUB
+EFR1+DCP+SGDI+6AT
+HEG1+EPS+IACC1
Standard Car+LUB
+EFR1+DCP+SGDI+8AT
+HEG1+EPS+IACC1
Standard Car+LUB
+EFR2+ATK2+DCP
+SGDI+6AT+HEG1+EPS
+IACC1
Standard Car+LUB
+EFR2+ATK2+DCP
+SGDI+8AT+HEG1+EPS
+IACC1
Standard Car+LUB
+EFR2+ATK2+CEGR
+DEAC+DCP+SGDI+8AT
+HEG2+EPS+IACC2
Standard Car+LUB
+EFR2+TURB24+CEGR
+DEAC+DCP+SGDI+8AT
+HEG2+EPS+IACC2
Standard Car+LUB
+EFR2+ATK2+CEGR
+DEAC+DCP+SGDI+8AT
+HEG2+EPS+IACC2
Standard Car+LUB
+EFR2+ATK2+CEGR
+DEAC+DCP+SGDI+8AT
+HEG2+EPS+IACC2
Standard Car+LUB
+EFR2+ATK2+CEGR
+DEAC+DCP+SGDI+8AT
+HEG2+EPS+IACC2
Standard Car+LUB
+EFR2+ATK2+CEGR
+DEAC+DCP+SGDI+8AT
+HEG2+EPS+IACC2
Standard Car+LUB
+EFR2+TURB24+CEGR
+DEAC+DCP+SGDI+8AT
+HEG2+EPS+IACC2
Mass
0%
0%
0%
0%
0%
0%
10%
0%
0%
10%
10%
Aero
0%
0%
0%
0%
0%
0%
0%
20%
0%
20%
20%
Roll
0%
0%
0%
0%
0%
0%
0%
0%
20%
20%
20%
ALPHA
Effectiveness
from
Reference
Package
0.0%
7.1%
4.9%
11.2%
26.9%
26.3%
30.5%
30.4%
30.3%
37.8%
37.3%
LPM
Effectiveness
from
Reference
Package
0.0%
6.9%
4.8%
11.2%
26.8%
26.2%
30.5%
30.3%
30.3%
37.5%
37.1%
Delta
Effectiveness
from
Reference
Package
0.0%
-0.2%
-0.1%
0.0%
-0.1%
-0.1%
0.0%
-0.1%
0.0%
-0.3%
-0.2%
LPM
Effectiveness
from Null
Package
16.5%
22.3%
20.5%
25.9%
38.9%
38.4%
42.0%
41.8%
41.8%
47.8%
47.4%
   The complete list of baseline fleet vehicles each incremented approximately fifty times results
in approximately 100,000 improved vehicles as input to the OMEGA model.
                                            5-274

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The effectiveness reductions and costs that are associated with applying a technology will
depend on the starting point technologies from which the cost and effectiveness improvements
are measured. For example, two vehicle models that start with different packages of technologies
will likely have different costs and effectiveness, even if both models finally arrive at the same
package combination of technologies. The agencies' recognition of the importance of clearly
specifying the point of comparison for cost and effectiveness estimates is consistent with the
NAS committee's finding "that understanding the base or null vehicle, the order of technology
application, and the interactions among technologies is critical for assessing the costs and
effectiveness for meeting the standards."

   As long as the point of comparison is  maintained consistently throughout the analysis for both
the baseline and future fleets, the decision of where to place an origin along the scale of cost and
effectiveness is inconsequential. For EPA's technology assessment, the origin is defined to
coincide with a "null technology package," which represents a technology floor such that all
technology packages considered in this assessment will have equal or greater effectiveness,
consistent with the FRM approach.  While other choices would have been equally valid, this
definition of a "null package"  has the practical benefits of avoiding technology packages with
negative effectiveness values, while also allowing for a direct comparison of effectiveness
assumptions with the FRM.

5.3.4   Data and Assumptions Used in  GHG Assessment

5.3.4.1 Engines: Data and Assumptions for this Assessment

   The majority of engine technologies used in this assessment are detailed in Section 5.2 of this
Draft TAR.  This section details engine technology information specific to the EPA GHG
analysis.

   In an effort to characterize  the efficiency and performance of late model vehicle powertrains,
and to update our engine data  from that used in the FRM, EPA tested several engines at the
National Vehicle and Fuel Emission Laboratory and contractor facilities. Depending on the
information required, the engines were tested with their factory and/or developmental engine
management  systems that allowed EPA engineering staff to calibrate engine control parameters.
Figure 5.97 illustrates  a typical engine test.
  Figure 5.97 2.0L 14 Mazda SKYACTIV-G engine Undergoing Engine Dynamometer Testing at the EPA-
                                      NVFEL Facility.
                                             5-275

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   In some cases, future engine configurations can be modeled using engine simulation software.
EPA used Gamma Technologies GT-POWER engine simulation software to model future engine
configurations based upon the Mazda 2.0L 14 SKYACTIV-G engine and the MAHLE
turbocharged/downsized 1.2L 13 GDI Di3 engine. Computer-aided engineering tools, including
GT-POWER, are commonly used during the initial stages of product development by automotive
manufacturers and academia to establish the potential performance of engine design features,
with respect to efficiency, emissions, and performance. GT-POWER is a physics based suite of
software that combines predictive diesel or spark-ignition combustion models; CAD-based,
preprocessed libraries of the physical layout of induction, exhaust and combustion systems;
models of chemical kinetics; wave dynamics models; turbocharger turbine and compressor
models with surge, reverse-flow and pressure wave prediction; induction turbulence models; a
kinetic knock model; injector spray models and an ability to apply minor adjustments to model-
predicted parameters using data from engine dynamometer measurements.  Engine dynamometer
data was also used to directly validate simulations of specific engine hardware configurations via
comparisons of measured vs. modeled values for knock intensity, combustion phasing, FMEP,
BTE and other parameters.
5.3.4.1.1
Low Friction Lubricants (LUB)
   Based on the analysis for the 2017-2025 FRM, the agencies estimated the effectiveness of
LUB to be 0.5 to 0.8 percent. EPA has reviewed this technology and finds the effectiveness
estimate remains applicable for this Draft TAR.

   The cost associated with making the engine changes needed to accommodate low friction
lubes is equivalent to that used in the 2012 FRM except for updates to 2013 dollars. The costs
are shown below.
    Table 5.53 Costs for Engine Changes to Accommodate Low Friction Lubes (dollar values in 2013$)
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$3
Low2

DMC: learning
curve
1C: near term thru
1
2018

2017


$3
$1
$4
2018


$3
$1
$4
2019


$3
$1
$4
2020


$3
$1
$4
2021


$3
$1
$4
2022


$3
$1
$4
2023


$3
$1
$4
2024


$3
$1
$4
2025


$3
$1
$4
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.1.2     Engine Friction Reduction (EFR1, EFR2)

   Based on the analysis for the 2017-2025 FRM, EPA estimated the effectiveness of EFR1 at
2.0 to 2.7 percent. Based on the analysis for the 2017-2025 FRM, EPA estimated the
effectiveness of EFR2 at 3.4 to 4.8 percent. EPA has reviewed this technology and finds the
effectiveness estimate remains applicable for this Draft TAR.

   The costs associated with engine friction reduction are equivalent to those used in the 2012
FRM except for updates to 2013 dollars. The costs are shown below first for engine friction
reduction level 1 and then for level 2.
                                            5-276

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
             Table 5.54 Costs for Engine Friction Reduction Level 1 (dollar values in 2013$)
Engine
13
14
V6
V8
13
14
V6
V8
13
14
V6
V8
Cost type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
DMC: base year cost
1C: complexity
$37
$50
$74
$99
Low2
Low2
Low2
Low2




DMC: learning curve
1C: near term thru
1
1
1
1
2018
2018
2018
2018
2018
2018
2018
2018
2017
$37
$50
$74
$99
$9
$12
$18
$24
$46
$62
$92
$123
2018
$37
$50
$74
$99
$9
$12
$18
$24
$46
$62
$92
$123
2019
$37
$50
$74
$99
$7
$10
$14
$19
$44
$59
$89
$118
2020
$37
$50
$74
$99
$7
$10
$14
$19
$44
$59
$89
$118
2021
$37
$50
$74
$99
$7
$10
$14
$19
$44
$59
$89
$118
2022
$37
$50
$74
$99
$7
$10
$14
$19
$44
$59
$89
$118
2023
$37
$50
$74
$99
$7
$10
$14
$19
$44
$59
$89
$118
2024
$37
$50
$74
$99
$7
$10
$14
$19
$44
$59
$89
$118
2025
$37
$50
$74
$99
$7
$10
$14
$19
$44
$59
$89
$118
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
             Table 5.55 Costs for Engine Friction Reduction Level 2 (dollar values in 2013$)
Engine
13
14
V6
V8
13
14
V6
V8
13
14
V6
V8
Cost type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
DMC: base year cost
1C: complexity
$81
$106
$155
$205
Low2
Low2
Low2
Low2




DMC: learning curve
1C: near term thru
1
1
1
1
2024
2024
2024
2024
2024
2024
2024
2024
2017
$81
$106
$155
$205
$20
$26
$38
$50
$101
$131
$193
$254
2018
$81
$106
$155
$205
$20
$26
$38
$50
$101
$131
$193
$254
2019
$81
$106
$155
$205
$20
$26
$38
$50
$101
$131
$193
$254
2020
$81
$106
$155
$205
$20
$26
$38
$50
$101
$131
$193
$254
2021
$81
$106
$155
$205
$20
$26
$38
$50
$101
$131
$193
$254
2022
$81
$106
$155
$205
$20
$26
$38
$50
$101
$131
$193
$254
2023
$81
$106
$155
$205
$20
$26
$38
$50
$101
$131
$193
$254
2024
$81
$106
$155
$205
$20
$26
$38
$50
$101
$131
$193
$254
2025
$81
$106
$155
$205
$16
$20
$30
$39
$97
$126
$185
$244
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.1.3
Cylinder Deactivation (DEAC)
   Within the analysis for the 2017-2025 FRM, EPA estimated an effectiveness of 6 percent for
DEAC. EPA has reviewed this technology and changed the effectiveness estimate to 3.9 to 5.3
percent for this Draft TAR.

   The costs associated with cylinder deactivation are equivalent to those used in the 2012 FRM
except for updates to 2013 dollars and use of a new learning curve (curve 24). Note that the 2012
FRM did not carry a cost for cylinder deactivation on an 1-4 engine. For this Draft TAR, we have
used half the cost of cylinder deactivation on a V8 engine. The costs are shown below.
                  Table 5.56 Costs for Cylinder Deactivation (dollar values in 2013$)
Engine
14
V6
V8
14
V6
V8
14
V6
V8
Cost type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
DMC: base year cost
1C: complexity
$86
$153
$172
Highl
Med2
Med2



DMC: learning curve
1C: near term thru
24
24
24
2018
2018
2018



2017
$82
$146
$164
$48
$59
$66
$130
$205
$230
2018
$80
$143
$161
$48
$59
$66
$129
$202
$227
2019
$79
$141
$158
$29
$44
$49
$108
$184
$207
2020
$78
$138
$155
$29
$44
$49
$107
$182
$205
2021
$76
$136
$153
$29
$44
$49
$106
$180
$202
2022
$75
$134
$151
$29
$44
$49
$105
$178
$200
2023
$74
$132
$149
$29
$44
$49
$104
$176
$198
2024
$73
$130
$147
$29
$44
$49
$103
$174
$196
2025
$72
$129
$145
$29
$43
$49
$102
$172
$194
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                               5-277

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.4.1.4
Intake Cam Phasing (ICP)
   Within the analysis for the 2017-2025 FRM, EPA estimated an effectiveness of 2.1 to 2.7
percent for ICP.  EPA has reviewed this technology and finds the effectiveness estimate remains
applicable for this Draft TAR.

   The costs associated with intake cam phasing are equivalent to those used in the 2012 FRM
except for updates to 2013 dollars and use of a new learning curve (curve 24). The costs are
shown below.
                  Table 5.57 Costs for Intake Cam Phasing (dollar values in 2013$)
Engine
OHC-I
OHC-V
OHV-V
OHC-I
OHC-V
OHV-V
OHC-I
OHC-V
OHV-V
Cost type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
DMC: base year cost
1C: complexity
$41
$81
$41
Low2
Low2
Low2



DMC: learning curve
1C: near term thru
24
24
24
2018
2018
2018



2017
$39
$78
$39
$10
$20
$10
$49
$97
$49
2018
$38
$76
$38
$10
$20
$10
$48
$96
$48
2019
$37
$75
$37
$8
$16
$8
$45
$90
$45
2020
$37
$73
$37
$8
$16
$8
$44
$89
$44
2021
$36
$72
$36
$8
$16
$8
$44
$88
$44
2022
$36
$71
$36
$8
$16
$8
$43
$87
$43
2023
$35
$70
$35
$8
$16
$8
$43
$86
$43
2024
$35
$69
$35
$8
$16
$8
$42
$85
$42
2025
$34
$68
$34
$8
$16
$8
$42
$84
$42
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                             5-278

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.4.1.5
Dual Cam Phasins (DCP)
   Based on the analysis for the 2017-2025 FRM, EPA estimated the effectiveness of DCP to be
between 4.1 to 5.5 percent. EPA has reviewed this technology and finds the effectiveness
estimate remains applicable for this Draft TAR.

   The costs associated with dual cam phasing are equivalent to those used in the 2012 FRM
except for updates to 2013 dollars and use of a new learning curve (curve 24). The costs are
shown below.
                   Table 5.58  Costs for Dual Cam Phasing (dollar values in 2013$)
Engine
OHC-I
OHC-V
OHC-I
OHC-V
OHC-I
OHC-V
Cost type
DMC
DMC
1C
1C
TC
TC
DMC: base year cost
1C: complexity
$74
$160
Med2
Med2


DMC: learning curve
1C: near term thru
24
24
2018
2018


2017
$71
$153
$29
$61
$100
$214
2018
$70
$150
$29
$61
$98
$211
2019
$68
$147
$21
$46
$90
$193
2020
$67
$145
$21
$46
$89
$190
2021
$66
$142
$21
$46
$87
$188
2022
$65
$140
$21
$46
$86
$186
2023
$64
$138
$21
$46
$86
$184
2024
$63
$136
$21
$46
$85
$182
2025
$63
$135
$21
$45
$84
$180
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.1.6     Discrete Variable Valve Lift (DWL)

   Based on the analysis for the 2017-2025 FRM, EPA estimated the effectiveness for DVVL at
4.1 to 5.6 percent. EPA has reviewed this technology and finds the effectiveness estimate
remains applicable for this Draft TAR.

   The costs associated with discrete variable valve lift are equivalent to those used in the 2012
FRM except for updates to 2013 dollars and use of a new learning curve (curve 24). The costs
are shown below.
              Table 5.59 Costs for Discrete Variable Valve Lift (dollar values in 2013$)
Engine
OHC-I
OHC-V
OHV-V
OHC-I
OHC-V
OHV-V
OHC-I
OHC-V
OHV-V
Cost type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
DMC: base year cost
1C: complexity
$127
$184
$263
Med2
Med2
Med2



DMC: learning curve
1C: near term thru
24
24
24
2018
2018
2018



2017
$122
$176
$252
$49
$71
$101
$171
$247
$353
2018
$119
$173
$247
$49
$71
$101
$168
$244
$348
2019
$117
$170
$243
$37
$53
$76
$154
$223
$318
2020
$115
$167
$239
$36
$53
$76
$152
$220
$314
2021
$113
$164
$235
$36
$53
$75
$150
$217
$310
2022
$112
$162
$231
$36
$53
$75
$148
$215
$307
2023
$110
$160
$228
$36
$53
$75
$146
$212
$303
2024
$109
$158
$225
$36
$53
$75
$145
$210
$300
2025
$107
$156
$222
$36
$53
$75
$144
$208
$297
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.1.7
Continuously Variable Valve Lift (CWL)
   Based on the analysis for the 2017-2025 FRM, EPA estimated the effectiveness for CVVL at
5.1 to 7.0 percent. EPA has reviewed this technology and finds the effectiveness estimate
remains applicable for this Draft TAR.

   The costs associated with continuously variable valve lift are equivalent to those used in the
2012 FRM except for updates to 2013 dollars and use of a new learning curve (curve 24). The
costs are shown below.
                                             5-279

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
             Table 5.60 Costs for Continuously Variable Valve Lift (dollar values in 2013$)
Engine
OHC-I
OHC-V
OHV-V
OHC-I
OHC-V
OHV-V
OHC-I
OHC-V
OHV-V
Cost type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
DMC: base year cost
1C: complexity
$191
$350
$381
Med2
Med2
Med2



DMC: learning curve
1C: near term thru
24
24
24
2018
2018
2018



2017
$182
$334
$365
$73
$135
$147
$256
$469
$512
2018
$179
$328
$358
$73
$134
$147
$252
$462
$504
2019
$176
$322
$351
$55
$100
$110
$230
$422
$461
2020
$173
$317
$345
$55
$100
$109
$227
$417
$455
2021
$170
$312
$340
$55
$100
$109
$225
$412
$449
2022
$167
$307
$335
$55
$100
$109
$222
$407
$444
2023
$165
$303
$330
$54
$100
$109
$220
$403
$439
2024
$163
$299
$326
$54
$100
$109
$217
$399
$435
2025
$161
$295
$322
$54
$100
$109
$215
$395
$431
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.1.8
Investigation of Potential Future Non-HEVAtkinson Cycle Engine Applications
   EPA initiated an internal study to investigate potential improvements in the incremental
effectiveness of Atkinson Cycle engines through the application of cooled EGR, an increase in
compression ratio,  and 2/4 cylinder deactivation. Cooled EGR offered the potential for
additional knock mitigation, increased compression ratio, and reduced pumping losses. The use
of cylinder deactivation held potential for additional pumping loss reduction under light-load
conditions. Initially, the potential for improvements was studied using 1-D gas dynamics/0-D
combustion simulation software.ss  A 2.0L Mazda SKYACTIV-G GDI Atkinson Cycle engine
was thoroughly benchmarked by EPA with the engine dynamometer test facilities at the EPA-
NVFEL laboratory in Ann Arbor, MI. Performance data and physical dimensions for the engine
and its gas exchange and combustion processes were used to build and validate the simulation.
Details of the study, including methods used to build the engine model, model validation and
initial engine modeling results are provided in Lee et al. 2016.52° Simulation results show
potential for an approximately 3 percent to 9 percent incremental effectiveness in areas of
operation of importance for the regulatory drive cycles using a combination of cooled EGR and a
1-point increase in  compression ratio (14:1), with the largest improvements (6 to 9 percent
incremental) occurring between 4-bar and 8-bar BMEP.
 ; Gamma Technologies "GT-Power."
                                             5-280

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
         - 16

         " 140

          120

          100
              1000    2000   3000   4000   5000   6000
                        Engine Speed (rpm)
1000   2000   3000   4000   5000   6000
           Engine Speed (rpm)
 Figure 5.98 Comparison of a 2.0L Mazda SKYACTIV-G engine with a 13:1 geometric compression ratio to
   engine simulation results of a comparable engine with a 1-point increase in geometric compression ratio
                               (14:1) and cooled, low-pressure EGR.TT

   Simulation results show potential for an approximately 3 percent to 12 percent incremental
effectiveness in areas of engine operation with significant importance for the regulatory drive
cycles using a combination of cooled EGR, a 1-point increase in compression ratio (14:1), and
with fixed (2-cylinder) cylinder deactivation below  5-bar BMEP and for engine speeds of 1000
rpm to 3000 rpm.  Simulation results also show an incremental effectiveness of approximately 3
percent to 7 percent when comparing the cooled EGR/higher geometric compression ratio results
with and without cylinder deactivation. This is consistent with other published results for both
production and proof-of-concept fixed (not dynamic) cylinder deactivation.525'526'527
                    2000   3000   4000   5000
                        Engine Speed (rpm)
                                                           \\
      2000    3000   4000   5000
           Engine Speed (rpm)
 Figure 5.99 Comparison of a 2.0L Mazda SKYACTIV-G engine with a 13:1 geometric compression ratio to
   engine simulation results of a comparable engine with a 1-point increase in geometric compression ratio
   (14:1), cooled, low-pressure EGR and cylinder deactivation with operation on 2 cylinders at below 5-bar
                                   BMEP and 1000 - 3000 rpm.
 T The simulation results presented in Figure 5.98 and Figure 5.99 include kinetic knock modeling and calibration of
  the simulation to knock induction comparable to the original engine configuration for both Tier 2 certification test
  fuel (EO, 96 RON) and LEV III certification test fuel (E10, 88 AKI, 91 RON).  An adequate representation of
  knock-limited torque within an engine simulation requires careful experimental validation of the kinetic knock
  model used by the simulation, which is currently under way at EPA-NVFEL. While the simulation results show
  comparable WOT torque between the different engine configurations, experimental validation of the achievable
  knock-limited torque at WOT was still underway at the time of publication of this assessment.
                                                 5-281

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The EPA internal study on Atkinson Cycle engines has entered a second phase involving
engine dynamometer validation of the simulation results using a EU-market version of the
Mazda SKYACTIV-G engine with increased geometric compression ratio (14:1), a proof-of
concept, low-pressure-loop, cooled EGR system, and the use of a dual-coil offset (DCO) ignition
system to improve EGR tolerance of the engine (see Figure 5.100).528'529 Initial results have
been promising.  The improved ignition characteristics of the DCO ignition system has allowed
an increase in the range of part-load engine operation at relatively high rates (approximately 20
percent) of cooled EGR beyond that of the relatively conservative, fixed EGR map used in the
simulation study. This allowed further reductions in part-load pumping losses while maintaining
a COV of EVIEPUU of less than 3-4 percent, which is comparable to that of the original engine
configuration.
 Figure 5.100 Mazda 2.0L SKYACTIV-G engine with 14:1 geometric compression ratio, cooled low-pressure
   external EGR system, DCO ignition system, and developmental engine management system undergoing
            engine dynamometer testing at the U.S. EPA-NVFEL facility in Ann Arbor, ML

   Future work will include validation of the engine model, particularly the kinetic knock model,
and proof-of-concept dynamometer testing of fixed cylinder deactivation of cylinder numbers 2
and 3.  Costs for this technology (future non-HEV Atkinson cycle, referred to as Atkinson-level
2 by EPA) are new as they were not part of the 2012 FRM. We have based our Atkinson-2
technology costs on the 2015 NAS report. Table S.2 of that report shows the cost estimates
presented below. Note that the NAS costs include the costs of gasoline direct injection (shown as
"DI" in the NAS report row header). EPA has removed those costs (using the NAS reported
values) since EPA accounts for those costs separately rather than including them in the Atkinson-
  Coefficient of variation of indicated mean effective pressure based on high-speed in-cy Under pressure
  measurements. This is a commonly used indicator of combustion instability and would typically be kept to values
  that are under 3% to 5% depending on operating conditions and engine application.
                                             5-282

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
2 costs. Note also that EPA always includes costs for direct injection, along with variable valve
timing and other costs when building an Atkinson-2 package.
           Table 5.61 Direct Manufacturing Costs (DMC) for Atkinson-2 Technology (2010$)
Tech



Stoichiometric Gasoline Direct Injection (NAS 2015)

Compression Ratio Increase (CR~13.1, exh. Scavenging, Dl (e.g.
SKYACTIV-G)) (NAS 2015)
EPA estimate (Row 2 minus Row 1)
Midsize
Car
14 DOHC

164

250

86
Large
Car
V6
DOHC
246

375

129
Large Light
Truck
V8OHV

296

500

204
Relative to



Previous
tech
Baseline

Stoich GDI
   Consistent with the NAS report, we have considered the NAS costs to be 2025 costs in terms
of 2010$. Adjusting to 2013$, applying a learning curve (22) that bases that cost in MY2025,
and applying medium 2 level complexity in calculating indirect costs results in the costs
presented below for each engine type in this Draft TAR analysis.
 Table 5.62  Costs for Atkinson-2 Technology, Exclusive of Enablers such as Direct Inject and Valve Timing
                             Technologies (dollar values in 2013$)
Engine
13
14
V6
V8
13
14
V6
V8
13
14
V6
V8
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
DMC: base year
cost
1C: complexity
$91
$91
$136
$215
Med2
Med2
Med2
Med2




DMC: learning
curve
1C: near term
thru
22
22
22
22
2024
2024
2024
2024




2017
$107
$107
$160
$253
$36
$36
$54
$85
$142
$142
$214
$338
2018
$104
$104
$157
$248
$36
$36
$54
$85
$140
$140
$210
$333
2019
$102
$102
$154
$243
$36
$36
$53
$85
$138
$138
$207
$327
2020
$100
$100
$150
$238
$36
$36
$53
$84
$136
$136
$204
$322
2021
$98
$98
$147
$233
$35
$35
$53
$84
$134
$134
$201
$317
2022
$96
$96
$145
$229
$35
$35
$53
$84
$132
$132
$198
$312
2023
$94
$94
$142
$224
$35
$35
$53
$84
$130
$130
$195
$308
2024
$93
$93
$139
$219
$35
$35
$53
$83
$128
$128
$192
$303
2025
$91
$91
$136
$215
$26
$26
$39
$62
$117
$117
$175
$277
       Note: DMC=direct manufacturing cost; IC=indirect cost; TC=total cost.
5.3.4.1.9
GDI, Turbochargins, Downsizing
   The TDS24 and TDS27 configurations used by EPA within the FRM analysis were originally
developed as part of engine and vehicle simulation work conducted by Ricardo, Inc. and SRA
Corporation under contract with EPA, hereto referred in the Draft TAR as the "Ricardo
Study."530 In recent years, Ricardo has developed a number of turbocharged and downsized
engine concepts with a number of characteristics in common 531>532>533>534

       •  Gasoline direct injection (GDI)
       •  Dual camshaft phasing and, in some cases, discrete variable valve lift
                                              5-283

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
       •   Relatively high boost and subsequently high levels of BMEP (over 30-bar in some
          cases)
       •   Cooled, external EGR
       •   Advanced turbocharger boosting systems


   Fuel mapping for different engine technologies was developed by Ricardo within the Study
using a combination of dynamometer test results, ID gas dynamics/OD combustion modeling,
application of correction factors for displacement scaling, and use of engineering judgment. The
development of fuel  maps for turbocharged GDI engines within the Ricardo Study began with
BSFC data obtained from Ricardo's EBDI engine development program.531 Specifications for
this engine are shown in Table 5.63 and a contour plot of BSFC versus engine speed and BMEP
is shown in Figure 5.101.
    Table 5.63 Specification of Ricardo 3.2L V6 Turbocharged, GDI "EBDI" Proof-of-concept Engine.
Base Engine
Swept Volume
Max Power @ 5,000 rpm
Max Torque @ 3,000 rpm
Target Max BMEP
Compression Ratio
Maximum Cylinder
Cam Phaser Authority
Intake Boosting System
Transient Torque Response Time
Prototype V6 with IEM
3190cc
450 hp on E85, 400 hp on 98 RON gasoline
900 Nm on E85, 775 Nm on 98 RON gasoline
35 bar on E85, 30 bar on Indolene (98 RON)
10.0:1
180 bar
50 degCA
Twin, sequential turbochargers with charge air cooling
after each boosting stage
<1.5s to 90% SS torque at 1,500 rpm
<1.0s to 90% SS torque at 2,000 rpm
                                            5-284

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment


                                                                                 450

                                                                                 400

                                                                                 350

                                                                                 300

                                                                                 260

                                                                                 250

                                                                                 245

                                                                                 240

                                                                                 230

                                                                                 220
      1000
2000           3000           4000
            Engine Speed (RPM)
5000
  Figure 5.101 Contour plot of BSFC in g/kW-hr versus engine speed and BMEP for the Ricardo "EBDI"
engine equipped with sequential turbocharging, DCP, DWL, cEGR, IEM, and with a 10:1 compression ratio
                                  using 98 RON Indolene.

   Although not captured within this map, Cruff et al.  show performance data up to 30-bar
BMEP with this engine configuration.
   Technical direction from EPA included a peak BMEP limit of 27-bar, which obviated the
necessity for some of the reciprocating assembly measures taken with the EBDI engine.  Taking
into account the capabilities of the combustion system, valvetrain configuration, EGR system,
and reduced BMEP levels, Ricardo recommended a small increase in compression ratio  (from
10:1 to 10.5:1) while maintaining protection for in-use fuel octanes of approximately 91 RON
(e.g. 87 AKIE10).  All fuel consumption results developed in the Ricardo Study assumed use of
U.S. Certification Gasoline (95 RON, EO). A fuel consumption improvement of 3.5 percent was
also applied to account for continued application of friction reduction from a combination of
technology advances, including piston ring-pack improvements, bore finish improvements, low-
friction coatings, improved valvetrain components, bearings improvements, and lower-viscosity
crankcase lubricants. BMEP levels were held approximately constant for particular classes of
engines within EPA's FRM analyses and analyses for the Draft TAR. A BSFC  correction was
applied as engine displacements were changed within an engine class in the Ricardo Study to
account for different vehicle applications. This correction was predominantly a correction of
thermal losses relative to combustion system surface-to-volume ratio and is expressed within the
displacement correction shown in Figure 5.102. Boosting requirements over the reduced
operational range for TDS24 (up to 24-bar BMEP) were assumed to be achievable using a VNT
within EPA's analyses for the FRM and the Draft TAR.  Sequential turbocharging was
maintained for TDS27 within EPA analyses for the FRM, but TDS27 was not included within
the analyses for the Draft TAR.
                                            5-285

-------
                             Technology Cost, Effectiveness, and Lead-Time Assessment
BSFC Increase [%]
BSFC Increase with Displacement Ratio
10 0°'
9.0%
8 0% -
7 n% -
6 0% -
5 0%
4.0%
3.0%
") n%
1.0%
n fw
\
\
\
\
\

\^
""-•V^
""^^^
^"^^^
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Displacement Ratio (NewlBaseline)
Figure 5.102 BSFC Multiplier Used For Scaling Engine Maps In The Ricardo Study Based On The Ratio:
                                  Displacement[New}
                                 Displacement[Base[ine]
                                            5-286

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
  BtardolEBCH 3.21VB, DCP.DWJ.. CEGR EM 1C 1 CR
  1000
         2000   3000    4000
           Engine Speed (RPM)
                            SC'OO
                                                                  ERft TDS27 2 34L V6
                                                     -
                                                                                           3QKW
                                                                                           ISkW
                                                          •oca
                                                           20M    3000   4000
                                                              Engine Spe&d IRPIV)
                                                                   BMEP reductd to S* bar
                                                                   Eooirw clung?!
                                                                    - tt rf ry reducxid
                                                                    D5pl pe? Ql reduced
                                                                                    5CDC
                 EWTDS24 t 16LI3
                                                                  EPATDS24 1.51LI4
  ,-
  -•'
  -•:-.
  IS
  16

I1"
*
200
        1000
               2000   3000    4000   5000
                    Engine Speed (RPM)
                                         60CO
                                                         1C DC
                                                          2000   3000    4MO   5000
                                                                Engine Speed fRPM)
                                                                                         6COO
                                        BMEP maintained
                                        Engine changes
                                          -#of cyl. reduced
                                          - Displ. per cyl. reduced
         Figure 5.103  Schematic Representation of the Development of BSFC Mapping for TDS24

   A graphical example of how BSFC maps were developed for varying displacements of
TDS24 are shown in Figure 5.103. The brake thermal efficiency (BTE) of the modeled and
                                                 5-287

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
corrected TDS24 engine maps are compared to contemporary turbocharged engines in Figure
5.104 through Figure 5.106.535,536,537,538,539 The Honcja 1.5L turbocharged GDI engine achieves
higher peak break thermal efficiency than TDS24, and has a larger area of operation above 35
percent BTE. TDS24 had improved efficiency at low-speed, light load conditions, possibly from
pumping loss improvements due to the use of discrete variable valve lift and cooled external
EGR. The 2017 VW EA211 TSIEVO engine appears to have a broader area of operation above
34 percent BTE than TDS 24 and the BTE reported at 2-bar, 2000 rpm of 30 percent is higher
than the corresponding operational point with TDS24. The coarseness of published BTE map for
the VW EA211 precludes further comparison.  The larger 2.0L VW EA888-3B engine was
compared  with a 1.51L variant of TDS24. The VW EA888 had a significantly larger area of
operation above 35 percent BTE.  Once again, TDS24 had improved efficiency at low-speed,
light load conditions; possibly due to pumping loss reduction due to the greater extent of
boosting and displacement downsizing and the use of discrete variable valve lift.  On the whole,
contemporary turbocharged engines can achieve higher peak BTE and high BTE over a broader
range of engine operating conditions than TDS24 modeling results. TDS24 shows improved
BTE at lower speeds and lighter loads. Further development of contemporary turbocharged
engines  from 2017 to 2022, including use of more advanced boosting systems (e.g., VNT or
series sequential turbochargers), engine downsizing to 22-bar BMEP or greater, use of external
cooled EGR, and use of variable valve lift systems would further improve low-speed, light load
pumping losses and allow such engines to meet or exceed the BTE modeled for TDS24.
    24

    22

    20

    18

    16

   I"
   CL12
                                                                                 15 kW
                                                                                 7.5 kW
                   3000   4000
                    Speed (RPM)
                             5000   6000
3000   4000   5000
   Engine Speed (rpm)
  Figure 5.104 Comparison between a 1.15L 13 version of TDS24 (left)W and the 1.5L turbocharged, GDI
                           engine used in the 2017 Civic (right)WW.

Dark green shading denotes areas of BTE>35%.
vv Adapted from Ricardo Study modeling results.
** Adapted from Wada et al. 2016 and Nakano et al 2016.
                                            5-288

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
                                           15 kW
                                           7.5 kW
IO
16

14

— 12
m
oTlO
LU

OJ g

6

4

2
n

200
180
.
160
' 140

. 120

100

±80

g-60

K40
20


/ / \
/ 34% ^^^
/ / 34%*v
II ^^V













/ / .37% 96kW \^
/ /
I y
I 34%
s^
//
\ /^
"^~~34%-__^
	

27%
27% .30%
_ _, _ - 16 %
       1000
             2000
                   3000    4000
                    Speed (RPM)
                                5000
                                      6000
                                                           1000
                                                                 2000
                               3000    4000
                                Speed (RPM)
                                                                                    5000
                                                                                          6000
  Figure 5.105 Comparison between a 1.15L 13 version of TDS24 (left)xx and the 2017 Golf 1.5L EA211 TSI
                                        EVO EngineYY.

Light-green shading denotes areas of BTE>34%. Dark green shading denotes areas of BTE>35%. The area of
BTE>3 5% for the VW EA211 is not discernable due to the coarseness of the data provided by the originally
published source.
                                                 20r
 514
0.12
Ul
m10
  8 300
    175
    150
   -I?25
160kW

140 kW

120kW

100kW

80 kW

60 kW

40 kW

20 kW
10kW
                   3000   4000
                    Speed (RPM)
              500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500
                             Engine Speed (rpm)
 Figure 5.106 Comparison between a 1.51L 13 version of TDS24 (left)xx and the 2017 Audi A3 2.0L 888-3B
                                        Engine (right)zz.
Dark green shading denotes areas of BTE>35%.
   Since the FRM, a significant amount of new information has become available from
production vehicles, industry data, benchmarking, and simulation to inform the effectiveness of
engine technologies. The most notable changes  from the FRM are the inclusion of non-hybrid
Atkinson engines, Miller Cycle engines, and the reduction in effectiveness of turbocharged
engines due to additional resolution in the ALPHA model. Table 5.64 compares the
effectiveness (percent CCh improvement from the null vehicle) of several FRM and Draft TAR
engine technology packages as used in OMEGA.
xx Adapted from Ricardo Study modeling results.
YY Adapted from Eichler et al. 2016.
zz
  Adapted from Wurms et al. 2015.
                                                5-289

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
          Table 5.64 FRM to Draft TAR Engine Technology Package Effectiveness Comparison
Engine Technology Package
PFIDOHC + VVT
SGDIDOHC + VVT
SGDI DOHC + VVT + DEAC + EFR1
18 Bar BMEP Turbo + SGDI
Atkinson + VVT + SGDI + EFR2
Atkinson + VVT + SGDI + CEGR + EFR2
24 Bar BMEP Turbo + SGDI + CEGR
Miller + SGDI + CEGR
Small Car
FRM - TAR
4.1-4.1
5.6-5.6
10.5-9.9
12.2-10.1
NA-11.7
NA-19.3
19.4-17.2
NA-23.0
Standard
Car
FRM - TAR
5.2-5.2
6.6-6.6
12.8-12.1
14.2 - 11.5
NA-12.9
NA-19.4
22.1-19.1
NA-23.3
Large Car
FRM - TAR
5.5-5.5
6.9-6.9
13.5-12.7
14.9 - 11.9
NA-13.3
NA-19.5
23.0-19.7
NA-23.4
Small MPV
FRM - TAR
4.1-4.1
5.5-5.5
10.4-9.8
12.1-10.0
NA-11.7
NA-19.3
19.3-17.1
NA-23.0
Large MPV
FRM - TAR
5.1-5.1
6.6-6.6
12.8-12.0
14.2 - 11.4
NA-12.9
NA-19.4
22.1-19.1
NA-23.3
Truck
FRM - TAR
4.9-4.9
6.3-6.3
12.1-11.4
13.6-11.1
NA-12.6
NA-19.4
21.3-18.6
NA-23.2
   Costs associated with gasoline direct injection are equivalent to those used in the FRM except
for updates to 2013 dollars and use of a new learning curve (curve 23). The GDI costs
incremental to port-fuel injection for 14, V6 and V8 engines are shown below.

        Table 5.65 Costs for Gasoline Direct Injection on an 13 & 14 Engine (dollar values in 2013$)
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$234
Med2

DMC: learning
curve
1C: near term thru
23
2018

2017


$211
$89
$301
2018


$208
$89
$297
2019


$205
$67
$272
2020


$202
$67
$269
2021


$200
$67
$266
2022


$197
$67
$264
2023


$195
$66
$261
2024


$193
$66
$259
2025


$190
$66
$257
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
          Table 5.66 Costs for Gasoline Direct Injection on a V6 Engine (dollar values in 2013$)
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$352
Med2

DMC: learning
curve
1C: near term thru
23
2018

2017


$319
$135
$454
2018


$314
$135
$448
2019


$309
$101
$410
2020


$305
$101
$405
2021


$301
$100
$401
2022


$297
$100
$397
2023


$293
$100
$394
2024


$290
$100
$390
2025


$287
$100
$387
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
          Table 5.67 Costs for Gasoline Direct Injection on a V8 Engine (dollar values in 2013$)
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$423
Med2

DMC: learning
curve
1C: near term thru
23
2018

2017


$383
$162
$545
2018


$377
$162
$539
2019


$372
$121
$493
2020


$367
$121
$487
2021


$362
$121
$482
2022


$357
$121
$478
2023


$353
$120
$473
2024


$349
$120
$469
2025


$345
$120
$465
   Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
   Costs associated with turbocharging are equivalent to those used in the FRM except for three
important updates: (1) we have updated costs to 2013 dollars; and, (2) we are using of a new
learning curve (curve 23); and, (3) we have added $44 (DMC, 2013$) to the costs of 24-bar
turbocharging (and Miller cycle turbocharging) to reflect the use of a variable geometry
                                               5-290

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
turbocharger which was not properly accounted for in the 2012 FRM costs. The turbo costs
incremental to naturally aspirated I-configuration and V-configuration engines are shown below.

      Table 5.68 Costs for Turbocharging, 18/21 bar, I-Configuration Engine (dollar values in 2013$)
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$443
Med2

DMC: learning
curve
1C: near term thru
23
2018

2017


$401
$170
$571
2018


$395
$169
$564
2019


$389
$127
$516
2020


$384
$126
$510
2021


$379
$126
$505
2022


$374
$126
$500
2023


$369
$126
$495
2024


$365
$126
$491
2025


$361
$126
$487
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.

      Table 5.69  Costs for Turbocharging, 18/21 bar, V-Configuration Engine (dollar values in 2013$)
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$747
Med2

DMC: learning
curve
1C: near term thru
23
2018

2017


$676
$286
$962
2018


$666
$286
$951
2019


$656
$213
$869
2020


$647
$213
$860
2021


$638
$213
$851
2022


$630
$213
$843
2023


$623
$212
$835
2024


$616
$212
$828
2025


$609
$212
$821
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
   Table 5.70 Costs for Turbocharging, 24 bar, I-Configu ration Engine & for Miller-cycle I-Configuration
                                  Engine (dollar values in 2013$)
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$708
Med2

DMC: learning
curve
1C: near term thru
23
2024

2017


$641
$271
$913
2018


$631
$271
$902
2019


$622
$271
$893
2020


$613
$270
$884
2021


$605
$270
$875
2022


$598
$269
$867
2023


$591
$269
$860
2024


$584
$269
$853
2025


$578
$201
$779
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
  Table 5.71 Costs for Turbocharging, 24 bar, V-Configuration Engine & for Miller-cycle V-Configuration
                                  Engine (dollar values in 2013$)
Cost
type



DMC
1C
TC
DMC: base
year cost
1C:
complexity

$1,208
Med2

DMC:
learning
curve
1C: near
term thru
23
2024

2017




$1,094
$463
$1,557
2018




$1,077
$462
$1,539
2019




$1,061
$461
$1,522
2020




$1,046
$461
$1,507
2021




$1,032
$460
$1,492
2022




$1,019
$459
$1,479
2023




$1,007
$459
$1,466
2024




$996
$458
$1,454
2025




$985
$343
$1,328
   Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
   Costs associated with engine downsizing are equivalent to those used in the FRM except for
updates to 2013 dollars and use of new learning curves (curve 23 and 28). The downsizing costs
incremental to the baseline engine configuration are shown below.
                                                5-291

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
     Table 5.72 Costs for Downsizing as part of Turbocharging & Downsizing (dollar values in 2013$)
Downsizing from & to
!4DOHCtol3
!4DOHCtol4
V6DOHCtol4
V6SOHCtol4
V6OHVtol4
V8DOHCtoV6
V8SOHC3VtoV6
V8SOHCtoV6
vsoHvtove
!4DOHCtol3
!4DOHCtol4
V6DOHCtol4
V6SOHCtol4
V6OHVtol4
V8DOHCtoV6
V8SOHC3VtoV6
V8SOHCtoV6
vsoHvtove
!4DOHCtol3
!4DOHCtol4
V6DOHCtol4
V6SOHCtol4
V6OHVtol4
V8DOHCtoV6
V8SOHC3VtoV6
V8SOHCtoV6
vsoHvtove
Cost type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
-$212
-$93
-$600
-$419
$296
-$300
-$170
-$92
$345
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2









DMC:
learning
curve
1C: near
term
thru
23
23
23
23
28
23
23
23
28
2018
2018
2018
2018
2018
2018
2018
2018
2018









2017
-$192
-$84
-$543
-$380
$289
-$272
-$154
-$83
$337
$81
$36
$230
$161
$114
$115
$65
$35
$133
-$111
-$49
-$313
-$219
$404
-$157
-$89
-$48
$471
2018
-$189
-$83
-$535
-$374
$283
-$268
-$152
-$82
$330
$81
$36
$229
$160
$114
$115
$65
$35
$133
-$108
-$47
-$305
-$213
$397
-$153
-$87
-$47
$463
2019
-$186
-$82
-$527
-$368
$278
-$264
-$150
-$81
$324
$61
$27
$172
$120
$85
$86
$49
$26
$99
-$125
-$55
-$355
-$248
$363
-$178
-$101
-$54
$423
2020
-$183
-$81
-$520
-$363
$273
-$260
-$147
-$80
$318
$60
$27
$171
$120
$85
$86
$49
$26
$99
-$123
-$54
-$348
-$243
$358
-$174
-$99
-$53
$417
2021
-$181
-$80
-$513
-$358
$268
-$257
-$145
-$78
$313
$60
$27
$171
$120
$85
$86
$49
$26
$99
-$120
-$53
-$342
-$239
$353
-$171
-$97
-$52
$412
2022
-$179
-$79
-$506
-$354
$264
-$253
-$144
-$78
$308
$60
$27
$171
$119
$85
$86
$49
$26
$99
-$118
-$52
-$335
-$234
$349
-$168
-$95
-$51
$407
2023
-$176
-$78
-$500
-$350
$260
-$250
-$142
-$77
$303
$60
$26
$171
$119
$85
$85
$48
$26
$99
-$116
-$51
-$330
-$230
$345
-$165
-$94
-$50
$402
2024
-$174
-$77
-$495
-$346
$256
-$248
-$140
-$76
$299
$60
$26
$171
$119
$85
$85
$48
$26
$99
-$114
-$50
-$324
-$226
$341
-$162
-$92
-$50
$398
2025
-$173
-$76
-$489
-$342
$253
-$245
-$139
-$75
$295
$60
$26
$170
$119
$84
$85
$48
$26
$99
-$112
-$49
-$319
-$223
$337
-$160
-$90
-$49
$394
       Note:
       DMC=direct manufacturing costs; IC=indirect costs; TC=total costs;
       the downsized configuration is always a DOHC.
   Costs associated with turbocharging combined with engine downsizing (TDS) are similarly
equivalent to those used in the FRM except for updates to 2013 dollars and use of new learning
curves (curve 23 and 28). The TDS costs incremental to the baseline engine configuration are
shown below. Note that the costs presented below do not include direct injection costs or other
possible technologies such as cooled EGR. The costs presented are simply the combination of
the above  turbo costs and downsizing costs.
                                             5-292

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                     Table 5.73 Costs for Turbocharging & Downsizing (2013$)
Turbo
TURB18-I
TURB18-I
TURB18-I
TURB18-I
TURB18-I
TURB18-V
TURB18-V
TURB18-V
TURB18-V
TURB24-I
TURB24-I
TURB24-I
TURB24-I
TURB24-I
TURB24-V
TURB24-V
TURB24-V
TURB24-V
Downsize
14 to 13
!4DOHCtol4
V6DOHCtol4
V6SOHCtol4
V6OHVtol4
V8DOHCtoV6
V8SOHC3VtoV6
V8SOHCtoV6
vsoHvtove
14 to 13
!4DOHCtol4
V6DOHCtol4
V6SOHCtol4
V6OHVtol4
V8DOHCtoV6
V8SOHC3VtoV6
V8SOHCtoV6
vsoHvtove

TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
2017
$460
$522
$257
$352
$974
$805
$873
$914
$1,433
$802
$864
$599
$694
$1,316
$1,400
$1,468
$1,509
$2,027
2018
$457
$517
$259
$351
$961
$798
$864
$904
$1,414
$795
$855
$597
$689
$1,300
$1,386
$1,452
$1,492
$2,002
2019
$390
$461
$160
$267
$878
$691
$768
$815
$1,292
$767
$838
$537
$644
$1,255
$1,344
$1,421
$1,468
$1,945
2020
$387
$456
$162
$267
$868
$685
$761
$806
$1,277
$761
$830
$535
$640
$1,241
$1,332
$1,408
$1,453
$1,924
2021
$384
$452
$163
$266
$858
$680
$754
$799
$1,263
$755
$822
$534
$636
$1,228
$1,321
$1,395
$1,440
$1,904
2022
$382
$448
$165
$266
$849
$675
$748
$791
$1,250
$749
$815
$532
$633
$1,216
$1,311
$1,384
$1,427
$1,886
2023
$379
$444
$166
$265
$840
$670
$742
$785
$1,237
$744
$809
$530
$630
$1,204
$1,301
$1,373
$1,416
$1,868
2024
$377
$441
$167
$265
$832
$666
$736
$778
$1,226
$739
$803
$529
$626
$1,194
$1,292
$1,362
$1,405
$1,852
2025
$375
$438
$168
$264
$824
$661
$730
$772
$1,215
$666
$729
$460
$556
$1,116
$1,169
$1,238
$1,279
$1,722
              Note: TC=total costs; the downsized configuration is always a DOHC.
   Costs associated with turbocharging combined with Atkinson-2 technology (i.e., Miller-cycle)
are presented below. Note that the costs presented below do not include direct injection costs or
other required technologies such as cooled EGR. The costs presented are simply the combination
of the above turbo costs and Atkinson-2 costs presented in Section 5.3.4.1.8. Note also that the
ATK2 engine as shown in the table is always a DOHC configuration engine so also not included
in the table are the costs associated with converting, for example, a SOHC or OHV engine to a
DOHC configuration. Those costs are presented below following the cooled EGR costs.
                            Table 5.74 Costs for Miller Cycle (2013$)
Turbo
TURB24-I
TURB24-I
TURB24-V
TURB24-V
ATK2 engine
13
14
V6
V8

TC
TC
TC
TC
2017
$1,055
$1,055
$1,770
$1,894
2018
$1,043
$1,043
$1,749
$1,871
2019
$1,031
$1,031
$1,729
$1,849
2020
$1,019
$1,019
$1,710
$1,829
2021
$1,009
$1,009
$1,693
$1,810
2022
$999
$999
$1,676
$1,791
2023
$990
$990
$1,661
$1,774
2024
$981
$981
$1,646
$1,757
2025
$896
$896
$1,504
$1,606
       Note: TC=total costs; the downsized configuration is always a DOHC.
   Costs associated with cooled EGR are equivalent to those used in the FRM except for updates
to 2013 dollars and use of new learning curve (curve 23). The cooled EGR costs incremental to
the baseline engine configuration are shown below.
                     Table 5.75 Costs for Cooled EGR (dollar values in 2013$)
Cost type
DMC
1C
TC
DMC: base year cost
1C: complexity
$258
Med2

DMC: learning curve
1C: near term thru
23
2024

2017
$233
$99
$332
2018
$230
$98
$328
2019
$226
$98
$325
2020
$223
$98
$321
2021
$220
$98
$318
2022
$217
$98
$315
2023
$215
$98
$313
2024
$212
$98
$310
2025
$210
$73
$283
              Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
   Costs associated with converting non-DOHC engines to a DOHC configuration without any
engine downsizing are equivalent to those used in the 2012 FRM except for updates to 2013$
and use of new learning curves (curves 23 and 28). These costs are used when converting a non-
DOHC engine to a DOHC configuration when downsizing is not also included. The primary
example for this Draft TAR analysis is converting to a DOHC  configuration to enable Atkinson-
                                              5-293

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
2 technology. The costs are presented below and do not include other potential technologies such
as variable valve timing or lift or cylinder deactivation, all of which are accounted for separately
by EPA.
     Table 5.76 Costs for Valvetrain Conversions from non-DOHC to DOHC (dollar values in 2013$)
Conversion
V6 SOHC to V6
DOHC
veoHvtove
DOHC
V8SOHC3VtoV8
DOHC
VSSOHCtoVS
DOHC
VSOHVtoVS
DOHC
V6 SOHC to V6
DOHC
veoHvtove
DOHC
V8SOHC3VtoV8
DOHC
VSSOHCtoVS
DOHC
VSOHVtoVS
DOHC
V6 SOHC to V6
DOHC
veoHvtove
DOHC
VSSOHCSVtoVS
DOHC
VSSOHCtoVS
DOHC
VSOHVtoVS
DOHC
Cost
type
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
$181
$518
$130
$208
$568
Med2
Med2
Med2
Med2
Med2





DMC:
learning
curve
1C: near
term
thru
23
28
23
23
28
2018
2018
2018
2018
2018





2017
$164
$506
$118
$189
$554
$69
$200
$50
$80
$219
$233
$706
$168
$269
$774
2018
$161
$496
$116
$186
$543
$69
$200
$50
$80
$219
$230
$695
$166
$266
$761
2019
$159
$486
$114
$183
$532
$52
$149
$37
$60
$163
$210
$635
$151
$243
$696
2020
$157
$477
$113
$181
$523
$52
$149
$37
$60
$163
$208
$626
$150
$240
$686
2021
$154
$469
$111
$178
$514
$52
$149
$37
$59
$163
$206
$618
$148
$238
$677
2022
$153
$462
$110
$176
$506
$52
$148
$37
$59
$162
$204
$610
$147
$235
$668
2023
$151
$455
$108
$174
$498
$51
$148
$37
$59
$162
$202
$603
$145
$233
$661
2024
$149
$449
$107
$172
$491
$51
$148
$37
$59
$162
$200
$597
$144
$231
$653
2025
$147
$443
$106
$170
$485
$51
$148
$37
$59
$162
$199
$591
$143
$229
$647
       Note:
       DMC=direct manufacturing costs; IC=indirect costs; TC=total costs;
       the downsized configuration is always a DOHC.
5.3.4.2 Transmissions: Data and Assumptions for this Assessment

   In assessing the effectiveness of transmission technology, EPA used multiple data sources.
These data sources include benchmarking activities, conducted at both the National Vehicle and
Fuel Emissions Lab (NVFEL) in Ann Arbor, Michigan and through contract work, technical
literature, technical conferences, vehicle certification data and stakeholder meetings.  To ensure
the data were consistent, it was important to understand the assumptions made in determination
                                              5-294

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
of the effectiveness. It is also important to note the engine with which the transmission is being
paired.  Since much of the effectiveness associated with advanced transmissions is in the
transmission's ability to alter the operation range of the engine, and thus minimize pumping
losses, the engine efficiency in the area of operation is a major part of the effectiveness
calculation. The National Academy of Science, in their 2015 report, noted that "as engines
incorporate new technologies to improve fuel consumption, including variable valve timing and
lift, direct injection, and turbocharging and downsizing, the benefits of increasing transmission
ratios or switching to a CVT diminish."540 This is not to say that transmissions are not an
important technology going forward, but rather a recognition that future engines will have larger
"islands" of low fuel consumption that potentially rely less on the transmission to improve the
overall efficiency of the vehicle.  Thus, effectiveness percentages reported for transmissions
paired with unimproved engines would be expected to be reduced when the same transmission is
paired with a more advanced engine.  Regardless of the engine with which a particular
transmission is mated, it is expected that vehicle manufacturers and suppliers  will continue to
improve the overall efficiency of the transmission itself by reducing friction and parasitic losses.

   This  approach to effectiveness calculation is consistent with the approach used in the analysis
contained in the FRM, and with EPA's lumped parameter model (LPM) in use during the
rulemaking. For example, in the  LPM, an advanced eight-speed AT (with optimized shift logic,
TC lockup, and high efficiency gearbox level 1) on a standard car had an effectiveness of 13.4
percent  when paired with a null engine.  When paired with an  improved PFI engine (with dual
cam phasing and engine friction reduction), the same transmission had an effectiveness of 11.7
percent. With a more advanced GDI engine (adding GDI, low friction lubrication, and more
engine friction reduction), the effectiveness was 11.1 percent.  Finally, with a turbo-downsized
engine with EGR, the transmission effectiveness was 8.6 percent.  Table 5.77 puts this example
in table  form.
                            Table 5.77 Standard Car Effectiveness
Engine Level
Null
Improved PFI Engine (with dual cam phasing and engine friction reduction)
Advanced GDI Engine (adding GDI, low friction lubrication, and more
engine friction reduction)
Turbo-Downsized Engine with EGR
Effectiveness for an
Advanced Eight-Speed AT
(with optimized shift logic,
TC lockup, and high
efficiency gearbox level 1)
13.4
11.7
11.1
8.6
5.3.4.2.1     Assessment of Automated Transmissions (AT, AMT, DCT, CVT)

   For this Draft TAR, EPA is assessing the baseline fleet in the following manner (MY2014):

       1) All manufacturers have incorporated some level of early torque converter lockup, as
          well as an appropriate level of advanced shift logic, into automatic transmissions with
          six speeds and above.
                                             5-295

-------
                       Technology Cost, Effectiveness, and Lead-Time Assessment
 2)  All manufacturers have incorporated some level gear of box efficiency improvements
    (called out in the FRM as "high efficiency gearbox" or HEG), and advanced shift
    logic (called out in the FRM as "advanced shift logic" or ASL) into automatic
    transmissions with six speeds and above.
 3)  All types of automated transmissions will improve between now and 2025 MY.  EPA
    expects that similar gains in efficiency can be made, independent of the transmission
    type. Figure  5.107 shows that all three of the main transmission types moving across
    their respective paths toward their ultimate level of efficiency. The term "Flexibility"
    here is denotes how well the transmission can keep the engine on its optimal
    efficiency line.

  Fuel  Economy improvement	1-lntroduction

s Transmission's performance potential can be expressed  in two-
   dimensional map, transmitting 'Efficiency' and ratio 'Flexibility'

                                                           Ultimate
                                                           2-Pedal
                                                           Transmission
                x
                0)
                                 Efficiency
          8th International CTI Symposium North America 2014, Rochester
            _ M Nakasakj Ja,co Ltd and Y. Qota. NISSAN Motor Co., Ltd. -
                                                     NISSAN MOTOR CORPORATION
                                                                      =CTi
                                                                     Car Turning Institute
            Figure 5.107 Comparison of the Different Transmission Types
 4)
          The incremental effectiveness and cost for all automated transmissions are based on
          data from conventional automatics.
   EPA does not believe that the technologies represented by FIEG and ASL have been
incorporated into all transmissions in the 2014 fleet, but are presumed included to be in both the
base 6- speed and 8-speed transmission (higher-gear transmissions) in the 2014 fleet.

   Under the premise that automated transmissions that are currently in the fleet demonstrate
different effectiveness, and with the expectation that all automated transmissions will be
improved between now and 2025 MY, 2014 transmissions were mapped to three different
designations Null, TRX11 and TRX21.  Table 5.78 shows the mapping between the existing
                                     5-296

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
transmissions in the 2014 baseline fleet and the transmission designations that have been
established for this Draft TAR analysis.  Note that manual transmission were left alone unless
the vehicle was determined to need electrification in order to comply in which case it would be
upgraded to either a hybrid or electric vehicle transmission. In the "TRX" numbering system the
first digit specifies the number of gears in the transmission and the second digit specifies the
HEG level. A "1" in the first digit represents an 8-speed transmission and a "2" in the first digit
represents an 8-speed. Similarly, a "1" in the second digit represents HEG1 and a "2" in the
second digit represents HEG2.  An important aspect of using the TRX system is that it meant to
estimate the effectiveness of both the current transmission technology and future transmission
technology. This is appropriate because it allows EPA to account for technology already found
in the baseline fleet, as well as apply future transmission technology as a means of improving
vehicle efficiency.  With the predominant transmission type in the 2014 MY baseline fleet (73.8
percent) being a conventional automatic transmission, EPA believes that this approach most
closely approximates the overall incremental effectiveness and cost associated with all
automated transmissions. In the future, if a particular transmission technology develops in such
a way that it becomes more  cost effective compared to our estimates, and it demonstrates the
capability of meeting vehicle functional  objectives, EPA expects that vehicle manufacturers may
adopt that technology instead.
                              Table 5.78  Transmission Level Map
Trans code from Data
A
A
A
A
A
A
AM
AM
AM
C
D
D
Transmission Type
Automatic
Automatic
Automatic
Automatic
Automatic
Automatic
Automated Manual
Automated Manual
Automated Manual
CVT
Dual Clutch
Dual Clutch
Number of Gears
4
5
6
7
8
9
5
6
7
0
6
7
Transmission Level
Null
Null
TRX11
TRX21
TRX21
TRX21
Null
TRX11
TRX21
TRX11
TRX11
TRX21
   The effectiveness associated with TRX11 is based on a benchmarked GM six-speed
transmission from the 2013 Malibu541.  The expectation is that transmission mapped to the
TRX11 can still be improved to a level that that would bring the transmission effectiveness to the
efficiency level of the TRX22 (with effectiveness based on a ZF 8 speed with HEG level 2).
Table 5.79 shows the effectiveness of a TRX11 level transmission vs. the null transmission on
the different vehicle types with a null engine. Table 5.79 also shows the effect of adding HEG
level 2 to the GM 6 speed giving us TRX12.
                                             5-297

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                     Table 5.79 TRX11 and TRX12 Null Engine Effectiveness
Vehicle Type
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Transmission Level
TRX11 (HEG1)
5.9
7.3
7.5
6.1
7.1
5.5
TRX12 (HEG2)
9.9
11.9
11.9
10.6
11.3
9.4
   The effectiveness of TRX21 is based on the benchmarked 845REeight-speed transmission (a
ZF licensed FCA clone) from the 2014 Dodge Ram542. The expectation is that transmission
mapped to the TRX21 can be improved to a level that that would bring the transmission
effectiveness to the efficiency level of the TRX22 (ZF 8 speed with HEG level 2). Table 5.80
shows the effectiveness of a TRX21 level transmission vs. the null transmission on the different
vehicle types with a null engine. Table 5.80 also shows TRX22 the effect of adding FLEG level 2
to the ZF which was modeled using EPA's ALPHA model based on information in various  SAE
papers from ZF describing how they intend to create a future higher efficiency version of their
current 8 speed transmission.
                     Table 5.80 TRX21 and TRX22 Null Engine Effectiveness
Vehicle Type
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Transmission Level
TRX21 (HEG1)
11.5
13.4
13.2
12.3
12.7
12.8
TRX22 (HEG2)
14
16.3
15.9
15.1
15.4
15.2
   The aggregation of effectiveness values represents the best data available to EPA for the Draft
TAR analysis.  EPA plans on performing extensive CVT benchmarking and a cost tear-down in
support of the Proposed Determination.  EPA feels that these effectiveness values are appropriate
since it allows a maximum of 9.7 percent improvement in effectiveness from a TRX11 to a
TRX22. A 9.7 percent improvement in effectiveness is achievable given that most transmission
can gain 6-10 percent from efficiency improvements alone, and designs for increased gear counts
and wider ratio spans  from 8-10 are expected.

   Currently available CVT transmissions in the 2014 MY baseline fleet have been characterized
as TRX11 level transmissions.  However, a limitation was added to vehicles with CVT
transmissions that prevented the transmissions from being improved to the TRX22 level.
                                            5-298

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
Vehicles with CVTs can increase to TRX21 which is about a 6 percent effectiveness
improvement.  Most CVT transmissions are 85 percent efficient and are expected to be 90-94
percent efficient by 2025. They are also expected to have their ratio span increase from the
current 6-7.3 to between 8 and 8.5.

   Effectiveness for all transmission types will be evaluated after the Draft TAR as more data is
available from the ALPHA model.
5.3.4.2.2
Technology Applicability and Costs
   For future vehicles, it was assumed that the costs for transitioning from one technology level
(TRX11-TRX22) to another level is the same for each transmission type (AT, AMT, DCT, and
CVT). The costs used are based on AT transmission which make up 73.8 percent of
transmissions in the 2014 fleet. This is a reasonable approach based on the costs used in the
FRM for the different transmission types.

   Transmission  technology costs are presented in Table 5.81.

        Table 5.81  Costs for Transmission Improvements for all Vehicles (dollar values in 2013$)
Tech
TRX11
TRX12
TRX21
TRX22
TRX11
TRX12
TRX21
TRX22
TRX11
TRX12
TRX21
TRX22
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
DMC: base
cost
1C:
complexity
$39
$252
$171
$384
Low2
Low2
Low2
Low2




DMC: learning
curve
1C: near term
thru
23
23
23
23
2018
2018
2024
2024




2017
$35
$228
$155
$348
$17
$111
$75
$169
$52
$339
$230
$516
2018
$35
$225
$152
$342
$17
$110
$74
$167
$52
$335
$227
$510
2019
$34
$222
$150
$337
$14
$89
$74
$166
$48
$310
$224
$504
2020
$34
$218
$148
$333
$14
$88
$73
$165
$47
$307
$221
$498
2021
$33
$216
$146
$328
$13
$88
$73
$164
$47
$303
$219
$492
2022
$33
$213
$144
$324
$13
$87
$72
$163
$46
$300
$217
$487
2023
$32
$210
$142
$320
$13
$87
$72
$162
$46
$297
$214
$483
2024
$32
$208
$141
$317
$13
$86
$72
$161
$45
$294
$212
$478
2025
$32
$206
$139
$313
$13
$86
$58
$131
$45
$291
$197
$444
   Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
   As a comparison to how the Draft TAR transmission, or TRX, costs presented above would
compare to the transmission costs EPA used in the FRM, see the table below. To construct this
table, EPA has added various FRM transmission technologies (updated to 2013$) together on a
year-over-year basis and presented them along with the conceptual intent behind the new TRX
structure discussed above. Note that the FRM costs were presented in 2010$ and, importantly,
EPA revised the FRM transmission costs in 2013 due to FEV-generated updates to the tear down
costs used in the 2012 FRM.543The FRM costs presented in the table below reflect the updates
made to the FRM costs by FEV. We present the updated values rather than the actual FRM
values  since the updated values, if they were being used in this Draft TAR analysis, are the
values  we would have used.
  Table 5.82 Comparison of Transmission Costs Using the 2012 FRM Methodology to Draft TAR Costs for
                                   Transmissions (2013$)
               Tech
              Cost type | 2017 | 2018 | 2019 | 2020 | 20211 2022 | 2023 | 2024 | 2025
                                             5-299

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
6sp DCT-dry+ASL2+HEGl
6sp DCT-wet+ASL2+HEGl
6spAT+ASL2+HEGl
TRX11

6sp DCT-dry+ASL2+HEG2
6sp DCT-wet+ASL2+HEG2
6spAT+ASL2+HEG2
TRX12

8sp DCT-dry+ASL2+HEGl
8sp DCT-wet+ASL2+HEGl
8spAT+ASL2+HEGl
TRX21

8sp DCT-dry+ASL2+HEG2
8sp DCT-wet+ASL2+HEG2
8spAT+ASL2+HEG2
TRX22
TC
TC
TC
TC

TC
TC
TC
TC

TC
TC
TC
TC

TC
TC
TC
TC
-$68
-$29
$24
$52

$192
$231
$285
$339

$89
$185
$120
$230

$349
$445
$380
$516
-$66
-$28
$24
$52

$190
$228
$280
$335

$88
$182
$119
$227

$344
$438
$374
$510
-$83
-$39
$23
$48

$169
$213
$275
$310

$87
$180
$111
$224

$339
$432
$363
$504
-$81
-$38
$23
$47

$167
$210
$271
$307

$86
$178
$110
$221

$334
$426
$358
$498
-$79
-$37
$22
$47

$166
$208
$267
$303

$85
$176
$108
$219

$330
$421
$353
$492
-$77
-$36
$22
$46

$164
$206
$264
$300

$84
$175
$107
$217

$326
$416
$349
$487
-$76
-$35
$22
$46

$163
$204
$261
$297

$83
$173
$106
$214

$322
$412
$345
$483
-$74
-$35
$22
$45

$162
$202
$258
$294

$82
$172
$105
$212

$319
$408
$341
$478
-$75
-$35
$20
$45

$149
$188
$243
$291

$76
$158
$103
$197

$300
$381
$326
$444
5.3.4.3 Electrification: Data and Assumptions for this Assessment

   As in the 2012 FRM analysis, this Draft TAR GHG assessment relies on estimates of cost and
effectiveness of each GHG-reducing technology in order to project its expected role in fleet
compliance with the standards. Electrification technologies represent a particularly broad range
of cost and effectiveness, ranging from relatively low-cost technologies offering incremental
degrees of effectiveness, such as stop-start and mild hybrids, to higher-cost, highly effective
technologies such as plug-in hybrids and pure electric vehicles.

   In this analysis, the costs associated with electrification are divided into battery and non-
battery costs.  The agencies' joint Section 5.2 reviewed industry developments in battery and
non-battery technology since the 2012 FRM As anticipated in the FRM, many of these
developments have resulted in cost reductions for both battery and non-battery components as
the industry has gained in experience and production scale.  For this Draft TAR analysis, EPA
has reviewed its 2012 FRM projections of electrification costs for the 2022-2025 time frame, and
revised them based on these developments.

   Also as anticipated in the FRM, many of these developments have resulted in gradual
improvements in effectiveness as the industry has continued to innovate and compete. EPA has
therefore reviewed its FRM projections of electrification effectiveness for the 2022-2025 time
frame, and have revised them based on these developments.
5.3.4.3.1
Cost and Effectiveness for Non-hybrid Stop-Start
   For the 2012 FRM analysis, the agencies' primary reference for effectiveness of stop-start
technology was the Ricardo simulation study. Based on this study the agencies estimated the on-
cycle effectiveness of stop-start technology to be in the range of 1.8 to 2.4 percent, depending on
vehicle class.
                                             5-300

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   As reviewed in Section 5.2, several new implementations of stop-start have been produced,
proposed or described in the literature since the 2012 FRM.  These examples have provided a
much broader picture of the potential effectiveness of stop-start technology. Based in part on
these examples, EPA has chosen to update the effectiveness estimates for stop-start for use in
this Draft TAR analysis to reflect an effectiveness of 3.0 to 4.0 percent depending on vehicle
class, as shown in Table 5.83.
                      Table 5.83 GHG Technology Effectiveness of Stop-Start
Technology
12V Stop-Start - 2012 FRM
12V Stop-Start - Draft TAR
Technology Effectiveness [%]
Small Car
1.8
3.0
Standard
Car
2.1
3.5
Large
Car
2.4
4.0
Small
MPV
2.2
3.7
Large
MPV
2.2
3.7
Small
Truck
1.8
3.0
Large
Truck
2.2
3.7
   We have assumed costs associated with stop-start equivalent to those used in the FRM except
for updates to 2013 dollars and use of new learning curves (curve 25). The costs incremental to
the baseline engine configuration for our different vehicle classes are shown below.
          Table 5.84  Costs for Stop-Start for Different Vehicle Classes (dollar values in 2013$)
Tech
Small car
Standard
car
Large car
Small MPV
Large MPV
Truck
Small car
Standard
car
Large car
Small MPV
Large MPV
Truck
Small car
Standard
car
Large car
Small MPV
Large MPV
Truck
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
DMC: base
cost
1C: complexity
$308
$308
$349
$349
$349
$383
Med2
Med2
Med2
Med2
Med2
Med2






DMC: learning
curve
1C: near term
thru
25
25
25
25
25
25
2018
2018
2018
2018
2018
2018






2017
$260
$260
$294
$294
$294
$323
$117
$117
$133
$133
$133
$146
$377
$377
$427
$427
$427
$469
2018
$246
$246
$279
$279
$279
$306
$116
$116
$132
$132
$132
$145
$362
$362
$411
$411
$411
$451
2019
$235
$235
$267
$267
$267
$293
$87
$87
$99
$99
$99
$108
$322
$322
$365
$365
$365
$401
2020
$227
$227
$257
$257
$257
$282
$87
$87
$98
$98
$98
$108
$313
$313
$355
$355
$355
$389
2021
$219
$219
$248
$248
$248
$273
$86
$86
$98
$98
$98
$107
$306
$306
$346
$346
$346
$380
2022
$213
$213
$241
$241
$241
$265
$86
$86
$98
$98
$98
$107
$299
$299
$339
$339
$339
$372
2023
$208
$208
$235
$235
$235
$258
$86
$86
$98
$98
$98
$107
$294
$294
$333
$333
$333
$365
2024
$203
$203
$230
$230
$230
$252
$86
$86
$97
$97
$97
$107
$289
$289
$327
$327
$327
$359
2025
$198
$198
$225
$225
$225
$247
$86
$86
$97
$97
$97
$107
$284
$284
$322
$322
$322
$354
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.3.2
Cost and Effectiveness for Mild Hybrids
   In the 2012 FRM analysis, the agencies based their cost and effectiveness estimates for mild
hybrid technology on an analysis of BISG technology as exemplified by the General Motors
eAssist. EPA sized the system using a 10 to 15 kW starter/generator and a 0.25 to 0.5 kWh Li-
ion battery pack.  The same effectiveness results were applied by both NHTSA and EPA.  The
absolute effectiveness for the CAFE analysis ranged from 8.5 to 11.6 percent depending on
                                              5-301

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
vehicle subclass. The effectiveness values include technologies that would be expected to be
incorporated with BISG, which are stop-start (MHEV) and improved accessories (IACC1 and
IACC2).  The effectiveness values did not include electric power steering (EPS).

   As reviewed in Section 5.2, several new implementations of mild hybrid technology have
emerged since the 2012 FRM. These examples provide a much broader picture of the potential
effectiveness of mild hybrid technology.

   For this Draft TAR analysis, EPA has updated the assumed BISG configuration to include a
12 kW electric machine. The Lumped Parameter Model estimates that a BISG with 12 kW
electric machine results in a GHG effectiveness estimate of 9.5 percent and 9.4 percent for small
cars and standard (mid-size) cars, respectively.  Based on this result as well as the examples
discussed in Section 5.2, EPA has updated the GHG effectiveness of CISG PI and TISG 48V P2
mild hybrids as shown in Table 5.85.
                    Table 5.85 GHG Technology Effectiveness of Mild Hybrids
Technology
High voltage Mild Hybrid - 2012 FRM
12-15 kW BISG 48-120V Mild Hybrid - Draft TAR
20 kW CISG/TISG 48-120V Mild Hybrid - Draft TAR
Technology Effectiveness [%]
Small
Car
7.4
9.5
15.2
Standard
Car
7.3
9.4
15.0
Large
Car
7.2
9.2
14.8
Small
MPV
6.9
8.8
14.2
Large
MPV
6.9
8.9
14.2
Small
Truck
6.8
8.2
12.0
Large
Truck
8.0
8.3
12.2
   For this Draft TAR analysis, EPA has updated the battery costs for high-voltage (non-48V)
mild hybrids, as described in Section 5.3.4.3.7.2. Non-battery costs for high-voltage mild hybrids
that were used in the 2012 FRM analysis have been retained for this analysis and updated to
2013$. In adding 48V mild hybrids to the analysis, new battery and non-battery costs were
developed as discussed in Section 5.2.
5.3.4.3.3
Cost and Effectiveness for Strong Hybrids
   In the 2012 FRM, P2 hybrid was the only hybrid architecture that was applied in the EPA
analysis. Although PSHEV and 2MHEV technology were discussed because they were present
in the market at the time of the FRM, they were not included in the analysis because the industry
was expected to trend toward more cost-effective hybrid configurations such as P2.

   The primary reference EPA used for strong hybrid effectiveness in the 2012 FRM was the
Ricardo modeling study which modeled a P2 with a future OCT. On this basis EPA estimated an
absolute CCh effectiveness for P2 strong hybrids ranging from 13.4 to 15.7 percent depending on
vehicle class (see 2012 RIA, p. 1-18).

   As reviewed in Section 5.2, several new production and research examples of strong hybrid
technology have emerged since the 2012 FRM. These examples provide a much broader picture
of the potential effectiveness of strong hybrid technology.

   The ANL-VOLPE analysis found about 34.3 percent total GHG effectiveness (including other
technologies present on the vehicle) for an input power-split HEV based on  the 2010 Toyota
                                            5-302

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Prius with a 1.8L PFI Atkinson cycle engine and a combined electric motor-generator power of
77kW. The 34.3 percent GHG effectiveness estimate is comparable to the 33.3 percent GHG
effectiveness of Toyota Camry power-split from the two-cycle combined results from
certification test data when comparing the 2015 Toyota Camry HEV to the non-HEV, 4-cylinder
version of the 2015 Camry. The ANL-VOLPE analysis also found approximately 32.6 percent
GHG effectiveness for a P2 parallel hybrid with a 30 kW traction motor. The 32.6 percent GHG
effectiveness of 30 kW P2 hybrid is comparable to 33.9 percent total GHG effectiveness of 2016
Hyundai Sonata P2 parallel hybrid calculated from a comparison of two-cycle combined
certification test data between the 2016 Hyundai Sonata Hybrid with a 2.0L Atkinson cycle
engine and a non-HEV 2015 Sonata with a 2.4L GDI engine. In the 2016 Hyundai Sonata P2
hybrid, a 38 kW traction motor and wet clutches are integrated into the transmission and replace
the torque converter in a planetary gearset six-speed automatic transmission. A second, 10.5 kW
high voltage Hybrid Starter Generator (HSG) BISG is incorporated for torque smoothing
between the engine and the traction motor, automatic engine re-starting, and battery charging at
idle in Hyundai Sonata hybrid.

   Many aspects of hybrid technology effectiveness can be estimated by means of computational
tools such as ANL-Autonomie, Gamma Technology GT-Power/GT-Suite, MSC EASY5, EPA-
ALPHA and other vehicle models.  A standalone hybrid vehicle model544 was used to correlate
recent ANL chassis dynamometer test data and 2010 Toyota Prius power-split hybrid and 2011
Hyundai Sonata P2 parallel hybrid model simulations over U.S. regulatory driving cycles. The
model was successfully validated using ANL test data within 5 percent of test cycle fuel
economy.

   EPA also calculated overall strong hybrid effectiveness by comparing the non-hybrid variants
from the same vehicle manufacturers.  For example, the 2015 2.5L 14 engine non-hybrid Camry
was used to estimate the overall effectiveness of 2015 2.5L Camry hybrid. The use of a PFI
Atkinson Cycle engine, improved aero-dynamics, and reduced tire rolling resistance technology
effectiveness were applied within the Lumped Parameter Model (LPM) to better estimate the
overall system effectiveness of strong hybrid  electrification since the Camry Hybrid vehicle
package includes these differences in addition to the power-split HEV system. Two-cycle fuel
economy (MPG) data over the city and highway drive cycles were used to estimate the relative
effectiveness improvement of the hybrid electric vehicles. Hybrid technology effectiveness can
then be estimated by subtracting the LPM/NRC-estimated effectiveness of non-hybrid
technologies present on the vehicle from the total effectiveness.

   Hybrid technology effectiveness of input power-split hybrids and P2 parallel hybrids appear
to be converging and this appears to be confirmed by the fuel economy achieved with the 2017
Hyundai IONIQ P2 hybrid with a highly hybrid-optimized 6 speed DCT transmission. Hence,
the GHG effectiveness was updated to 20.1 percent for mid-size standard car strong hybrids
compared to the 15.5 percent effectiveness used in the 2012 FRM, as shown in Table 5.86.
                                            5-303

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                   Table 5.86 GHG Technology Effectiveness of Strong Hybrids
Technology
P2 Full Hybrid Drivetrain - 2012 FRM
Strong Hybrid - Draft TAR
Technology Effectiveness [%]
Small Car
15.5
19.0
Standard
Car
15.5
20.1
Large
Car
15.4
19.9
Small
MPV
14.6
18.8
Large
MPV
14.6
19.1
Small
Truck
13.4
17.2
Large
Truck
15.7
17.7
   For this Draft TAR analysis, EPA has updated the battery costs for strong hybrids, as
described in Section 5.3.4.3.7.2. Non-battery costs for strong hybrids that were used in the 2012
FRM analysis have been retained for this analysis and updated to 2014$.
5.3.4.3.4
Cost and Effectiveness for Plug-in Hybrids
   Plug-in hybrid electric vehicles (PHEVs) utilize two sources of energy, electricity and liquid
fuel, which are accounted for differently according to the effectiveness accounting methods
established in the 2012 FRM.

   As discussed in the 2012 TSD that accompanied the FRM, the overall GHG effectiveness
potential of PHEVs depends on many factors, the most important being the energy storage
capacity designed into the battery pack, and the vehicle's ability to provide all electric range to
the operator.  Section 3.4.3.6.4 of the TSD detailed the methods by which EPA and NHTSA
estimated PHEV effectiveness.  According to the method used by EPA, which estimates
effectiveness based on the SAE J1711 utility factor calculation, the AER, and the vehicle class,
the assumed effectiveness for a PHEV20 would be approximately 58 percent GHG reduction for
a midsize car and approximately 47 percent  GHG reduction for a large truck.

   The 2012 FRM established an incentive multiplier for compliance purposes for PHEVs sold
in MYs 2017 through 2021. This multiplier approach means that each PHEV would count as
more than one vehicle in the manufacturer's compliance calculation. The multiplier value for
PHEVs starts at 1.6 in MY2017 and phases down to a value of 1.3 in MY2021.  There is no
PHEV multiplier for MYs 2022-2025.

   The 2012 FRM also set the tailpipe compliance value for the electricity portion of PHEV
energy usage to 0 g/mi for MYs 2017-2021, with no limit on the quantity of vehicles eligible for
0 g/mi tailpipe emissions accounting. For MYs 2022-2025, 0 g/mi will only be allowed up to a
per-company cumulative sales cap: 1) 600,000 vehicles for companies that sell 300,000
BEV/PHEV/FCVs in MYs 2019-2021; 2) 200,000 vehicles for all other manufacturers.  For
sales above these thresholds, manufacturers  will be required to account for the net upstream
GHG emissions for the electric portion of operation, using accounting methodologies set out in
the FRM.

   As with other electrified vehicles, costs for PHEVs are separated into battery and non-battery
costs. EPA has updated these costs as described in Sections 5.3.4.3.6 and 5.3.4.3.7.
5.3.4.3.5
Cost and Effectiveness for Electric Vehicles
   The 2012 FRM established an incentive multiplier for compliance purposes for BEVs sold in
MYs 2017 through 2021.  This multiplier approach means that each BEV counts as more than
                                            5-304

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
one vehicle in the manufacturer's compliance calculation.  The multiplier value for BEVs starts
at 2.0 in MY2017 and phases down to a value of 1.5 in MY2021. There is no BEV multiplier for
MYs 2022-2025.

   The 2012 FRM also set the tailpipe compliance value for the electricity usage of BEVs to 0
g/mi for MYs 2017-2021, with no limit on the quantity of vehicles eligible for 0 g/mi tailpipe
emissions accounting. For MYs 2022-2025,  0 g/mi will only be allowed up to a per-company
cumulative sales cap: 1) 600,000 vehicles for companies that sell 300,000 BEV/PHEV/FCVs in
MYs 2019-2021; 2) 200,000 vehicles for all other manufacturers. For sales above these
thresholds, manufacturers will be required to account for the net upstream GHG emissions for
the electric portion of operation, using accounting methodologies set out in the FRM. In this
Draft TAR analysis, the GHG effectiveness of BEVs is unchanged from that used in the FRM,
which is 100 percent GHG reduction.

   As with other electrified vehicles, costs for BEVs are separated into battery and non-battery
costs. EPA has updated these costs as described in Sections 5.3.4.3.6 and 5.3.4.3.7.

5.3.4.3.6      Cost of Non-Battery Components for xEVs

   At this time, EPA is continuing  to use the  2012 FRM cost assumptions for non-battery
components as a basis for draft OMEGA runs. Costs for electric motors are slightly modified by
changes in motor sizing resulting from the revised battery sizing methodology described below,
but are based on the underlying motor cost assumptions of the FRM.

   The 2015 NAS report correctly noted that raw material costs for propulsion motors tends to be
a stronger function of torque output than of power output, and recommended that the agencies
scale motor costs on a torque basis. While EPA acknowledges the technical basis of this
recommendation, practical considerations make it difficult to do so while remaining compatible
with other aspects of the analysis that require motors to be characterized by power output.
Accurately converting between  a torque basis and a power basis would require a greater amount
of information to be specified about the individual propulsion systems and drivelines of each of
the modeled PHEVs,  possibly limiting the applicability of the analysis to a narrower range of
configurations than intended. Further, through additional research and through stakeholder
meetings with OEMs, EPA has found that it is not unusual to encounter motor cost projections or
targets being expressed in terms of power, such as dollars per kilowatt. The US DRIVE cost
targets for electric motors published by the Department of Energy are also expressed in dollars
per kilowatt. Finally, the cost of the power electronics that accompany a propulsion motor
system are closely related to the power specification of the propulsion motor, and are also
commonly projected or targeted as a function of power.  For these reasons, EPA has chosen to
continue to scale motor and power  electronics costs in terms of power rather than torque.

   Several possible sources for updated non-battery costs may become available after the June
2016 publication of this Draft TAR but prior to the proposed determination.

   In May 2016, CARB commissioned a study on non-battery costs for strong HEVs and
PHEVs.545  Initial results from this study may become available in late 2016  and will be
considered for future  inclusion in the EPA non-battery cost model. EPA is also considering
commissioning a teardown study of a BEV or PHEV through a contractor, with the goal of
further quantifying non-battery costs for these vehicles.
                                            5-305

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
   EPA is also studying the possibility of using US DRIVE cost targets for motors and power
electronics, based on information gained through stakeholder meetings that suggests that some
OEMs may already be meeting or exceeding some of these targets, or are on track to do so
within the time frame of the rule.

   EPA has also reviewed many cost estimates by applying engineering judgement informed by
ongoing survey of industry literature, announcements of new products, and discussions with
OEMs and suppliers.

   For this Draft TAR, EPA has continued to use the same non-battery costs as used in the 2012
FRM with two exceptions: costs have been updated to 2013$; and, MFIEV48V non-battery costs
are new since they were not considered in the 2012 FRM. All applicable non-battery costs are
presented in the tables below, first in terms of cost curves as were presented in the 2012 FRM,
and then for each vehicle class at various mass reduction levels.
Table 5.87 Linear Regressions of Strong & Plug-in Hybrid Non-Battery System Direct Manufacturing Costs
                       vs Net Mass Reduction Applicable in MY2012 (2013$)
Vehicle Class
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Strong HEV
-$277x+$l,766
-$412x+$l,958
-$737x+$2,293
-$349x+$l,874
-$533x+$2,164
-$683x+$2,287
PHEV20
-$426x+$2,122
-$672x+$2,443
-$l,390x+$3,214
-$601x+$2,344
n/a
n/a
PHEV40
-$852x+$2,597
-$l,343x+$3,175
-$2,780x+$4,705
-$l,203x+$2,997
n/a
n/a
                     Note: "x" in the equations represents the net weight reduction as a percentage.
  Table 5.88 Linear Regressions of Battery Electric Non-Battery System Direct Manufacturing Costs vs Net
                          Mass Reduction Applicable in MY2016 (2013$)
Vehicle Class
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
EV75
-$978x+-$134
-$l,542x+$526
-$3,190x+$l,365
-$l,381x+-$516
n/a
n/a
EV100
-$978x+-$134
-$l,542x+$526
-$3,190x+$l,365
-$l,381x+-$516
n/a
n/a
EV200
-$978x+-$133
-$l,542x+$527
-$3,190x+$l,366
-$l,381x+-$516
n/a
n/a
                     Note: "x" in the equations represents the net weight reduction as a percentage.
              Table 5.89 Costs for MHEV48V Non-Battery Items (dollar values in 2013$)
Vehicle Class
All
All
All
Cost type
DMC
1C
TC
DMC: base year cost
1C: complexity
$440
Med2

DMC: learning curve
1C: near term thru
23
2018

2017
$398
$168
$567
2018
$392
$168
$560
2019
$386
$126
$512
2020
$381
$126
$506
2021
$376
$125
$501
2022
$371
$125
$496
2023
$367
$125
$492
2024
$362
$125
$487
2025
$359
$125
$483
              Note:
              DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                              5-306

-------
                  Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.90 Costs for Strong Hybrid Non-Battery Items (dollar values in 2013$)
Vehicle
Class
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
WRtech
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
WRnet
5
10
15
5
10
15
5
10
15
5
10
15
6
11
16
6
11
16
5
10
15
5
10
15
5
10
15
5
10
15
6
11
16
6
11
16
5
10
15
5
10
15
5
10
15
5
10
15
6
11
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
$1,752
$1,738
$1,725
$1,937
$1,917
$1,896
$2,257
$2,220
$2,183
$1,857
$1,839
$1,822
$2,132
$2,105
$2,079
$2,246
$2,212
$2,178
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl














DMC:
learning
curve
1C: near
term
thru
23
23
23
23
23
23
23
23
23
23
23
23
23
23
23
23
23
23
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018














2017
$1,587
$1,574
$1,561
$1,754
$1,736
$1,717
$2,043
$2,010
$1,976
$1,681
$1,665
$1,649
$1,930
$1,906
$1,882
$2,034
$2,003
$1,972
$977
$969
$961
$1,080
$1,069
$1,057
$1,258
$1,237
$1,217
$1,035
$1,025
$1,016
$1,189
$1,174
$1,159
$1,252
$1,233
$1,214
$2,563
$2,543
$2,523
$2,834
$2,804
$2,774
$3,301
$3,247
$3,193
$2,716
$2,691
$2,665
$3,119
$3,080
2018
$1,562
$1,549
$1,537
$1,727
$1,708
$1,690
$2,011
$1,978
$1,945
$1,655
$1,639
$1,624
$1,900
$1,876
$1,853
$2,002
$1,971
$1,941
$975
$968
$960
$1,078
$1,067
$1,055
$1,256
$1,235
$1,215
$1,033
$1,024
$1,014
$1,187
$1,172
$1,157
$1,250
$1,231
$1,212
$2,537
$2,517
$2,497
$2,805
$2,775
$2,745
$3,267
$3,214
$3,160
$2,688
$2,663
$2,637
$3,087
$3,048
2019
$1,539
$1,526
$1,514
$1,701
$1,683
$1,665
$1,981
$1,949
$1,917
$1,630
$1,615
$1,600
$1,872
$1,849
$1,825
$1,972
$1,942
$1,912
$598
$594
$589
$662
$655
$648
$771
$758
$745
$634
$628
$622
$728
$719
$710
$767
$755
$744
$2,137
$2,120
$2,103
$2,363
$2,338
$2,313
$2,752
$2,707
$2,662
$2,264
$2,243
$2,222
$2,600
$2,568
2020
$1,517
$1,505
$1,493
$1,678
$1,660
$1,642
$1,954
$1,922
$1,890
$1,608
$1,592
$1,577
$1,846
$1,823
$1,800
$1,945
$1,915
$1,886
$598
$593
$588
$661
$654
$647
$770
$757
$745
$633
$627
$621
$727
$718
$709
$766
$755
$743
$2,115
$2,098
$2,082
$2,338
$2,314
$2,289
$2,724
$2,679
$2,635
$2,241
$2,220
$2,199
$2,573
$2,541
2021
$1,497
$1,485
$1,474
$1,655
$1,638
$1,620
$1,928
$1,897
$1,865
$1,586
$1,572
$1,557
$1,822
$1,799
$1,776
$1,919
$1,890
$1,861
$597
$592
$588
$660
$653
$646
$769
$756
$744
$633
$627
$621
$727
$717
$708
$765
$754
$742
$2,094
$2,078
$2,061
$2,316
$2,291
$2,266
$2,697
$2,653
$2,609
$2,219
$2,198
$2,177
$2,548
$2,516
2022
$1,479
$1,467
$1,455
$1,635
$1,617
$1,600
$1,904
$1,873
$1,842
$1,567
$1,552
$1,537
$1,799
$1,777
$1,754
$1,895
$1,866
$1,838
$597
$592
$587
$660
$653
$646
$768
$756
$743
$632
$626
$620
$726
$717
$708
$765
$753
$741
$2,075
$2,059
$2,042
$2,294
$2,270
$2,246
$2,672
$2,629
$2,585
$2,199
$2,178
$2,157
$2,525
$2,493
2023
$1,461
$1,449
$1,438
$1,615
$1,598
$1,581
$1,882
$1,851
$1,820
$1,548
$1,533
$1,519
$1,778
$1,755
$1,733
$1,873
$1,844
$1,816
$596
$591
$587
$659
$652
$645
$767
$755
$742
$631
$626
$620
$725
$716
$707
$764
$752
$741
$2,057
$2,041
$2,024
$2,274
$2,250
$2,226
$2,649
$2,606
$2,562
$2,179
$2,159
$2,138
$2,503
$2,471
2024
$1,444
$1,433
$1,422
$1,597
$1,580
$1,563
$1,860
$1,830
$1,799
$1,530
$1,516
$1,502
$1,757
$1,735
$1,714
$1,852
$1,823
$1,795
$595
$591
$586
$658
$651
$644
$767
$754
$742
$631
$625
$619
$724
$715
$706
$763
$752
$740
$2,040
$2,024
$2,008
$2,255
$2,231
$2,207
$2,627
$2,584
$2,541
$2,161
$2,141
$2,121
$2,482
$2,451
2025
$1,429
$1,418
$1,406
$1,580
$1,563
$1,546
$1,840
$1,810
$1,780
$1,514
$1,500
$1,485
$1,738
$1,717
$1,695
$1,831
$1,804
$1,776
$595
$590
$586
$658
$651
$644
$766
$754
$741
$630
$624
$619
$724
$715
$706
$763
$751
$739
$2,024
$2,008
$1,992
$2,238
$2,214
$2,190
$2,606
$2,564
$2,521
$2,144
$2,124
$2,104
$2,462
$2,432
                                 5-307

-------
                          Technology Cost, Effectiveness, and Lead-Time Assessment
LgMPV
Truck
Truck
Truck
20
10
15
20
16
6
11
16
TC
TC
TC
TC








$3,041
$3,286
$3,236
$3,186
$3,009
$3,252
$3,202
$3,153
$2,535
$2,739
$2,698
$2,656
$2,509
$2,711
$2,670
$2,629
$2,485
$2,685
$2,644
$2,603
$2,462
$2,660
$2,620
$2,579
$2,440
$2,637
$2,597
$2,557
$2,420
$2,615
$2,575
$2,535
$2,401
$2,594
$2,555
$2,515
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
  Table 5.91 Costs for 20 Mile Plug-in Hybrid Non-Battery Items (dollar values in 2013$)
Vehicle
Class
SmCar
SmCar
StCar
StCar
LgCar
LgCar
SmMPV
SmMPV
LgMPV
LgMPV
Truck
Truck
SmCar
SmCar
StCar
StCar
LgCar
LgCar
SmMPV
SmMPV
LgMPV
LgMPV
Truck
Truck
SmCar
SmCar
StCar
StCar
LgCar
LgCar
SmMPV
SmMPV
LgMPV
LgMPV
Truck
Truck
WRtech
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
WRnet
6
11
6
11
5
10
6
11
4
9
6
11
6
11
6
11
5
10
6
11
4
9
6
11
6
11
6
11
5
10
6
11
4
9
6
11
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
$2,097
$2,075
$2,402
$2,369
$3,145
$3,075
$2,307
$2,277
$2,797
$2,750
$2,943
$2,884
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl












DMC:
learning
curve
1C: near
term
thru
23
23
23
23
23
23
23
23
23
23
23
23
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018












2017
$1,898
$1,879
$2,175
$2,145
$2,847
$2,784
$2,089
$2,062
$2,533
$2,490
$2,664
$2,611
$1,169
$1,157
$1,339
$1,321
$1,753
$1,714
$1,286
$1,270
$1,559
$1,533
$1,640
$1,608
$3,067
$3,036
$3,514
$3,465
$4,600
$4,499
$3,376
$3,332
$4,092
$4,022
$4,305
$4,219
2018
$1,869
$1,850
$2,141
$2,111
$2,803
$2,741
$2,056
$2,030
$2,493
$2,450
$2,623
$2,571
$1,167
$1,155
$1,337
$1,318
$1,750
$1,712
$1,284
$1,267
$1,557
$1,530
$1,638
$1,605
$3,035
$3,005
$3,478
$3,429
$4,553
$4,452
$3,341
$3,297
$4,050
$3,981
$4,260
$4,176
2019
$1,841
$1,822
$2,109
$2,080
$2,761
$2,700
$2,026
$2,000
$2,456
$2,414
$2,584
$2,533
$716
$709
$820
$809
$1,074
$1,050
$788
$778
$955
$939
$1,005
$985
$2,557
$2,531
$2,930
$2,889
$3,835
$3,751
$2,814
$2,778
$3,411
$3,353
$3,589
$3,518
2020
$1,815
$1,797
$2,080
$2,051
$2,723
$2,663
$1,998
$1,972
$2,422
$2,381
$2,548
$2,497
$715
$708
$819
$808
$1,073
$1,049
$787
$777
$954
$938
$1,004
$984
$2,531
$2,505
$2,900
$2,859
$3,796
$3,712
$2,785
$2,749
$3,376
$3,319
$3,552
$3,481
2021
$1,792
$1,773
$2,053
$2,024
$2,687
$2,628
$1,972
$1,946
$2,390
$2,349
$2,514
$2,465
$714
$707
$819
$807
$1,072
$1,048
$786
$776
$953
$937
$1,003
$983
$2,506
$2,481
$2,871
$2,831
$3,759
$3,676
$2,758
$2,722
$3,343
$3,286
$3,517
$3,447
2022
$1,769
$1,751
$2,027
$1,999
$2,654
$2,595
$1,947
$1,922
$2,360
$2,320
$2,483
$2,434
$714
$707
$818
$806
$1,071
$1,047
$786
$775
$952
$936
$1,002
$982
$2,483
$2,458
$2,845
$2,805
$3,724
$3,642
$2,733
$2,697
$3,312
$3,256
$3,485
$3,416
2023
$1,748
$1,730
$2,003
$1,975
$2,622
$2,564
$1,924
$1,899
$2,332
$2,293
$2,454
$2,405
$713
$706
$817
$806
$1,070
$1,046
$785
$775
$951
$935
$1,001
$981
$2,461
$2,436
$2,820
$2,781
$3,692
$3,610
$2,709
$2,673
$3,284
$3,228
$3,454
$3,386
2024
$1,728
$1,711
$1,980
$1,953
$2,592
$2,535
$1,902
$1,877
$2,306
$2,267
$2,426
$2,378
$712
$705
$816
$805
$1,069
$1,045
$784
$774
$950
$934
$1,000
$980
$2,441
$2,416
$2,797
$2,758
$3,661
$3,580
$2,686
$2,651
$3,256
$3,201
$3,426
$3,358
2025
$1,710
$1,692
$1,959
$1,931
$2,564
$2,508
$1,882
$1,857
$2,281
$2,242
$2,400
$2,352
$712
$705
$816
$804
$1,068
$1,044
$783
$773
$950
$934
$999
$979
$2,421
$2,397
$2,775
$2,736
$3,632
$3,552
$2,665
$2,630
$3,230
$3,176
$3,399
$3,331
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                         5-308

-------
                         Technology Cost, Effectiveness, and Lead-Time Assessment
  Table 5.92 Costs for 40 Mile Plug-in Hybrid Non-Battery Items (dollar values in 2013$)
Vehicle
Class
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
WRtech
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
WRnet
6
5
3
7
0
5
6
5
3
7
0
5
6
5
3
7
0
5
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
$2,546
$3,108
$4,622
$2,912
$3,850
$4,133
Highl
Highl
Highl
Highl
Highl
Highl






DMC:
learning
curve
1C: near
term
thru
23
23
23
23
23
23
2018
2018
2018
2018
2018
2018






2017
$2,305
$2,814
$4,185
$2,637
$3,486
$3,742
$1,419
$1,733
$2,577
$1,624
$2,147
$2,304
$3,724
$4,547
$6,761
$4,261
$5,633
$6,046
2018
$2,269
$2,770
$4,119
$2,596
$3,432
$3,683
$1,417
$1,730
$2,572
$1,621
$2,143
$2,300
$3,685
$4,500
$6,692
$4,216
$5,575
$5,984
2019
$2,235
$2,729
$4,058
$2,557
$3,381
$3,629
$869
$1,062
$1,578
$995
$1,315
$1,411
$3,105
$3,791
$5,637
$3,552
$4,696
$5,041
2020
$2,204
$2,691
$4,002
$2,522
$3,334
$3,579
$868
$1,060
$1,577
$993
$1,314
$1,410
$3,073
$3,752
$5,579
$3,515
$4,648
$4,989
2021
$2,175
$2,656
$3,949
$2,489
$3,290
$3,532
$867
$1,059
$1,575
$992
$1,312
$1,408
$3,043
$3,715
$5,524
$3,481
$4,602
$4,940
2022
$2,148
$2,623
$3,900
$2,457
$3,249
$3,487
$867
$1,058
$1,573
$991
$1,311
$1,407
$3,015
$3,681
$5,473
$3,449
$4,560
$4,894
2023
$2,122
$2,592
$3,854
$2,428
$3,210
$3,446
$866
$1,057
$1,572
$991
$1,310
$1,406
$2,988
$3,649
$5,426
$3,419
$4,520
$4,852
2024
$2,098
$2,562
$3,810
$2,401
$3,174
$3,407
$865
$1,056
$1,571
$990
$1,308
$1,404
$2,963
$3,618
$5,381
$3,390
$4,482
$4,811
2025
$2,076
$2,534
$3,769
$2,375
$3,140
$3,370
$864
$1,055
$1,569
$989
$1,307
$1,403
$2,940
$3,590
$5,338
$3,364
$4,447
$4,773
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                        5-309

-------
                         Technology Cost, Effectiveness, and Lead-Time Assessment
       Table 5.93 Costs for 75 Mile BEV Non-Battery Items (dollar values in 2013$)
Vehicle
Class
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
WRtech
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
WRnet
10
15
20
10
15
20
10
15
20
10
15
20
5
10
15
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
5
10
15
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
5
10
15
10
15
20
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
-$232
-$281
-$329
$371
$294
$217
$1,046
$886
$727
-$654
-$723
-$792
$359
$250
$141
-$653
-$787
-$921
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2


















DMC:
learning
curve
1C: near
term
thru
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024


















2017
-$226
-$274
-$322
$363
$288
$212
$1,022
$866
$710
-$639
-$707
-$774
$351
$244
$137
-$638
-$769
-$900
$178
$216
$253
$285
$226
$167
$803
$681
$558
$503
$556
$609
$276
$192
$108
$502
$605
$708
-$48
-$59
-$69
$648
$514
$379
$1,825
$1,547
$1,268
-$136
-$151
-$165
$626
$436
$245
-$136
-$164
-$192
2018
-$222
-$268
-$315
$355
$282
$208
$1,000
$848
$695
-$626
-$692
-$758
$343
$239
$134
-$624
-$753
-$881
$178
$215
$253
$285
$226
$167
$802
$680
$557
$502
$555
$608
$275
$192
$108
$501
$604
$706
-$44
-$53
-$62
$640
$507
$374
$1,802
$1,527
$1,252
-$124
-$137
-$150
$618
$430
$242
-$124
-$149
-$175
2019
-$217
-$263
-$309
$348
$276
$204
$981
$831
$681
-$614
-$678
-$743
$337
$234
$132
-$612
-$738
-$864
$177
$215
$252
$284
$225
$166
$800
$678
$556
$501
$554
$607
$275
$191
$108
$500
$602
$705
-$40
-$48
-$57
$633
$501
$370
$1,781
$1,509
$1,238
-$113
-$125
-$137
$611
$425
$239
-$112
-$136
-$159
2020
-$213
-$258
-$303
$342
$271
$200
$963
$816
$669
-$603
-$666
-$730
$331
$230
$129
-$601
-$725
-$848
$177
$214
$252
$284
$225
$166
$799
$677
$555
$500
$553
$606
$274
$191
$107
$499
$601
$704
-$36
-$44
-$52
$626
$496
$366
$1,762
$1,493
$1,225
-$103
-$113
-$124
$605
$421
$237
-$102
-$123
-$144
2021
-$210
-$254
-$298
$336
$267
$197
$947
$802
$658
-$593
-$655
-$718
$325
$226
$127
-$591
-$713
-$834
$177
$214
$251
$283
$225
$166
$798
$676
$555
$499
$552
$605
$274
$191
$107
$498
$601
$703
-$33
-$40
-$47
$620
$491
$363
$1,745
$1,479
$1,213
-$93
-$103
-$113
$599
$417
$235
-$93
-$112
-$131
2022
-$207
-$250
-$294
$331
$262
$194
$932
$790
$648
-$583
-$645
-$706
$320
$223
$125
-$582
-$701
-$821
$177
$214
$251
$283
$224
$166
$797
$675
$554
$499
$551
$604
$273
$190
$107
$498
$600
$702
-$30
-$36
-$43
$614
$487
$359
$1,729
$1,465
$1,201
-$85
-$93
-$102
$593
$413
$232
-$84
-$102
-$119
2023
-$203
-$246
-$289
$326
$258
$191
$918
$778
$638
-$575
-$635
-$696
$315
$219
$123
-$573
-$691
-$809
$176
$214
$251
$283
$224
$165
$796
$674
$553
$498
$551
$603
$273
$190
$107
$497
$599
$701
-$27
-$33
-$39
$609
$483
$356
$1,714
$1,453
$1,191
-$77
-$85
-$93
$588
$409
$230
-$76
-$92
-$108
2024
-$201
-$243
-$285
$322
$255
$188
$905
$767
$629
-$567
-$626
-$686
$311
$216
$122
-$565
-$682
-$798
$176
$213
$250
$282
$224
$165
$795
$674
$552
$497
$550
$602
$273
$190
$107
$496
$598
$700
-$24
-$30
-$35
$604
$479
$353
$1,700
$1,441
$1,182
-$69
-$76
-$84
$584
$406
$229
-$69
-$83
-$97
2025
-$198
-$240
-$282
$317
$252
$186
$894
$757
$621
-$559
-$618
-$677
$307
$213
$120
-$558
-$673
-$787
$113
$137
$161
$182
$144
$106
$512
$434
$356
$320
$354
$388
$176
$122
$69
$320
$385
$451
-$85
-$102
-$120
$499
$396
$292
$1,405
$1,191
$977
-$239
-$264
-$289
$482
$336
$189
-$238
-$287
-$336
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                        5-310

-------
                         Technology Cost, Effectiveness, and Lead-Time Assessment
      Table 5.94 Costs for 100 Mile BEV Non-Battery Items (dollar values in 2013$)
Vehicle
Class
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
WRtech
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
WRnet
8
13
18
7
12
17
8
13
18
7
12
17
3
8
13
7
12
17
8
13
18
7
12
17
8
13
18
7
12
17
3
8
13
7
12
17
8
13
18
7
12
17
8
13
18
7
12
17
3
8
13
7
12
17
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
-$212
-$261
-$310
$418
$341
$264
$1,110
$950
$791
-$613
-$682
-$751
$403
$293
$184
-$572
-$707
-$841
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2


















DMC:
learning
curve
1C: near
term
thru
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
28
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024


















2017
-$207
-$255
-$303
$408
$333
$257
$1,084
$928
$772
-$599
-$666
-$734
$393
$287
$180
-$559
-$690
-$821
$163
$201
$238
$321
$262
$202
$853
$730
$607
$471
$524
$577
$309
$225
$142
$440
$543
$646
-$44
-$54
-$65
$729
$594
$460
$1,936
$1,658
$1,380
-$128
-$142
-$157
$703
$512
$322
-$119
-$147
-$175
2018
-$203
-$250
-$296
$399
$326
$252
$1,061
$909
$756
-$586
-$652
-$718
$385
$281
$176
-$547
-$676
-$804
$163
$200
$238
$320
$261
$202
$851
$729
$606
$470
$523
$576
$309
$225
$141
$439
$542
$645
-$40
-$49
-$59
$720
$587
$454
$1,912
$1,637
$1,362
-$116
-$129
-$142
$694
$506
$317
-$108
-$134
-$159
2019
-$199
-$245
-$291
$392
$319
$247
$1,040
$891
$741
-$575
-$640
-$704
$378
$275
$173
-$537
-$663
-$788
$162
$200
$237
$320
$261
$202
$849
$727
$605
$469
$522
$575
$308
$225
$141
$438
$541
$644
-$37
-$45
-$53
$711
$580
$449
$1,890
$1,618
$1,346
-$106
-$118
-$129
$686
$500
$314
-$99
-$122
-$145
2020
-$195
-$240
-$285
$385
$314
$243
$1,022
$875
$728
-$564
-$628
-$692
$371
$270
$170
-$527
-$651
-$774
$162
$199
$237
$319
$260
$201
$848
$726
$604
$468
$521
$574
$308
$224
$141
$437
$540
$643
-$33
-$41
-$49
$704
$574
$444
$1,870
$1,601
$1,332
-$96
-$107
-$118
$678
$494
$310
-$90
-$111
-$132
2021
-$192
-$236
-$281
$378
$308
$239
$1,005
$860
$716
-$555
-$618
-$680
$365
$266
$167
-$518
-$640
-$761
$162
$199
$236
$319
$260
$201
$847
$725
$603
$468
$520
$573
$307
$224
$141
$437
$539
$642
-$30
-$37
-$44
$697
$568
$440
$1,851
$1,585
$1,319
-$87
-$97
-$107
$672
$490
$307
-$81
-$101
-$120
2022
-$189
-$233
-$276
$372
$304
$235
$989
$847
$705
-$546
-$608
-$669
$359
$261
$164
-$510
-$630
-$749
$162
$199
$236
$318
$260
$201
$846
$724
$602
$467
$520
$572
$307
$224
$140
$436
$538
$641
-$27
-$34
-$40
$691
$563
$436
$1,834
$1,571
$1,307
-$79
-$88
-$97
$666
$485
$305
-$74
-$91
-$109
2023
-$186
-$229
-$272
$367
$299
$231
$974
$834
$694
-$538
-$599
-$659
$353
$258
$162
-$503
-$620
-$738
$161
$199
$236
$318
$259
$201
$844
$723
$602
$467
$519
$572
$306
$223
$140
$436
$538
$640
-$25
-$31
-$36
$685
$558
$432
$1,819
$1,557
$1,296
-$72
-$80
-$88
$660
$481
$302
-$67
-$83
-$98
2024
-$184
-$226
-$268
$362
$295
$228
$961
$823
$685
-$531
-$591
-$650
$349
$254
$160
-$496
-$612
-$728
$161
$198
$236
$318
$259
$200
$843
$722
$601
$466
$518
$571
$306
$223
$140
$435
$537
$639
-$22
-$28
-$33
$679
$554
$429
$1,804
$1,545
$1,285
-$65
-$72
-$79
$655
$477
$300
-$61
-$75
-$89
2025
-$181
-$223
-$265
$357
$291
$225
$948
$812
$676
-$524
-$583
-$642
$344
$251
$157
-$489
-$604
-$718
$104
$128
$152
$204
$167
$129
$543
$465
$387
$300
$334
$368
$197
$144
$90
$280
$346
$412
-$77
-$95
-$113
$561
$458
$354
$1,491
$1,277
$1,062
-$224
-$249
-$274
$541
$394
$248
-$209
-$258
-$307
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                        5-311

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
           Table 5.95 Costs for 200 Mile BEV Non-Battery Items (dollar values in 2013$)
Vehicle
Class
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
WRtech
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
WRnet
8
8
10
8
4
8
8
8
10
8
4
8
8
8
10
8
4
8
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
-$211
$403
$1,047
-$626
$380
-$597
High2
High2
High2
High2
High2
High2






DMC:
learning
curve
1C: near
term
thru
28
28
28
28
28
28
2024
2024
2024
2024
2024
2024






2017
-$206
$394
$1,023
-$612
$372
-$583
$162
$310
$805
$481
$292
$459
-$44
$704
$1,828
-$131
$664
-$125
2018
-$202
$386
$1,001
-$599
$364
-$571
$162
$309
$803
$480
$292
$458
-$40
$695
$1,804
-$119
$655
-$113
2019
-$198
$378
$982
-$587
$357
-$560
$161
$309
$802
$479
$291
$457
-$36
$687
$1,784
-$108
$648
-$103
2020
-$194
$371
$964
-$577
$350
-$550
$161
$308
$800
$479
$291
$456
-$33
$679
$1,765
-$98
$641
-$94
2021
-$191
$365
$948
-$567
$344
-$540
$161
$308
$799
$478
$290
$455
-$30
$673
$1,747
-$89
$635
-$85
2022
-$188
$359
$933
-$558
$339
-$532
$161
$307
$798
$477
$290
$455
-$27
$667
$1,731
-$81
$629
-$77
2023
-$185
$354
$920
-$550
$334
-$524
$160
$307
$797
$477
$289
$454
-$25
$661
$1,716
-$73
$623
-$70
2024
-$183
$349
$907
-$542
$329
-$517
$160
$307
$796
$476
$289
$454
-$22
$656
$1,703
-$66
$619
-$63
2025
-$180
$345
$895
-$535
$325
-$510
$103
$197
$513
$307
$186
$292
-$77
$542
$1,407
-$229
$511
-$218
    Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.





Table 5.96 Costs for In-Home Charger Associated with 20 Mile Plug-in Hybrid (dollar values in 2013$)
Vehicle Class
All
All
All
Cost type
DMC
1C
TC
DMC: base year cost
1C: complexity
$32
Highl

DMC: learning curve
1C: near term thru
26
2024

2017
$52
$19
$72
2018
$49
$19
$68
2019
$46
$19
$65
2020
$44
$19
$63
2021
$42
$19
$61
2022
$40
$19
$59
2023
$39
$19
$57
2024
$37
$19
$56
2025
$36
$11
$47
    Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
Table 5.97 Costs for In-Home Charger Associated with 40 Mile Plug-in Hybrid (dollar values in 2013$)
Vehicle Class
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
Cost type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
DMC: base year cost
1C: complexity
$169
$197
$215
$215
$215
$215
Highl
Highl
Highl
Highl
Highl
Highl






DMC: learning curve
1C: near term thru
26
26
26
26
26
26
2024
2024
2024
2024
2024
2024






2017
$273
$317
$347
$347
$347
$347
$102
$119
$130
$130
$130
$130
$375
$436
$477
$477
$477
$477
2018
$256
$298
$326
$326
$326
$326
$101
$117
$128
$128
$128
$128
$357
$415
$454
$454
$454
$454
2019
$242
$282
$308
$308
$308
$308
$100
$116
$127
$127
$127
$127
$343
$398
$435
$435
$435
$435
2020
$231
$268
$293
$293
$293
$293
$99
$116
$126
$126
$126
$126
$330
$383
$419
$419
$419
$419
2021
$220
$256
$280
$280
$280
$280
$99
$115
$125
$125
$125
$125
$319
$371
$405
$405
$405
$405
2022
$211
$246
$268
$268
$268
$268
$98
$114
$125
$125
$125
$125
$310
$360
$393
$393
$393
$393
2023
$203
$236
$258
$258
$258
$258
$98
$113
$124
$124
$124
$124
$301
$350
$382
$382
$382
$382
2024
$196
$228
$249
$249
$249
$249
$97
$113
$123
$123
$123
$123
$294
$341
$373
$373
$373
$373
2025
$190
$221
$241
$241
$241
$241
$59
$69
$75
$75
$75
$75
$249
$289
$316
$316
$316
$316
    Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                               5-312

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
        Table 5.98 Costs for In-Home Charger Associated with All BEVs (dollar values in 2013$)
Vehicle Class &
Range

All
All
All
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$215
Highl

DMC: learning
curve
1C: near term thru
26
2024

2017


$347
$130
$477
2018


$326
$128
$454
2019


$308
$127
$435
2020


$293
$126
$419
2021


$280
$125
$405
2022


$268
$125
$393
2023


$258
$124
$382
2024


$249
$123
$373
2025


$241
$75
$316
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.

   Table 5.99 Costs for Labor Associated with All In-Home Chargers for Plug-in & BEV (dollar values in
                                         2013$)
Vehicle Class &
Range


All
All
All
Cost
type


DMC
1C
TC
DMC: base year
cost
1C: complexity

$1075
None

DMC: learning
curve
1C: near term
thru
1
n/a

2017



$1,075
$0
$1,075
2018



$1,075
$0
$1,075
2019



$1,075
$0
$1,075
2020



$1,075
$0
$1,075
2021



$1,075
$0
$1,075
2022



$1,075
$0
$1,075
2023



$1,075
$0
$1,075
2024



$1,075
$0
$1,075
2025



$1,075
$0
$1,075
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.3.7
Cost of Batteries for xEVs
   In order to develop cost estimates for electrified vehicles, it is necessary to determine the
specifications of battery and non-battery components that can deliver the desired energy
management, driving range and acceleration performance goals. Once known, their properties
can then be input to a costing methodology to develop detailed projections of their cost.

   Battery costs have many drivers, and future cost projections derived by any methodology are
subject to significant uncertainties. The choice of costing methodology is therefore an important
consideration. For costing of battery components, EPA uses BatPaC,546 a peer-reviewed battery
costing model developed by Argonne National Laboratory (ANL).  As described later in Section
5.3.4.3.7.3, the ANL BatPaC model employs a rigorous, bottom-up, bill-of-materials approach to
battery cost analysis, and has undergone continual development  and review since the 2012 FRM.

   BatPaC requires numerous input assumptions, including battery  energy capacity,  battery
output power, and many other assumptions describing the chemistry, construction, and other
aspects of the battery.

   A first step in this process is the determination of battery energy capacity and battery output
power. The following sections describe: (a) how EPA determined battery energy capacity and
power for a population of modeled electrified vehicles; (b) how EPA selected other input
assumptions to BatPaC that influence battery cost, and (c) how the inputs and assumptions that
EPA employed in the FRM analysis were updated for this Draft  TAR analysis. Source data for
many of the charts in this section are available in the Docket.547

5.3.4.3.7.1    Battery Sizing Methodology for BEVs and PHEVs

   This section discusses how EPA sized the batteries for BEVs and PFLEVs (referred to
collectively here as PEVs). For FtEVs, EPA used a different methodology that is described in
the next section.
                                             5-313

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Sizing a PEV battery pack primarily involves determining the necessary energy storage
capacity (in kWh) and power capability (in kW) to provide a desired driving range and level of
acceleration performance. Energy storage capacity has a strong influence on the weight of the
pack as well as its overall cost because it determines the amount of active energy storage
material that must be included in the battery.  Power capability has an influence on weight and
also has a strong influence on cost because it determines how the materials are arranged as well
as the relative proportion of active materials to inactive materials in each cell.

   Because most PEV battery chemistries are known to experience degradation in power and
energy capacity over time (also known as power fade and capacity loss respectively), it is also
important to consider how performance at end-of-life might differ from beginning-of-life, and
consider the need for increasing the target capacity or power to ensure that performance goals
can be met for the life of the vehicle.

   The choice of battery energy capacity is primarily a function of the energy efficiency of the
vehicle and the target driving range. Because range may decline over time due to battery
degradation, this raises the question of whether the target range should be considered a
beginning-of-life  or end-of-life criterion. Current regulatory practice, as exemplified by the EPA
labeling guidelines for PHEVs and BEVs,548 measures range at beginning-of-life and omits any
adjustment for future capacity degradation. For PHEVs, however, current regulatory practice for
the EPA GHG standards  effectively requires vehicle manufacturers to consider degradation in
range as it will directly affect the calculated in-use emissions when tested for compliance at any
time during full useful life.AAA  Accordingly, for PHEVs, manufacturers typically use a
combination of battery oversizing and an energy management strategy that provides for a
consistent range throughout the useful life.  For BEVs, however, rather than oversizing the
battery sufficiently to maintain the original EPA range over time, manufacturers have tended to
make the customer aware of the possibility of range loss and in some cases have warranted the
battery to a specified degree of capacity  retention over a specified period of time. For example,
Nissan warrants their 24-kWh Leaf battery to retain nine of 12 capacity bars (corresponding to
about 70 percent capacity) for 60 months or 60,000 miles, and warrants their 30-kWh battery for
96 months or 100,000 miles.  As another example, Tesla does not warrant against a specific
degree of capacity loss but makes it clear that some capacity loss is normal and provides the
customer with recommendations for preserving battery capacity.

   The choice of battery power capability is primarily governed by vehicle performance
expectations.  In  the case of BEVs and many longer-range PHEVs, the battery  is sufficiently
large that its power capability is likely to naturally exceed that needed for acceleration
performance alone.  These batteries effectively have a power reserve that provides a natural
buffer against power fade.  Smaller batteries, such as those of shorter-range PHEVs, may lack
this advantage and may need to be sized deliberately to meet a target power capability, in which
AAA As noted in Section 5.3.4.3.4, PHEV GHG emissions are calculated using the SAE J1711 utility factor and
  AER. Accordingly, if range degrades during useful life, the utility factor correction would change and thus, the
  calculated GHG emissions would increase. As EPA's GHG emission standards are full useful life standards and
  vehicles are considered noncompliant if their emissions exceed the certified emission level by more than 10
  percent during the useful life, manufacturers must account for degradation or risk exceeding the GHG standards
  in-use.
                                              5-314

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
case power fade should be factored in to the sizing process because it could lead to loss of
performance and loss of utility factor over the life of the vehicle.

   At the time of the 2012 FRM, the task of assigning battery capacity and power for the many
PEV configurations to be analyzed was a very difficult task, with few well-developed techniques
and tools available.  Further, it was necessary to choose assumptions to reflect an expected state
of technology in the 2020-2025 time frame, even though few production vehicles were available
at the time to either serve as a reference for the current state of technology or to establish trends
for its advancement. As described below, the EPA methodology therefore employed a wide
variety of simplifying assumptions and estimation methods in order to conduct the effort in a
practical way while using calculation tools that are easily accessible to external reviewers.

   For the FRM analysis, EPA determined battery energy capacities and power capabilities for
modeled PEVs using a spreadsheet-based sizing methodology that was described in Section
3.4.3.8.1 of the 2012 TSD. Because battery capacity and power requirements are strongly
influenced by vehicle weight,  and battery weight is a function of capacity and power while also
being a large component of vehicle weight, sizing the battery for a BEV or PHEV requires an
iterative solution.  This problem is well suited to the iteration function available in common
spreadsheet software. A spreadsheet-based methodology was therefore selected as being
sufficiently powerful while remaining accessible to public inspection using standard
commercially available software. EPA used Microsoft Excel for this purpose, with the Iteration
setting enabled and set to 100  iterations.

   This Draft TAR analysis is based on the same methodology, with  significant refinements to
reflect developments in the industry since the FRM and to improve the fidelity of the sizing
estimates. The general methodology is reviewed below, followed by a review of the
refinements.

   EPA built a battery and motor sizing methodology to estimate the required battery capacity
and power output capability for a large array of modeled PEVs. The array included five
electrified vehicle types (EV75, EV100, EV200, PHEV20, and PHEV40), six baseline vehicle
classes of different curb weights (Small Car, Standard Car, Large Car, Small MPV, Large MPV,
and Truck); and five levels of target curb weight reduction (0, 2, 7.5, 10, and 20 percent).  This
resulted in a total of 150 PEV vehicle instances,BBB each characterized by a driving range, a
baseline curb weight, and a level of target curb weight reduction, as shown in Figure 5.108.  A
sizing spreadsheet determined battery energy capacities and battery power requirements for each
vehicle, in conjunction with ANL BatPaC which determined battery specific energy (kWh/kg)
for use by the sizing spreadsheet, and ultimately a pack cost estimate. Pack cost, electric drive
power ratings, and the necessary level of mass reduction applied to the glider (the baseline
vehicle minus powertrain components) for each vehicle were then utilized by the OMEGA
model.
               i5Q vehicles, two battery cathode chemistries (NMC622 and blended LMO/NMC) and four
  production volumes (50K, 125K, 250K and 450K) were also considered, resulting in the generation of 1,200
  individual battery cost estimates.
                                             5-315

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                    5 electrified
                    vehicle types
                       EV75
                      EV100
                      EV200
                      PHEV20
                      PHEV40
    X
 6 baseline
curb weights
   SmCar
   StdCar
   LgCar
  SmMPV
   LgMPV
   Truck
                  5 levels of
                 curb weight
                  reduction
X
 0%
 2%
7.5%
10%
20%
150 vehicle instances,
  characterized by
    i     I   I
                                              apply
                                                      up to 20% MR

Battery sizing
assumptions
kWh/kg
| Battery design
! assumptions

range
I CV\
1 J
%WR
hi .
- %MR -
to glider
r j
Sizing
spreadsheet
_//CV
i i
battery battery motor
kW kWh kW
i i i


f



ANL BatPaC


n
w\ y
1/ba-e ^^T| 	 *"
' CWPEV
1 — +•


to OMEGA
total applied
%MR
net %WR
from CWbase
electric drive
power (kW)
battery pack
DMC in 2025
(current $)
                    Figure 5.108 EPA PEV Battery and Motor Sizing Method
   Method for Sizing of Battery Energy Capacity

   Battery energy capacity was considered to be a function of desired driving range (mi) and
vehicle energy consumption (Wh/mi).

   Driving range was defined by the various range configurations (EV75, EV100, EV200,
PHEV20, and PHEV40) and was considered to be an approximate real-world, EPA-label range.
The 2012 FRM analysis considered PHEV range to be an all-electric range without assistance
from the engine under any vehicle operating conditions, and therefore all PHEVs in that analysis
were modeled with a range-extended electric vehicle (REEV) architecture rather than a blended-
operation architecture.  The Draft TAR analysis modifies this approach by adopting a blended
configuration for PHEV20 but retaining REEV configuration for PHEV40.

   Energy consumption had to be estimated by an appropriate method that took into account the
weight of the battery necessary to deliver this range, and many other factors.

   To estimate energy consumption for a given PEV instance, first its curb weight was estimated
as equal to the curb weight CWbase of the corresponding baseline conventional vehicle, modified
                                            5-316

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
by any applicable curb weight reduction WRtarget (2, 7.5, 10, or 20 percent), and further modified
by subtraction of the weight of conventional powertrain components (for BEVs) and addition of
the weight of electric content  (for BEVs and PHEVs), as shown in Equation 2 through Equation
5.
       Equation 2.  Target curb weight reduction
                               WRtarget = %WR*CWbase

       Equation 3. Weight-reduced curb weight

                           CWbase reduced = CWbase - WRtarget
       Equation 4.  Raw curb weight of BEV
                CWBEV — CWbase reduced ~~ WlCE_powertrain + Welectric

       Equation 5.  Raw curb weight of PHEV

                        CWPHEV = CWbase reduced + Weiectric content
:_content
   The curb weights CWbase of conventional baseline vehicles (detailed in Table 5.109 on page
5-331) were derived from the applicable MY baseline fleet (MY2008 in the FRJVI, updated to
MY2014 in this Draft TAR analysis) for each vehicle class (Small Car, Standard Car, Large Car,
Small MPV, Large MPV, and Truck).

   The assumed weights of the removed conventional powertrain components (WicE_powertrain)
varied for the six vehicle classes and are shown in Table 5.100.
         Table 5.100 Baseline ICE-Powertrain Weight Assumptions (Pounds), By Vehicle Class
Class
Small car
Std car
Large car
Small MPV
Large MPV
Truck
Engine
250
300
375
300
400
550
Transmission*
125
150
175
150
200
200
Fuel system*
50
60
70
60
80
100
Engine mounts*
25
25
25
25
25
25
Exhaust
20
25
30
25
30
40
12V battery
25
30
35
30
40
50
Total
495
590
710
590
775
965
       Note:
       Transmission minus differential; fuel system 50% fill; engine mounts include NVH treatments.
   Electric content weight (Weiectdc content) consisted of estimated battery weight and electric drive
weight (motor and power electronics). Since the weight of this content is strongly influenced by
total vehicle weight and many other variables, it is not a constant figure but is iteratively
computed by the spreadsheet. The computation included estimates of battery specific energy and
motor specific power applicable to the 2020 time frame. While the FRJVI used a fixed value for
specific energy, this Draft TAR analysis utilizes a direct link to BatPaC to pull in dynamically
updated values, as described later.  For BEVs, a gearbox weight of 50 pounds was also added.

   The "raw" curb weight calculations of Equation 4 and Equation 5, if used directly, would
typically generate estimated PEV curb weights that are significantly larger than the curb weights
of the baseline vehicles on which they are based, due to the added weight of the large battery
which may weigh more than the removed components. For several reasons noted below, EPA
                                              5-317

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
chose to further constrain the iteration by forcing the projected curb weight (CWBEv or CWPHEV)
of each PEV to match the curb weight (CWbase reduced) of the corresponding baseline vehicle. In
order to achieve this objective, EPA solved for the exact percentage of mass reduction that would
need to be applied to the glider in order to offset the difference in curb weight, and applied that
level of mass reduction to cause the curb weights to match.  In cases where more than 20 percent
mass reduction technology would have been necessary to offset the difference, it was capped at
20 percent and only in these cases was the curb weight of the electrified vehicle allowed to vary.

   In part, EPA chose to constrain the PEV curb weights in this way because it helps to
differentiate between "applied" mass reduction and "net" curb weight reduction throughout the
analysis. EPA differentiates between applied and net reduction because they are used in
different ways in the analysis. Net curb weight reduction refers to a reduction in curb weight,
and is used for estimating energy consumption. Applied mass reduction refers to percentage
mass reduction applied to the glider, and is used for estimating the cost of mass reduction
technology that has been embodied in the vehicle. Often, to achieve a given amount of net curb
weight reduction, more mass reduction technology might need to be applied to electrified
vehicles than to conventional vehicles because of the added weight of the electric content.

   For example, the FRM analysis indicated that a typical EV150 battery pack and associated
motors and other BEV-specific equipment may increase curb weight by roughly  18 percent.  As
a result, as shown in Table 5.101, an EV150 that applied 20  percent mass reduction technology
to the glider would have a net curb weight reduction of only about 2 percent.  In such a case,
EPA would base the estimate of EV150 mass reduction technology costs on a 20 percent applied
mass reduction, while basing the estimate of EV150 battery  and motor costs on battery and
motor sizings that are based on the energy and power requirements associated with only a 2
percent net curb weight reduction.
Table 5.101  Example Net Curb Weight Reduction for BEVs and PHEVs With 20% Applied Mass Reduction
                                       Technology

EV75
EV100
EV150
PHEV20
PHEV40
Actual %MR vs. base vehicle: 2008 Baseline (FRM)
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
19%
18%
19%
18%
18%
19%
14%
13%
13%
13%
13%
14%
2%
2%
2%
1%
1%
3%
12%
12%
12%
12%
12%
11%
7%
7%
7%
7%
7%
6%
   In theory, rather than constraining the PEV curb weights, a similar result could have been
achieved by applying the various weight reduction cases directly to the glider and allowing the
curb weights to grow as they might. This would have generated a different set of applied and net
reduction data points, with more data points representing little or no applied mass reduction,
higher curb weight, and higher energy consumption and larger batteries as a result. However,
because the high cost of battery capacity tends to improve the cost effectiveness of mass
reduction technology in PEV applications, EPA expects that manufacturers are likely to
implement significant mass reduction in most PEVs, meaning that cases with  little or no applied
mass reduction are of limited interest to the analysis.  The chosen method generates a greater
                                             5-318

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
density of points at the higher percentages of applied weight reduction that are most likely to
represent industry practice.

   After determining the PEV curb weight (which in most cases was constrained to match the
baseline curb weight, but now carries a specific degree of applied mass reduction in order to do
so), EPA then computed the loaded vehicle weight (also known as inertia weight or equivalent
test weight (ETW)) by adding 300 pounds to the curb weight:
       Equation 6. Equivalent test weight (ETW) of PE Vs
                                                        300

   EPA then used this test weight to develop an energy consumption estimate.  First, EPA
estimated the fuel economy (mi/gal) for a conventional light-duty vehicle (LDV) of that test
weight by a regression formula derived from the relationship between 2-cycle fuel economy and
inertia weight as described in the EPA Trends Report for MY2008 (from Table M-80 of the 2008
Trends Report).  Figure 5.109 depicts fuel economy trendlines derived from this source for all
LDVs, and also for cars and SUVs alone.
                                                                             A  mpgallSUV
                                                                             O  mpgallLDV

                                                                             <0>  mpgall cars
                                                                              --All SUV

                                                                           	•  _:••.

                                                                           	All Cars
     15CC'
2500
3500         4500

   Inertia weight (Ib)
55CC
6500
    Figure 5.109 Average LDV Fuel Economy Based On Inertia Weight from MY2008 FE Trends Data

   The MY2008 trendline was retained for this Draft TAR analysis because it represents the null
technology case, relative to which improvements in road load technology such as aerodynamic
drag and rolling resistance are accounted for.  As will be discussed later, electrified vehicles are
assumed to include a specific degree of aerodynamic drag and rolling resistance improvement
relative to the 2008 baseline.

   EPA used the All LDV fuel economy trendline (the solid black line) to characterize the
relationship between ETW and fuel economy for this analysis. Because the LDV fuel economy
trendline is derived from all MY2008 light-duty vehicles, it does not account for potential
differences in aerodynamic drag coefficients and frontal areas among the various vehicle classes
(for example, cars and MPVs, which are likely to have different  frontal area and aerodynamic
                                             5-319

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
features).  However, the All Cars trendline agrees well with the All LDV trendline, suggesting
that use of the LDV trendline is accurate for the car classes within the range of weights modeled.
Within the range of vehicle weights represented by MY2008 SUVs (3500 pounds and higher),
the differences in fuel economy are also small, suggesting that the LDV trendline is also
reasonably applicable to MPVs. EPA then derived a regression formula for the All LDV fuel
economy trendline, which is  shown in Equation 7.
       Equation 7. MY2008 conventional LDV fuel economy regression formula

      FEconv(mi/gal) = 0.0000017894 X ETWPEV2 - 0.0219693 X ETWPEV + 85.988

   This was then converted to a gross Wh/mile figure, assuming 33,700 Wh of energy per gallon
of gasoline as shown in Equation 8:
       Equation 8. Gross energy consumption (Wh/mile)

                         Egrossjn-pWh/mi) = (—^—) x 33,700
                                               F ^conv
   This figure was then brought into electrified vehicle space by applying a series of adjustments
representing assumed differences in energy losses between conventional vehicles and electrified
vehicles. This required making assumptions for several powertrain efficiencies:

   (a) Brake efficiency: For conventional vehicles, this is the percentage of chemical fuel energy
converted to energy at the engine crankshaft. For electrified vehicles, it is the percentage of
stored battery energy converted to shaft energy entering the transmission.  It therefore includes
battery discharge efficiency and inverter  and motor efficiency.

   (b) Driveline efficiency: the percentage of brake energy entering the transmission and
delivered through the driveline to the wheels. It includes transmission efficiency and
downstream losses (such as wheel bearing, axle, and brake drag losses), but not tire rolling
resistance.

   (c) Cycle efficiency: the percentage of energy delivered to the wheels that is used to overcome
road loads in moving the vehicle (that is, the portion of wheel energy that is not later lost to
friction braking).  This efficiency is larger for vehicles with regenerative braking.

   The efficiencies assumed  for baseline conventional vehicles were based on efficiency terms
derived from EPA's lumped  parameter model. Brake efficiency for conventional vehicles was
estimated at 24 percent, driveline efficiency at 81.3 percent, and cycle efficiency at 76.9 percent.

   In the FRM, brake efficiency for BEVs was estimated at 85 percent (the result of assuming a
roughly 95 percent efficiency for each of the battery (discharge),  motor, and power electronics).
Driveline efficiency was estimated at 93  percent (based on the value calculated by the lumped
parameter model for an advanced 6-speed dual-clutch transmission). Cycle efficiency was
estimated at 97 percent (representing regenerative braking recovering the bulk of braking energy
rather than dissipating it in friction brakes). EPA has since revised some of these values for the
current analysis as described later.

   PEV road  loads were also adjusted relative to conventional vehicles to  represent assumed
reductions in aerodynamic drag, rolling resistance, and vehicle weight applicable to these
                                             5-320

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
vehicles. All PEVs modeled for the 2012 FRM analysis were given a 10 percent reduction in
both aerodynamic drag and rolling resistance, in addition to the varying levels of net and applied
mass reduction. For example, in the case of an EV100 with a 20 percent mass reduction applied
to the glider (resulting in about 15 percent net curb weight reduction) and an assumed 10 percent
reduction in rolling resistance and aerodynamic drag, road loads (as calculated by the LP model)
were reduced to about 87 to 88 percent of the baseline conventional vehicle.

   The estimated energy consumption of each PEV is therefore derived from the energy
consumption of a corresponding baseline conventional vehicle by applying a ratio of the road
loads of the PEV (%RoadloadpEv) to those of the baseline vehicle (%Roadloadconv =1) and a
ratio of the assumed efficiencies of the respective powertrains, as shown in Equation 9.
       Equation 9. PEV unadjusted energy consumption

                     n,71,   .,    „           (%RoadloadP/EV   r]vehicle conv \
            EP/EV Frp(Wh/mi) = Egross FTP *   ————	*	-	
              1  ~                    ~      \%Roadloadconv   r]vehicleP/EV)

   Equation 9 yields a laboratory (unadjusted) two-cycle FTP energy consumption estimate. To
represent a real-world energy consumption, the 2012 FRM analysis applied a derating factor of
70 percent to convert unadjusted fuel economy to real-world fuel  economy. This is consistent
with the EPA 5-cycle fuel economy labeling rule as well as the EPA range labeling rule, both of
which specify a default derating factor for converting two-cycle figures to five-cycle figures.
The EPA range labeling rule specifies a default derating factor of 70 percent, with provisions for
using a different (custom) factor based on optional 5-cycle testing.

   In energy consumption space, a 70 percent derating of fuel economy corresponds to a 43
percent increase in energy consumption (1/0.70). Applying this factor (as shown in Equation 10)
results in the PEV on-road energy consumption estimate that EPA used to determine the required
battery pack capacity for the vehicle.ccc
       Equation 10.  PEV on-road energy consumption
                          Eonroad(Wh/mi) = EP/EV_FTP * (—)
       Finally, as shown by
   Equation 11, EPA determined the required battery energy capacity (BEC) as the on-road
energy consumption in Wh/mile, multiplied by the desired range in miles, divided by the usable
portion of the battery capacity, or usable SOC design window. The assumed usable SOC design
window (SOC%) varied between BEVs and PHEVs and is discussed in a later section.
ccc As described later, this Draft TAR analysis uses a 70 percent factor for most PEVs but applies a custom derating
  factor of 80 percent for EV200 based on examples of recent industry practice.


                                             5-321

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                                             Wh
                    Equation 11.  Required battery pack energy capacity for PEVs
   As mentioned previously, the intensively iterative nature of the battery capacity sizing
problem means that all of the preceding calculations are constructed in a spreadsheet as circular
references and performed iteratively by the spreadsheet software until the estimated weights,
ranges, and energy consumption figures converge.

   Method for Sizing of Battery Power Capability

   Another input to the battery sizing process is the required power capability of the battery.
Battery power capability was derived from an assigned peak motor power, which in turn was
considered to be a function of desired acceleration performance.

   In this analysis, PHEV40 was conceptualized as a range-extended electric vehicle, with a
motor and battery sized to be capable of providing pure all-electric range in all driving situations,
while PHEV20 was modeled as a blended-operation vehicle where the motor is often assisted by
the engine during the charge depletion phase. This means that PHEV40 motor power ratings in
this analysis are likely to be higher than would  apply to a blended-operation PHEV40. PHEVs
were configured with a single propulsion motor, in contrast to some production PHEV designs
that split the total power rating between two motors. Most PHEVs also include a second electric
machine used primarily as a generator.  The analysis does not explicitly assign a weight to this
component but considers it as part of the weight of the conventional  portion of the powertrain,
which retains its original weight despite the likelihood of downsizing in a PHEV application.

   In the FRM analysis, acceleration performance was represented by the average power-to-
weight ratio of conventional vehicles in each vehicle class. This meant that once the curb weight
for a PEV was estimated, a simple linear calculation determined the peak motor power needed to
meet the target power-to-weight ratio.  The battery power was then estimated as 15 percent
greater than the peak motor power, to account for losses in the motor. As with battery capacity,
motor and battery power both interact with battery and vehicle weight, and the calculation must
be performed iteratively in the spreadsheet as part of the overall battery sizing process.

   In preparation for this Draft TAR analysis, EPA studied trends  in PEV motor sizing in
production vehicles and used this information to improve the method for determining the
assigned peak motor power as a function of acceleration performance goals.  Other assumptions
were also revised. These improvements, along with those affecting capacity sizing, are described
below.

   Improvements to Battery Sizing Assumptions and Methodology

   Since the 2012 FRM, the emergence of a variety of production  PEVs has provided an
opportunity to validate the assumptions and methods of the 2012 FRM analysis. Further, the
industry appears to have begun proceeding toward stabilizing certain variables of PEV design
that help to constrain the battery sizing problem.  As a result, EPA has significantly updated and
refined the methods and input assumptions for assigning battery capacity, battery power, motor
power, and other aspects of the PEV modeling problem.  The major changes include:
                                             5-322

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
       (a) improvements to weight estimation for non-battery components;

       (b) improvements to weight estimation for battery packs;

       (c) improvements to the assignment of electric drive motor power;

       (d) updated curb weights, representing a 2014 baseline;

       (e) increase in usable battery capacity for BEVs and some PHEVs;

       (f) an increase in the assumed electric drive efficiency;

       (g) an increase in the battery power rating for PHEVs;

       (h) an increase in battery power to compensate for battery power degradation;

       (i) an increase in applied aerodynamic drag and rolling resistance reduction;

       (j) a change in range derating factor for EV200; and

       (k) a change in PHEV20 motor sizing to represent a blended PHEV configuration.

   These changes are described in detail in the following subsections (a) through (k).

   (a) Improved weight estimation for non-battery components

   At the time of the 2012 FRM, little data was available to characterize the weight of PEV non-
battery components (propulsion motor, power electronics, and cabling) due to the limited number
of PEV models being produced.  Weight of non-battery components was therefore estimated in
the 2012 FRM analysis as a function of total battery capacity, on the expectation that larger
vehicles with larger battery packs would generally require larger non-battery components.  The
FRM analysis thus estimated the combined weight of electric content (battery and non-battery
components together) by assuming an overall specific energy of 120 Wh/kg, assessed on total
battery capacity.  This figure embodied an assumed battery specific energy of 150 Wh/kg
combined with nominal estimates for the weight of non-battery content as  suggested by teardown
data and other sources.

   Ideally, the weight of electric power components would more properly be estimated by means
of a specific power metric (such as kW/kg) applicable to the component in question. An
appropriate metric could be determined by teardown study of a variety of electrified vehicles of
varying power capability.  Although EPA was unable to conduct additional teardown studies of
specific PHEVs or BEVs in time for this analysis, in the time since the FRM additional options
have become available for characterizing the specific power of non-battery components.

   Performance targets for non-battery components published by US DRIVE provide one
reference point.  US DRIVE549 is a consortium involving the U.S. Department of Energy,
USCAR (an organization of the major U.S. automakers), and several other organizations
including major energy companies and public energy utilities.  This industry collaboration has
established a number of cost and performance targets for automotive traction motors, inverters,
chargers, and other power electronics components for the 2015 and 2020 time frames.550 These
include targets for specific power of electric propulsion motors and power electronics, both
separately and alone, as shown in Table 5.102. These metrics are particularly relevant to the
problem of component sizing.
                                            5-323

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
           Table 5.102 U.S. Drive Targets for Non-Battery Specific Power for 2015 and 2020
Component
Electric motor and power electronics
Electric motor alone
Power electronics alone
U.S Drive Target (kW/kg)
2015
1.2
1.3
12
2020
1.4
1.6
14.1
   Since the EPA battery sizing methodology does not distinguish the power rating of the power
electronics from that of the drive motor, the US DRIVE target that would be most relevant to the
EPA analysis is the specific power of electric motor and power electronics combined, which US
DRIVE places  at 1.4 kW/kg for the 2020 time frame.

   This figure has some support in the literature.  A presentation by Bosch551 at The Battery
Show 2015 states that the electric motor and power electronics for a 100 kW, 20 kWh BEV
system in the 2025 time frame is expected to comprise about 37 percent of electric content
weight, with battery weight comprising the remaining 63 percent.  Assuming the 20 kWh battery
pack has a specific energy of about 140 Wh/kg (as indicated by BatPaC for an NMC622 pack at
115 kW net battery power), and a corresponding weight of 143 kg, the non-battery content would
be estimated at about 53 kg.  The 100 kW system would then  represent 100 kW/53 kg or 1.88
kW/kg, making the US DRIVE figure of 1.4 kW/kg appear conservative.

   Although the US DRIVE figures are targets and therefore not necessarily indicative of
industry status, EPA has confidence that the targets for specific power represent attainable goals
during the 2022-2025 time frame.  This is based in part on the observation that the 2020 specific
power target for electric motor and  power electronics combined is very close to levels that were
already being attained by some production vehicles at the time they were set.552 Also,
confidential business information conveyed to EPA through private stakeholder meetings with
OEMs conducted since the FRM suggests that some of these targets are already being met or
exceeded in production components today, or are expected to be met within the time frame of the
rule.

   This Draft TAR analysis therefore estimates the weight of non-battery PEV components using
the 2020 US DRIVE specific power target for motor and power electronics  combined, at 1.4
kW/kg.

   As mentioned above, teardown studies would be another source of validation.  As an
alternative to conducting its own teardown studies, EPA has collected data on xEV component
weights from a comprehensive teardown database produced by A2Macl,553 an automotive
benchmarking firm.  This database includes detailed weight analyses for the battery and non-
battery electrical components of several BEVs and PHEVs produced for U.S. and global markets
up to 2015. It therefore could provide a good source of data for the specific power of non-
battery components that were produced in the 2012-2015 time frame, for comparison with the
1.4 kW/kg US DRIVE target. Although EPA was unable to complete this analysis in time to
include it as part of this Draft TAR analysis, EPA plans to complete the analysis prior to the
proposed determination.

   (b) Improved weight estimation for battery components
                                            5-324

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   In the 2012 FRM analysis, EPA had estimated battery pack weights by applying a constant
specific energy value of 120 Wh/kg to account for the combined mass of the battery pack,
electric motor, wiring, and power electronics. This factor was applied to BEVs and PHEVs of
all driving ranges and was based in part on an assumed specific  energy of 150 Wh/kg for the
battery pack alone.

   In practice, the specific energy of a battery pack will vary depending on its power-to-energy
(P/E) ratio and its energy capacity. In general, smaller more power-optimized batteries tend to
show a lower specific energy than larger energy-optimized batteries.

   For this Draft TAR analysis, EPA therefore modified the method to allow the weight estimate
for the battery pack to be more sensitive to the P/E ratio of the battery. This was done by
directly linking the battery sizing spreadsheets to the BatPaC model to retrieve the specific
power computed by BatPaC for each individual battery pack. This greatly improves the
accuracy of the battery weight calculation. This adjustment causes the battery weight calculation
to increase slightly for PHEVs due to their typically higher P/E ratio,  and to decrease slightly for
longer-range BEVs.

   Accordingly, as shown by the selected examples in Table 5.103 and Table 5.104, the pack-
level specific energy figures EPA uses in this Draft TAR analysis  vary significantly, ranging
from about 140 to 180 Wh/kg for EV75 to EV200 (assuming NMC622 cathode), to about 140 to
145 Wh/kg for PHEV40 (also NMC622), and about 110 to 125 Wh/kg for PHEV20 (assuming
blended NMC/LMO cathode).
Table 5.103 Examples of Pack-Level Specific Energy Calculated By BatPac for Selected PEV Configurations
                                        (0% WR)


Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
EV75
(NMC622-G)
Wh/kg
142.4
146.3
141.5
150.1
159.8
161.0
P/E ratio
4.43
5.56
8.97
4.67
5.63
6.04
EV100
(NMC622-G)
Wh/kg
153.5
158.9
157.6
162.0
167.9
173.6
P/E ratio
3.32
4.17
6.73
3.50
4.23
4.53
EV200
(NMC622-G)
Wh/kg
162.0
170.6
171.1
169.3
175.6
180.5
P/E ratio
2.01
2.52
4.07
2.12
2.56
2.74
PHEV20(NMC75%/
LMO25%-G)
Wh/kg
117.9
118.1
111.2
120.2
124.3
125.4
P/E ratio
6.69
8.41
13.56
7.05
8.52
9.13
PHEV40
(NMC622-G)
Wh/kg
146.3
139.1
107.3
147.8
138.5
137.6
P/E ratio
7.17
9.01
14.54
7.56
9.13
9.79
                                             5-325

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.104 Examples of Pack-Level Specific Energy Calculated By BatPac for Selected PEV Configurations
                                       (20% WR)


Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
EV75
(NMC622-G)
Wh/kg
142.3
142.0
141.8
145.7
154.7
158.5
P/E ratio
4.12
5.17
8.42
4.32
5.20
5.52
EV100
(NMC622-G)
Wh/kg
149.1
154.8
157.0
157.9
163.9
169.1
P/E ratio
3.12
3.91
6.40
3.27
3.95
4.21
EV200
(NMC622-G)
Wh/kg
160.1
168.6
170.4
167.5
173.8
178.4
P/E ratio
1.94
2.45
4.01
2.05
2.47
2.63
PHEV20 (NMC75%/
LMO25%-G)
Wh/kg
116.5
117.2
111.2
118.8
123.9
125.4
P/E ratio
6.46
8.07
13.12
6.76
8.19
8.78
PHEV40
(NMC622-G)
Wh/kg
146.6
141.1
107.3
147.7
138.9
137.9
P/E ratio
7.00
8.81
14.54
7.37
8.93
9.65
   While these figures may appear very aggressive compared to batteries seen in 2012-2016MY
applications, it should be noted that the technology assumptions in BatPaC are forecasts for the
2020 time frame. For comparison, in January 2016, GM announced that the 60 kWh Chevy Bolt
BEV pack weighs 435 kg, suggesting that this EV200 pack has already achieved a specific
energy of 138 Wh/kg today.554 The same specific energy was already seen in the 85 kWh Tesla
Model S as early as 2012.555 Similarly, the 18.4 kWh pack of the 2016 Chevy Volt PHEV
weighs 183 kg, suggesting this PHEV53  pack has achieved 101 Wh/kg today. As has occurred
in the time since the FRM, the level of industry activity in battery development suggests that
similar advances are likely to continue through the 2022-2025 time frame.

   (c) Improved method for assignment of electric drive motor power

   In the FRM, in order to maintain acceleration performance equivalent to that of conventional
vehicles, EPA assigned power-to-weight ratios for PEVs to be equal to those of MY2008
conventional vehicles of their respective classes. Weight was modeled as equivalent test weight
(ETW), which is curb weight plus 300 pounds payload. Table 5.105 below shows the power-to-
ETW ratios assigned in the FRM for each vehicle class.
                  Table 5.105  Power-to-ETW Ratios Assigned to xEVs in the FRM
Class
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
hp/lb ETW
0.04364
0.05269
0.08101
0.04266
0.05289
0.05825
kW/kg ETW
0.07175
0.08662
0.13318
0.07013
0.08695
0.09576
   These ratios were derived from published engine power ratings of conventional vehicles.
However, it is well known that electric motors develop torque and power differently from
internal combustion engines, and so may translate a rated power to an acceleration performance
differently as well. Therefore, EPA conducted further analysis to determine whether targeting
PEV acceleration performance by sizing PEV motor power ratings based on engine power
ratings is appropriate.
                                            5-326

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   One of the most common metrics of acceleration performance is the time it takes a vehicle to
accelerate from zero to sixty miles per hour, also known as the O-to-60 time. Although there are
other metrics that describe acceleration performance, including metrics such as O-to-30 time, 30-
to-60 time,  and quarter-mile time (and grade-ability metrics as well), O-to-60 time is likely the
most familiar metric for understanding the acceleration performance of a vehicle.

   While in widespread popular use, this metric is not reported by manufacturers to EPA nor is
its measurement subject to uniform standards.  As an alternative, acceleration times of vehicles
with conventional powertrains are sometimes estimated by means of a methodology developed
by Malliaris et al.556 The Malliaris methodology predicts O-to-60 time as a function of the
power-to-ETW ratio of the vehicle and two numerical coefficients empirically obtained from a
least-squares fit of vehicle performance data. The Malliaris equation is depicted in Equation 12
below, with the coefficients 0.892 and 0.805 representing conventional vehicles with automatic
transmissions.
                                t =
                                                   -0.805
                      Equation 12. Malliaris equation for 0-60 acceleration time in seconds
   At the time of the FRM, EPA had historically used this equation and coefficients to estimate
acceleration performance of vehicles for pre-2014 editions of the annual Trends Report.557
Subsequent editions have used a newer method developed by MacKenzie et al.558 that EPA
believes to be more accurate, particularly for newer vehicles. The latter method relies on a more
detailed set of input parameters and tends to estimate slightly faster O-to-60 times than the
previous method.   By the MacKenzie method, average O-to-60 time for cars in MY2008 was at
8.9 seconds and fell to 8.4 seconds in MY2014 (with trucks falling from 9.0 seconds to 8.1
seconds).  The MacKenzie method is not directly applicable to electric powertrains due to the
requirement for ICE-specific inputs.

   The existence of these methods means that power-to-ETW ratios assigned to PEVs in the
FRM can therefore be converted to  approximate acceleration times (for the ICE-powered
conventional vehicles on which they were based).  Since the Malliaris method was in effect at
the time of the FRM, that method is used to estimate the  0-60 times depicted in Table 5.106
below. By this method, the power-to-weight ratios assigned to PEVs in the FRM analysis were
equivalent to 0-60 acceleration times between 8.8 and  11.3 seconds, with Large Car an outlier at
6.75 seconds.
 Table 5.106 Estimated 0-60 mph Target Acceleration Times Corresponding to FRM Assumptions for xEV
                                        hp/lb ETW
Class
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
hp/lb ETW
0.04364
0.05269
0.08101
0.04266
0.05289
0.05825
0-60 mph
(sec)
11.1
9.5
6.8
11.3
9.5
8.8
                                             5-327

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The practice of using ICE-based hp/lb to size the electric propulsion motor of an xEV
assumes that the power ratings of electric powertrains translate to acceleration times in the same
way as the power ratings of conventional powertrains.  At the time of the FRM, the small
number of production BEVs made it difficult to validate this assumption.

   Since the FRM, a significant number of BEV models have entered the market and now
provide an opportunity to better predict BEV acceleration performance as a function of motor
power and weight. Although comprehensive estimates of 0-60 acceleration time are not
published by any single authority, estimates for many PEVs have been published by
manufacturers and press organizations and provide a readily available source of empirical data.

   Figure 5.110 plots the approximate 0-60 mph acceleration times of MY2012-2016 BEVs and
PHEVs as a function of their power-to-ETW ratio, as expressed by rated peak motor power (kW)
divided by test weight (the published curb weight in kg, plus 136 kg payload).DDD  Acceleration
times were collected from publicly available sources including manufacturers and  press
organizations, and in some cases were averaged when estimates from different sources had slight
variation.  PHEVs for which an all-electric (battery only) acceleration time could not be
established were not included.

   An empirical trendline was derived from this data and is shown in the Figure as a thin orange
line. For comparison, the acceleration times that would be predicted by the Malliaris equation
for the same range of power-to-ETW ratios is shown in the Figure as a heavy black line.  As
shown by Equation 13, the empirical trendline has the same equation form as the Malliaris
equation, but with different coefficients  of 0.9504 and 0.795 that result from a least-squares fit to
the PEV data as expressed  in SI units for power and weight.

           is
g
g
-
2
:
                                                                    • EV and PHEV

                                                                   	ICE (Maliaris)
                   y = 0.9504x-° ™
             D  002 0.04 0.06 008 0.1 0.12 0 14 0.16 0.18 0.2 0.22 0.24 0.26
                           Power-to-ETW ratio (kW/kg)

    Figure 5.110 Acceleration Performance of MY2012-2016 PEVs Compared To Targets Generated By
                                    Malliaris Equation
DDD Tesla high-performance vehicles represented by 85 kWh Model S.
                                             5-328

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                               t = 0.9504
                                             kW
                                                    -0.795
       Equation 13. Empirical equation for 0-60 all-electric acceleration time of MY2012-2016 PEVs
   The plot of Figure 5.110 suggests that use of the Malliaris equation to size the motor power
rating of an electric powertrain results in higher power ratings and faster acceleration times for
PEVs than intended in the FRM. For example, to target a 0 to 60 mph acceleration time of 10
seconds, the Malliaris equation (shown by the heavy line) would indicate that the motor should
be sized to achieve a power-to-ETW ratio of 0.08 kW/kg. However, the empirical PEV trendline
indicates that this power-to-ETW ratio would actually provide an electric powertrain with an
acceleration time of about 7 seconds.  The degree to which 0-60 performance was likely over
specified in the FRM is shown in Table 5.107.  It appears that the 2012 FRM therefore assumed
significantly greater motor power ratings (and by extension,  battery power ratings) than required
for the intended acceleration times.
   Table 5.107 PEV Acceleration Performance Intended in the FRM and Projected Probable Performance

Class
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
0-60 mph time (sec)
FRM intent
11.1
9.5
6.8
11.3
9.5
8.8
FRM actual
7.7
6.6
4.7
7.9
6.6
6.1
   One option for improving the assignment of PEV power ratings would adopt the empirical
trendline of Equation 13 in place of the Malliaris equation to assign the necessary motor power
to match the originally targeted performance levels for each vehicle class. According to the EPA
Trends Report for 2015, average O-to-60 time for cars in MY2014 as estimated by the Malliaris
method was equal to that of MY2008 at 9.6 seconds  (with trucks showing a slight performance
increase from 9.7 seconds to 9.1 seconds), suggesting that the original power-to-ETW ratios
targeted in the FRM remain reasonably valid for the  current time frame.

   A second option would adopt the empirical trendline of Equation 13 while also updating the
power-to-ETW ratios to values more representative of today's fleet. This option retains good
comparability with the original methodology,  while allowing the performance targets to be
updated to reflect changes in the fleet since MY2008.

   EPA has therefore updated the power-to-ETW targets for each PEV vehicle class to values
derived from the MY2014 baseline.  These new values are shown in Table 5.108.
                                             5-329

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
         Table 5.108 Changes in PEV Power-To-Weight Ratios and 0-60 Targets for Draft TAR

Class
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Power-to-weight ratio
(hp/lb ETW)
FRM
0.04364
0.05269
0.08101
0.04266
0.05289
0.05825
Draft TAR
0.04718
0.05916
0.09740
0.05000
0.06205
0.06569
Estimated
equivalent
0-60 time (sec)
FRM
11.1
9.5
6.8
11.3
9.5
8.8
Draft
TAR
10.4
8.7
5.8
9.9
8.4
8.0
   The revised power-to-ETW values are slightly greater than the values assumed in the 2012
FRM, leading to slightly faster acceleration times.  EPA has carefully considered whether it is
appropriate to target greater power levels in this Draft TAR analysis, since this would tend to
divert some of the anticipated GHG benefit of the modeled vehicles toward vehicle performance
rather than GHG reduction. However, increased performance has in many cases been a factor in
the marketing of some PEVs, with many production and announced PEVs targeting faster
acceleration times than similarly appointed conventional vehicles.

   This adjustment to motor sizing should therefore allow the EPA PEV modeling methodology
to better match the power-to-weight ratios and acceleration performance that PEV manufacturers
appear to be following.  Assigning a more accurate power rating to PEV powertrains will allow
greater fidelity in the projected cost of both the battery and non-battery components of PEVs.
Further, basing the motor power sizing explicitly on an empirically derived estimate of 0-60
acceleration time for each modeled vehicle will more clearly demonstrate the performance
neutrality of the modeled PEVs.

   (d) Updated baseline curb weights

   For the FRM, the target curb weights for the six vehicle classes were based on the MY2008
baseline.  For this Draft TAR, the baseline was updated to MY2014. Also, PEVs were removed
from the sample to better represent the weight and performance of conventional vehicles alone.
Accordingly, the curb weights serving as inputs to the battery pack sizing analysis were updated
to these non-PEV  MY2014 values. Most curb weights increased, with the exception of Small
Car and Standard Car which declined slightly. The new weights are shown in Table 5.109
below.
                                            5-330

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
       Table 5.109 Changes to Baseline Curb Weights from FRM MY2008 to Draft TAR MY2014
Vehicle Class
Small Car
Standard car
Large car
Small MPV
Large MPV
Truck
Curb weight (Ib)
FRM
(MY2008)
2633 Ib
3306 Ib
3897 Ib
3474 Ib
4351 Ib
5108 Ib
MY2014
2628 Ib
3296 Ib
4117 Ib
3500 Ib
4448 Ib
5161 Ib

Change
-0.19%
-0.30%
+5.65%
+0.75%
+2.22%
+1.04%
   (e) Increase in usable battery capacity for BEVs and some PHEVs

   Based on observations of trends in recent BEV and PHEV usable capacity (discussed in
Section 5.2), the usable battery capacity was increased to 85 percent for EV75 and EV100, and
to 90 percent for EV200. The use of 90 percent for EV200 was chosen on the recognition of two
advantages associated with particularly high-capacity battery packs. First, because the total
available range is significantly larger than the average daily trip distance, vehicles with a long
driving range may on average utilize a smaller portion of the total battery capacity on a daily
basis, leading to generally shallower charge-discharge cycles. Also, these longer-range vehicles
require fewer charge-discharge cycles over the life of the battery to achieve a given lifetime
mileage.  Both factors may act to widen the usable portion of the battery for the purpose of
measuring maximum range without unduly affecting battery life in typical use.

   Since the battery of a PHEV40 is similar in size to that of a BEV, and based in part on the
Chevy Volt example, the usable capacity for PHEV40 was increased from 70 percent to 75
percent. PHEV20 remained at 70 percent due to the  smaller size of the battery.

   (f) Increase in electric powertrain brake and driveline efficiency

   In the 2012  FRM, brake efficiency and driveline efficiency for electric powertrains was
assumed to be  85 percent and 93 percent respectively (or 79 percent combined).  Since the 2012
FRM, some evidence has emerged that  some electric powertrains are already performing beyond
these levels. In 2013, a GM executive described the  drive unit of the yet-to-be-released Chevy
Spark EV as having an average DC  current-to-wheels efficiency of 85  percent in the city cycle
and 92 percent in the highway cycle559.  This current-to-wheels metric appears similar to the
product of brake and driveline efficiency, but neglecting battery discharge efficiency. Assuming
an average  battery discharge efficiency  of 95 percent, and a standard 55/45 city/highway
weighting (amounting to 88.15 percent  combined), the product of brake and driveline efficiency
for this powertrain would be about 83.75 percent.

   To bring the FRM assumptions closer to this figure, for this Draft TAR analysis EPA adjusted
the assumed brake and driveline efficiencies for BEVs to 87 percent and 95 percent respectively,
or 82.7 percent combined. Because the charge-depleting mode of a PHEV with AER is similar
in nature to BEV operation, brake efficiency for PHEVs was also increased to 87 percent, with
driveline  efficiency remaining at 93 percent to reflect the more complex nature of the PHEV
driveline.
                                             5-331

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   (g) Increase in PHEV battery power target

   In the 2012 FRM, the battery pack power requirement for BEVs was assigned as 15 percent
greater than the motor power rating. This adjustment represented estimated energy losses in the
electric motor, assuming an 85 percent motor efficiency. Battery sizing for PHEVs did not
employ this adjustment on the assumption that the engine could assist with acceleration.  In
retrospect, this assumption is inconsistent with PHEVs that operate as range-extended vehicles,
where all acceleration must be achieved by the battery and electric motor alone. Further, since
the FRM it has also appeared that some manufacturers of shorter-range, blended-operation
PHEVs are trending toward providing a stronger electric drivetrain capable of keeping the engine
off in a broader range of driving conditions. For these reasons, use of the adjustment factor has
been extended to PHEV battery sizing as well in order to better reflect an increased capability of
electric-only propulsion. Also, to reflect the assumed improvements in brake efficiency
described above, the factor for both BEVs and PHEVs is reduced from 15 percent to 10 percent
to reflect a 90 percent motor efficiency.

   (h) Allowance for power fade in battery power calculation

   As mentioned above, in the FRM analysis, the method of assigning motor power resulted in
motor and battery power sizing that was significantly greater than that observed in later
production PEVs.  Having modified the method to result in more representative (lower) motor
power ratings, battery power ratings are therefore also lower in the new analysis. This makes it
more critical to account for power fade during the life of the battery, since the new analysis no
longer over-sizes the battery as before.

   Battery power targets for PEVs were therefore nominally increased by an oversizing factor of
20 percent to compensate for power fade. In cases where a sufficiently large PEV battery
naturally results in an excess power capability greater than 20 percent, the oversizing factor does
not have an impact on the design of the battery.

   (i) Increase in applied aerodynamic drag and rolling resistance reduction

   In the construction of technology packages for the OMEGA analysis in the FRM, BEV and
PHEV technology packages  included an aerodynamic drag reduction of 20 percent (the
technology case known as AERO2), and a tire rolling resistance reduction of 20 percent (the case
known as LRRT2). This was based in part on the expectation that manufacturers would find
these technology improvements to be more cost effective for plug-in vehicles than for
conventional vehicles due to the potential to reduce the size and cost of the battery.  The package
costs thus reflected the cost of application of AERO2 and LRRT2 relative the 2008 baseline.
However, the battery  sizing methodology of the FRM applied only a 10 percent reduction in each
(AERO1 andLRRTl).

   For consistency with the rest of the analysis, EPA has now revised the battery sizing
methodology to apply AERO2 and LRRT2 in determining PEV energy consumption
requirements. This adjustment causes the assumed costs to be more representative of the
assumed level of technology application, and also tends to slightly reduce the estimated battery
capacity for a given range target.

   (i) Increase in derating factor for EV200
                                            5-332

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   For certification purposes, EPA allows manufacturers to either use a default derating factor of
70 percent to convert a two-cycle range test result to a label value, or to derive a custom derating
factor by undergoing complete five-cycle testing. Since the FRM, EPA certification data for
2012-2016MY EVs indicates that most BEV manufacturers have chosen to apply the default 70
percent derating factor in their certification tests. Tesla Motors is the only BEV manufacturer
that has elected to use a custom derating factor derived from 5-cycle testing.  Tesla has used a
factor of 79.6 percent for the standard Model S configurations from 60 kWh to 90 kWh, and a
factor ranging from 73 to 75 percent for higher-performance and AWD configurations of the
Model S and Model X.  Since the nearest current production example of an EV200 is the Tesla
Model S standard configuration, this Draft TAR analysis adopts a derate factor of 80 percent for
EV200.  Because manufacturers of EV75 and EVIOO-type vehicles have only used the default 70
percent derating factor and have not derived custom factors, EPA has retained the 70 percent
derating factor for EV75 and EV100. While these derating factors therefore represent the most
recent trends in industry practice since the 2012 FRM, their appropriateness in modeling the
label range of future PEVs will depend on the degree to which manufacturers continue to follow
this pattern in selecting the derating factors used for certification.

   (k) PFEV20 motor sized  for blended operation rather than EREV with AER

   Primarily in order to accommodate the high power requirements of the Large Car class as
modeled in this analysis, the PFLEV20 was assigned a lower motor power rating more in line
with a blended-architecture PFLEV rather than the EREV configuration of PFLEV40. The blended
motor power requirement was estimated as half of the power that would have been assigned to an
EREV configuration. Modeling of PFLEV20 as a blended PFLEV is also consistent with the
observation that many sub-20 mile PHEVs operate with at least a partially blended operating
strategy rather than a strict EREV strategy that allows all-electric operation in all driving
conditions. The reduction in motor power also allows the battery for Large Car to be sized with
reasonable power requirements compatible with the specific chemistry formulations modeled in
BatPaC.

   Summary of Changes to Battery Sizing Assumptions

   Table 5.110 reviews the major input assumptions to the battery sizing method and the changes
that were made for this Draft TAR analysis.
                                            5-333

-------
                    Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.110 PEV Battery Sizing Assumptions and Changes from FRM to Draft TAR
Assumption
Small Car base curb weight
Standard car base curb weight
Large car base curb weight
Small MPV base curb weight
Large MPV base curb weight
Truck base curb weight
Applied aero reduction from 2008 baseline
Applied tire reduction from 2008 baseline
Applied mass reduction to glider from 2008 baseline
Short range BEV (mi)
Mid-range BEV (mi)
Long range BEV (mi)
Short range PHEV (mi)
Long range PHEV (mi)
Usable battery capacity, HEV
Usable battery capacity, PHEV20
Usable battery capacity, PHEV40
Usable battery capacity, EV75
Usable battery capacity, EV100
Usable battery capacity, EV150/200
Electric content specific energy
Battery specific energy
Non-battery specific power
Motor sizing
Brake efficiency, PEV
Driveline efficiency, BEV
Cycle efficiency, PEV
BEV battery power as fn of motor power
PHEV battery power as fn of motor power
Allowance for power fade
Road loads, PEV
2-cycle to 5-cycle derating factor, PHEV and EV75/100
2-cycle to 5-cycle derating factor, EV200
PHEV20 motor sizing basis
2012 FRM
2633 Ib
3306 Ib
3897 Ib
3474 Ib
4351 Ib
5108 Ib
10%
10%
Varies; max 20%
EV75
EV100
EV150
PHEV20
PHEV40
40%
70%
70%
80%
80%
80%
120 Wh/kg
included with
electric content
included with
electric content
Based on MY2008
baseline ICE hp/lb
for each vehicle
class
85%
93%
97%
1.15x
Ix
none
from LPM
70%
70%
EREV
2016 Draft TAR
2628 Ib
3296 Ib
4117 Ib
3500 Ib
4448 Ib
5161 Ib
20%
20%
unchanged
unchanged
unchanged
EV200
unchanged
unchanged
unchanged
unchanged
75%
85%
85%
90%
N/A
Wh/kg computed by BatPaC
1.4 kW/kg
Based on MY2014 baseline
0-60 performance estimate
and new empirical equation
for PEVs
87%
95%
unchanged
l.lx
l.lx
20%
unchanged
unchanged
80%
blended
                                  5-334

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Analysis of Changes

   The changes above result in significant changes to the projected sizing of PEV batteries and
motors compared to those of the FRM. Table 5.111  shows examples of the battery capacities
and motor power ratings generated by the revised sizing methodology and compares them to the
corresponding estimates generated by the FRM analysis.

   It can be seen that battery capacity estimates have declined under the new methodology. It can
also be seen that estimated motor power ratings have declined in all cases (even for EV200,
despite the increase in range and vehicle weight vs. EV150). The declines in motor power are
largely the result of using the empirical trendline equation to assign the motor power rating
necessary  for the desired acceleration performance. For PHEV20, the motor power declines are
also the result of adopting a blended powertrain architecture in place of an EREV architecture,
which leads to lower motor power rating.
                                            5-335

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
  Table 5.111 Example Changes in Projected PEV Battery Capacity and Motor Power, FRM to Draft TAR
                                (20% weight reduction case)

EV75
EV100
EV150*/200**
PHEV20
PHEV40
FRM (2008 baseline)

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Battery
(kWh)
20.5
25.2
29.9
26.7
33.6
38.6
Motor
(kW)
77.5
115.5
206.2
98.7
150.0
189.6
Battery
(kWh)
28.2
34.7
41.1
36.7
46.5
53.0
Motor
(kW)
82.6
123.0
219.6
105.1
160.0
201.8
Battery
(kWh)*
45.3
55.8
66.2
59.0
74.8
85.3
Motor
(kW)*
94.0
139.6
249.5
119.2
181.9
229.1
Battery
(kWh)
6.5
8.0
9.5
8.4
10.7
12.4
Motor
(kW)
84.3
124.8
223.5
105.5
161.6
209.6
Battery
(kWh)
13.4
16.4
19.5
17.3
21.9
25.4
Motor
(kW)
88.9
131.4
235.4
111.1
170.3
220.6
Draft TAR analysis

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Battery
(kWh)
17.3
21.4
27.7
22.7
29.3
33.0
Motor
(kW)
54.0
83.8
176.8
74.5
115.3
138.3
Battery
(kWh)
23.5
29.1
37.4
30.9
39.8
44.6
Motor
(kW)
55.6
86.2
181.6
76.6
119.0
142.3
Battery
(kWh)**
41.2
50.2
65.0
53.7
69.2
77.6
Motor
(kW)**
60.6
93.4
197.4
83.4
129.5
154.7
Battery
(kWh)
6.1
7.5
9.5
7.9
10.2
11.7
Motor
(kW)
29.7
45.8
94.6
40.6
63.2
78.0
Battery
(kWh)
11.7
14.4
18.8
15.1
19.7
22.6
Motor
(kW)
61.9
96.3
206.7
84.6
133.5
165.0
Change from FRM

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Battery
(kWh)
-15.6%
-15.1%
-7.4%
-15.0%
-12.8%
-14.5%
Motor
(kW)
-30.3%
-27.4%
-14.3%
-24.5%
-23.1%
-27.1%
Battery
(kWh)
-16.7%
-16.1%
-9.0%
-15.8%
-14.4%
-15.8%
Motor
(kW)
-32.7%
-29.9%
-17.3%
-27.1%
-25.6%
-29.5%
Battery
(kWh)t
-9.1%
-10.0%
-1.8%
-9.0%
-7.5%
-9.0%
Motor
(kW)t
-35.5%
-33.1%
-20.9%
-30.0%
-28.8%
-32.5%
Battery
(kWh)
-6.2%
-6.3%
0.0%
-6.0%
-4.7%
-5.6%
Motor
(kW)tt
-64.8%
-63.3%
-57.7%
-61.5%
-60.9%
-62.8%
Battery
(kWh)
-12.7%
-12.2%
-3.6%
-12.7%
-10.0%
-11.0%
Motor
(kW)
-30.4%
-26.7%
-12.2%
-23.9%
-21.6%
-25.2%
       Notes:
       *ForEV150
       **ForEV200
       f Compares EV200 (Draft TAR) to EV150 (FRM)
       f f Compares blended PHEV20 (Draft TAR) to EREV PHEV20 (FRM)
   The following figures compare the newly projected battery capacities to those observed in
MY2012-2016 BEVs and PHEVs.  Both figures show that the revised methodology produces
capacity estimates that center more accurately on the 2012-2016 trendline than did the analogous
FRM estimates reviewed in Section 5.2).
                                            5-336

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                                                                      •  MY 2012-2016 BEVs
                                                                      O  draft TAR
                                                                     	Trend ine
                           100       150       200
                           EPA estimated range (mi)
                                250
300
     Figure 5.111 Comparison of Draft TAR Projected BEV Battery Capacities to MY2012-2016 BEVs
                                                                     o  draft TAR
                                                                     •  MY 2012-2016 PHEVs
                                                                    	Trend line
               1C
20     30      40     50     60
  EPA estimated electric range (mi)
3C
   Figure 5.112  Comparison of Draft TAR Projected PHEV Battery Capacities to MY2012-2016 PHEVs
   To compare the Draft TAR capacity projections to specific production vehicles, Table 5.112
and Table 5.113  show the projected battery capacities and assumed curb weights for each
electrified vehicle type and vehicle class at 0 percent and 20 percent nominal weight reduction,
                                              5-337

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
respectively.  These tables are useful for drawing comparisons of the projected battery capacities
to those of specific production BEVs and PHEVs. In the battery sizing analysis, differences in
energy consumption among the six vehicle classes (Small Car to Truck) is primarily derived
from differences in curb weight. Therefore matching a production vehicle's curb weight, range
and capacity to the values in these tables provides a fair comparison regardless of whether the
indicated classification or weight reduction case matches that of the vehicle.

  Table 5.112 Draft TAR Projected Battery Capacities and Assumed Curb Weights, 0% Nominal Weight
                                        Reduction


Small Car
Std Car
LgCar
SmMPV
LgMPV
Truck
EV75 (NMC622)
Curb wt
(Ib)
2628
3296
4117
3500
4448
5161
kWh
19.5
23.9
30.2
25.4
32.7
37.2
EV100(NMC622)
Curb wt
(Ib)
2628
3296
4117
3500
4448
5161
kWh
26.0
31.9
40.3
33.9
43.7
49.6
EV200(NMC622)
Curb wt
(Ib)
2628
3296
4117
3500
4448
5161
kWh
42.9
52.7
66.6
56.0
72.1
82.0
PHEV20
(25NMC/75LMO)
Curb wt
(Ib)
2628
3296
4117
3500
4448
5161
kWh
6.4
7.9
10.0
8.4
10.8
12.3
PHEV40 (NMC622)
Curb wt
(Ib)
2628
3296
4146
3500
4448
5161
kWh
12.0
14.8
18.8
15.7
20.2
23.0
  Table 5.113 Draft TAR Projected Battery Capacities and Assumed Curb Weights, 20% Nominal Weight
                                        Reduction


Small Car
Std Car
LgCar
Sm MPV
LgMPV
Truck
EV75 (NMC622)
Curb wt
(Ib)
2119
2689
3505
2849
3617
4134
kWh
17.3
21.4
27.7
22.7
29.3
33.0
EV100(NMC622)
Curb wt
(Ib)
2192
2775
3607
2940
3741
4263
kWh
23.5
29.1
37.4
30.9
39.8
44.6
EV200 (NMC622)
Curb wt
(Ib)
2419
3029
3948
3227
4100
4660
kWh
41.2
50.2
65.0
53.7
69.2
77.6
PHEV20
(25NMC/75LMO)
Curb wt
(Ib)
2363
2968
3773
3129
3994
4701
kWh
6.1
7.5
9.5
7.9
10.2
11.7
PHEV40 (NMC622)
Curb wt
(Ib)
2474
3132
4148
3277
4237
4992
kWh
11.7
14.4
18.8
15.1
19.7
22.6
   In most cases, the projected capacities are reasonably close to those of production vehicles,
although somewhat larger. As one example, the 30 kWh trim of the Nissan Leaf was recently
announced as achieving an EPA range of 107 miles at a curb weight of 1515 kg (3340 Ib).  On a
curb weight basis, the closest match in the tables above would be EV100 Standard Car (Table
5.112) at 3296 Ib. The projected battery capacity for this vehicle is 31.9 kWh. While this figure
is larger than the 30 kWh capacity of the Leaf, it represents a vehicle with a 20 percent reduction
in  aerodynamic drag and rolling resistance from a 2008 baseline vehicle. If the Leaf applies
more reduction than this, it could achieve its 107 mile range with a smaller battery.

   As another example, the Chevy Bolt EV was announced in 2016 as  an EV200 with a 60 kWh
battery and a curb weight of 3580 Ib.  On a curb weight basis, the closest match in the tables
above would be EV200 Small MPV at 3500 Ib (this is also consistent with GM's description of
this vehicle as a "crossover").  The projected battery capacity is 56 kWh, compared to the 60
kWh of the Bolt. While the projected capacity is lower than that of the Bolt, the Bolt is 80
                                             5-338

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
pounds heavier than the example, and may have a driving range in excess of 200 miles (the
driving range of the Bolt has not been rated by EPA but is commonly described as possibly
exceeding 200 miles).

   As a third example, the 60 kWh version of the Tesla Model S achieved an EPA range of 208
miles (EV200) at an advertised curb weight of 1961 kg (4323 Ib). The closest EV200 match to
this curb weight in the tables above would be about halfway between the two examples of Large
MPV at 4100 and 4448 pounds (projected at 69.2 and 72.1 kWh respectively). The average
battery capacity of the two is 70.65 kWh.  While larger than the 60 kWh Tesla provides, part of
the difference might be explained by the slightly larger 208-mile range of the vehicle

   As a final example, the 2016 Chevy Volt PHEV achieves an EPA AER of 53 miles with an
18.4 kWh battery at a curb weight of 1607 kg (3543 Ib).  The closest match is to the PHEV40, 0
percent, Small MPV at 3500 Ib, which projects a 15.7 kWh battery.  The greater range of the
Volt (53 miles vs. 40 miles) obscures the comparison, but is directionally correct.

   By these examples, it is clear that the methodology tends to predict somewhat larger BEV
battery capacity than 2012-2016 MY production BEVs, leading to a conservative assessment on
the basis of battery capacity alone.

   This trend is more clearly shown by normalizing the projected capacities to curb weight.
Figure 5.113 compares the BEV battery capacity per unit curb weight (kWh/kg CW) projected
by the revised methodology against that of production BEVs that are most comparable to the
modeled vehicles. This comparison removes the effect of weight differences and more clearly
expresses the efficiency with which  gross battery capacity is converted to label range for a given
vehicle weight.  For the purpose of this plot, comparable BEVs are defined as BEVs that were
available as 2016MY vehicles, but with D variants of the Tesla vehicles excluded (due to their
dual-motor architecture which differs from other BEVs, and because only non-D variants were
certified using a range derating factor similar to the 0.8 factor that was assumed for EV200).
The Tesla Roadster, although not a 2016 vehicle, is included because of its powertrain
similarities with other single-motor Tesla vehicles.

   It is clear from this plot that the revised battery sizing methodology has significantly
improved its prediction of battery capacity per unit curb weight compared to the methodology
used in the 2012 FRM analysis. However, it does continue to assign BEVs a slightly higher
battery capacity per unit weight than seen in production BEVs of the same range.
                                            5-339

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                                                        • Comparable BEVs
                                                                        • draft TAR projected
                                                                   	2012 FRM
          0      50     100     150     200     250     300     350
                             EPA label range (mi)

   Figure 5.113 Projected BEV Battery Capacity per Unit Curb Weight Compared To Comparable BEVs

   Seen another way, the plot suggests that at least some current production vehicles have been
able to deliver a given range with slightly less battery capacity than this Draft TAR analysis
predicts for a future time frame. While this supports a conservative estimate, this trend deserves
further examination because the goal of the Draft TAR is to represent a future state of technology
in 2022-2025.

   There are several potential reasons why the capacity estimates generated by the battery sizing
methodology may not match the capacities observed in specific  production vehicles.

   As previously observed, there could be differences in assumed powertrain efficiencies or
differences in application of road load reducing technologies (mass reduction, aerodynamic drag
reduction, and rolling resistance reduction) between the production vehicles and the modeled
vehicles. For example, if xEV manufacturers are applying more than the 20 percent reduction in
aerodynamic drag and rolling resistance (from a baseline vehicle) assumed in the analysis, or are
applying more mass reduction, it could result in substantially smaller battery capacity
requirements. Also, the larger battery capacity of longer-range BEVs may slightly improve their
discharge efficiency relative to shorter range vehicles, because discharge would take place at a
lower C rate. Efficiency of regenerative braking might also improve slightly for these vehicles.
These factors could account for some of the disparity for longer-range vehicles.

   While it is tempting to consider calibrating the battery sizing methodology to the observed
2012-2016MY battery capacities (perhaps by simply assigning battery capacities based on the
2012-2016MY trendline shown above), this would compromise the analysis' accounting for the
cost of applied road load reduction technology, because the level of road load technologies
applied to the vehicles that compose the trendline is not known,  and probably varies from vehicle
to vehicle. For example, even if the application level for one EV75 were known, the larger
battery and weight of an EV100 or EV200 may have incentivized greater reductions which
would have to be accounted for accurately as well.
                                             5-340

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   In contrast, the current methodology applies known levels of road load reduction technology
in order to clearly account for its cost and allow extrapolation to other application levels.  If the
cost of applying road load technologies in excess of these levels is similar to the value of the
battery capacity saved, it is possible that smaller battery sizes could result, but not necessarily at
a lower net vehicle cost.

5.3.4.3.7.2    Battery Sizing Methodology for HEVs

   HEV battery packs were sized using a simpler methodology described below. This method is
continued in the current analysis.

   Because there is no "all-electric range" requirement for HEVs, battery pack sizes are
relatively consistent for a given weight class.  Furthermore, because battery pack sizes are at
least an order of magnitude smaller for HEVs than for all-electric vehicles, the sensitivity of
HEV vehicle weight (and hence energy consumption) to battery pack size is relatively
insignificant. For these reasons, a more direct approach (rather than an iterative process) works
for battery sizing of HEVs.

   In the FRM analysis and the current analysis, HEV batteries were scaled similarly to the 2010
Fusion Hybrid battery, based on a metric of nominal battery energy per pound of equivalent test
weight (ETW).  Although the Fusion battery utilized a nickel-metal hydride (Ni-MH) chemistry
in contrast to the lithium-ion  chemistries of the current analysis, the  energy window required for
hybrid operation and thus gross battery sizing is expected to be similar for either chemistry.

   The Fusion Hybrid Ni-MH battery had an ETW ratio of 0.37 Wh/lb. The battery was
understood to utilize a 30 percent usable SOC window.  The FRM analysis and the current
analysis assumes 40 percent for HEVs in the 2020 time frame. The rationale for this assumption
is outlined in more detail in Section 5.2.4.4.3. This results in a 25 percent reduction of the
energy capacity of the base Fusion battery, or a 0.28 Wh/lb ETW ratio.  This  value was used to
size strong HEV batteries for the analysis.

   In comparing anecdotal data for HEVs, EPA assumed a slight weight increase of 4-5 percent
for HEVs compared to baseline non-hybridized vehicles. The added weight of the Li-ion pack,
motor and other electric hardware were offset partially by the reduced size of the base engine.

5.3.4.3.7.3    ANL BatPaC Battery Design and Cost Model

   The U.S. Department of Energy (DOE) has established long term industry goals and targets
for advanced battery systems as it does for many energy efficient technologies. Prior to the  2012
FRM, Argonne National Laboratory (ANL) was funded by DOE to provide an independent
assessment of Li-ion battery costs because of their expertise in the field as one of the primary
DOE National Laboratories responsible for basic and  applied battery energy storage technologies
for future HEV, PHEV and BEV applications. This led to the development of a Li-ion battery
cost model, later named BatPaC.

   A basic description of the  battery cost model that formed the basis of BatPaC was published
in a peer-reviewed technical paper presented at EVS-24.560  ANL later extended the model to
include analysis of manufacturing costs for BEVs and HEVs as well has PHEVs.561 In early
2011, ANL issued a draft report detailing the methodology, inputs and outputs of their Battery
Performance and Cost (BatPaC) model.562  Soon after, EPA contracted a complete independent
                                             5-341

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
peer-review of the BatPaC model and its inputs and results for HEV, PHEV and BEV
applications.563  ANL also provided EPA with an updated report documenting the BatPaC model
that fully addressed the issues raised within the peer review.564 ANL has continued to develop
the model on an ongoing basis, adding several new features and refinements to the latest
version.565  For this Draft TAR analysis, EPA used Version 3.0 of BatPac, which was provided to
EPA on December 17, 2015.566

   BatPaC is based on a bill of materials approach in addition to specific design criteria for the
intended application of a battery pack. The costs include materials,  manufacturing processes, the
cost of capital equipment, plant area, and labor for each manufacturing step. The design criteria
include detailed parameters such as power and energy storage capacity requirements, cathode
and anode chemistry, and the number of cells per module  and modules per battery pack.  The
model assumes use of a stiff-pouch, laminated multi-layer prismatic cell, and battery modules
consisting of double-seamed rigid containers.  The model  supports both liquid-cooling and air-
cooling, with appropriate accounting for the resultant structure, volume, cost, and heat rejection
capacity of the modules.  The model takes into consideration the cost of capital  equipment, plant
area and labor for each step in the manufacturing process for battery packs and places relevant
limits on electrode coating thicknesses and other processes limited by existing and near-term
manufacturing processes.  The ANL model also takes into consideration annual pack production
volume and economies of scale for high-volume production.

   EPA chose to adopt the ANL BatPaC model for the following reasons.  First, BatPaC has
been described and presented in the public domain and does not rely upon confidential business
information (which would therefore not be reviewable by  the public).  The model was developed
by scientists at ANL who have significant experience in this area.  The model uses a bill of
materials methodology which the agencies believe is the preferred method for developing cost
estimates. BatPaC appropriately considers the target power and energy requirements of the
vehicle, which are two of the fundamental parameters when designing a lithium-ion battery for
an HEV, PHEV, or BEV.  BatPaC can estimate high volume production costs, which the
agencies believe is appropriate for the 2025 time frame.  Finally, its cost estimates are consistent
with some of the supplier cost estimates EPA received from large-format lithium-ion battery
pack manufacturers.  A portion of that data was received  from EPA on-site visits to vehicle
manufacturers and battery suppliers in 2008.

   Since the FRM, EPA has worked closely with ANL to  test new versions of BatPaC and to
guide the development of features that would support the midterm review and this Draft TAR
analysis.  ANL has since published several iterations of the model that incorporate updated costs,
improved costing methods and other improvements.

   EPA has worked closely with ANL since 2010 to evaluate each successive version of the
BatPaC model, to make suggestions for its improvement, and to specifically request features to
assist with its use for the purpose of battery costing for the rule. EPA also worked with ANL to
arrange for  an independent peer review of the model in 2011. This peer review along with EPA
input led to many improvements that were described in the TSD that accompanied the 2012
FRM.  ANL has continued to make improvements and add new features since the FRM, many at
EPA request.  Recent development has included: support for additional battery module
topologies,  improved modeling of impedance and electrode thickness, improved evaluation of
battery thermal capabilities, revised electrode chemistries  such as NMC622, improved
                                            5-342

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
accounting for plant costs and overhead, improved cost accounting for solvent recovery,
customization of cell thickness parameters, generation of US ABC parameters, and updated costs
for all constituent cell materials.

   To conduct this Draft TAR analysis, in December 2015 ANL provided EPA with a beta copy
of BatPaC Version 3. After testing and evaluation, this version was used in this Draft GHG
Assessment. A copy of this file is available in Docket EPA-HQ-OAR-2015-0827.

   Basic user inputs to BatPaC include performance goals (power and energy capacity), choice
of battery chemistry (of several predefined chemistries), the vehicle type for which the battery is
intended (HEV, PHEV, or BEV), the desired number of cells and modules and their layout in the
pack, and the volume of production.  BatPaC then designs the electrodes, cells, modules, and
battery pack, and provides a complete, itemized cost breakdown at the specified production
volume.

   BatPaC provides  default values for engineering properties and material costs that allow the
model to operate without requiring the user to supply detailed technical or experimental data. In
general, the default properties and costs represent what the model authors consider to be
reasonable values representing the state of the art expected to be available to large battery
manufacturers in the year 2020.  Users are able to edit these values as necessary to represent their
own expectations or their own proprietary data.

   In using BatPaC,  it is extremely important that the user monitor certain properties of the cells,
modules, and packs that it generates, to ensure that they stay within practical design guidelines,
adjusting related inputs if necessary.  In particular, pack voltage and individual cell capacity
should be limited to  appropriate ranges for the application. These design guidelines are not
rigidly defined, but approximate ranges are beginning to emerge in the industry.

   The cost outputs used by EPA to determine 2025  HEV, PHEV and BEV battery costs were
based on the inputs and assumptions described in the next section.  For engineering properties
and material costs, and for other parameters not identified below, EPA used the defaults provided
in the model.

5.3.4.3.7.4    Assumptions and Inputs to BatPaC

   EPA chose basic  user inputs to BatPaC as follows.

   For performance  goals, EPA used the power  and energy requirements derived from the
battery sizing analysis described in the previous section.  Additional inputs include battery
chemistry, vehicle type (BEV, PHEV, or HEV), cell and module layout, and production
volumes, as outlined below.

   In addition to these inputs, EPA monitored certain outputs to ensure that the resultant cell and
pack specifications were realistic. In particular, pack voltages, electrode dimensions, cooling
capability, and individual cell capacities were monitored to ensure that they were consistent with
current and anticipated industry practice.

   Additionally, EPA did not include warranty costs computed by BatPaC in the total battery
cost because these are accounted for elsewhere in the analysis by means of indirect cost
multipliers (ICMs).
                                             5-343

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Battery chemistry

   In the 2012 FRM analysis, chemistries were chosen due to their known characteristics and to
be consistent with both publicly available information on current and near term HEV, PHEV and
BEV product offerings from OEMs as well as confidential business information on future
products currently under development.  Therefore in that analysis, BEV and PHEV40 packs were
configured with NMC441 cathode chemistry, and PHEV20 and HEV packs were configured
with LMO cathode chemistry.  Although EPA considered NMC to be the preferred future
chemistry for all xEV packs at the time of the FRM, the choice of LMO was necessary due to the
relatively high power-to-energy ratio of PFIEV20 and FIEV, which precluded use of NMC as
modeled by BatPaC. All packs had a graphite anode chemistry.  These represented the most
appropriate chemistry choices among those offered in Version 2 of BatPaC at the time.

   Version 3 of BatPaC replaces NMC441 with NMC622, a more commonly cited formulation
of NMC567 with a long cycle life.568 A blended NMC/LMO cathode option was also added,
representing increasing popularity of blended cathodes over pure LMO.  Therefore in this Draft
TAR analysis, EPA selected NMC622 for BEV and PHEV40 packs, and a blended cathode (25
percent NMC and 75 percent LMO, the BatPaC default value) for PHEV20 and HEV packs.
Although most current Li-Ion HEV packs are reported to be using NMC cathodes,569 EPA found
it necessary to model a blended HEV cathode because the default NMC formulations modeled by
BatPaC  did not always support the power-to-energy ratios required by some of the modeled HEV
configurations. This might be due to variations in NMC formulations and particle morphologies
that manufacturers might employ to optimize the chemistry for HEV use but which are not
represented in the BatPaC default formulations.

   Pack topology and cell capacity

   In the 2012 FRM, the number  of cells per module for all packs had been fixed at 32 cells and
the pack topology (number of modules and their arrangement in rows) followed nominal rules
and was not optimized. In this  Draft TAR analysis,  EPA optimized the pack topology for BEVs
and PHEVs by choosing values for cells per module, number of modules and arrangement of
modules to target a preferred cell  capacity.

   Since the number of modules per pack must be a whole number, varying the number of cells
per module allows the number of cells per pack and their capacities to be better targeted. EPA
varied the number of cells per module to between 24 and 36. Based in part on the 55.5 Ampere-
hour cells that appear to be used by Nissan and GM in their recently announced 60-kWh packs,
and larger cell sizes currently produced or recently announced by leading suppliers, EPA
targeted an individual cell capacity of 60 A-hr for BEV packs (not to  exceed 75 A-hr) and 45 A-
hr for PHEV packs (not to exceed 50 A-hr). Although constraints such as pack voltage and pack
capacity prevent matching these targets exactly, cell capacities now cluster more closely to the
preferred values than in the 2012  FRM analysis. In  many cases this tends to reduce pack costs
by tending toward smaller numbers of cells of a larger capacity than assumed in the FRM.  HEV
packs, which consist of a single module, are configured with 32 cells  as before.

   Thermal management

   In the FRM, BEV and PHEV packs were modeled with liquid cooling while HEV packs were
modeled with passive air cooling.  Since BatPaC did not provide an option for passive air
                                           5-344

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
cooling in which only the outside of the pack is cooled, EPA substituted the BatPaC cooling
costs with costs derived from an FEV teardown of an HEV.570

   As before, the current version of BatPaC continues to provide an option for active air cooling
in which individual cells are separated by air passages through which cabin air or cooled air is
circulated. Use of this option results in package volumes that are much larger than for a liquid
cooled pack. Although passive air cooling continues to be prevalent in HEV packs at the time of
this writing, some industry sources have indicated that liquid cooling may also be preferable for
HEV packs in order to improve utilization of capacity and increase service life.  Minimization of
underhood package volume is also a growing concern.  EPA therefore chose to utilize liquid
cooling for HEV packs as well as BEV and PHEV packs for this Draft TAR analysis.

   Pack voltage

   For this Draft TAR analysis, EPA limited BEV and PHEV voltages to a slightly narrower
range to reflect expected standardization of power electronics voltages. Based on knowledge of
current practices and developing trends of battery manufacturers and OEMs, supplemented by
discussions with the BatPaC authors, EPA targeted allowable pack voltage to approximately
120V for HEVs (except 48V HEVs) and approximately  300-400V for BEVs and PHEVs.

   Electrode dimensions

   For electrode coating thickness, the 100 micron maximum limit used in the FRM analysis is
retained in this Draft TAR analysis.

   Recent developments in pack design (as described in Section 5.2.4.4.6,  Electrode
Dimensions) suggest that the industry may be moving toward low-profile or flat floor-mounted
packs. For this reason, in this Draft TAR analysis EPA has revised the 1.5:1 aspect ratio used in
the FRM analysis and now adopts the BatPaC default aspect ratio of 3:1.

   Manufacturing volumes

   The assumed manufacturing volume for BEV, PHEV and HEV battery packs was retained at
450,000 per year as in the FRM. For a full discussion of considerations with regard to the
assumed manufacturing volume, please refer to Section  5.2.4.4.7, Pack Manufacturing Volumes.

   Summary of Battery Design Assumptions

   Table 5.114 shows a summary of battery design assumptions used in the FRM and those
adopted for the Draft TAR analysis.
                                            5-345

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
 Table 5.114 Battery Design Assumptions Input to BatPaC and Changes from 2012 FRM to 2016 Draft TAR
Assumption
EV75 chemistry
EV100 chemistry
EV150/200 chemistry
PHEV20 chemistry
PHEV40 chemistry
HEV chemistry
Pack topology
Maximum cell capacity (A-hr)
Cells per module
BEV thermal medium
PHEV thermal medium
HEV thermal medium
BEV pack voltage range (V)
PHEV pack voltage range (V)
HEV pack voltage range (v)
Maximum electrode thickness (microns)
Electrode aspect ratio
BEV battery 2020 annual mfg volume
PHEV battery 2020 annual mfg volume
HEV battery 2020 annual mfg volume
2012 FRM
NMC441-G
NMC441-G
NMC441-G
LMO-G
NMC441-G
LMO-G
varies
80
32
Liquid
Liquid
Air
300V-600V
300-400
~120V
100
1.5:1
450000
450000
450000
2016 Draft TAR
NMC622-G
NMC622-G
NMC622-G
25%NMC/75%LMO-G
NMC622-G
25%NMC/75%LMO-G
optimized to target
preferred cell capacity
BEV: target 60, max 75
PHEV: target 45, max 50
24 to 32
unchanged
unchanged
Liquid
300V to 400V
300V to 400V
unchanged
unchanged
3:1
unchanged
unchanged
unchanged
5.3.4.3.7.5    Battery Cost Projections for xEVs

   Table 5.116 through Table 5.121 show the battery pack direct manufacturing costs (DMC)
that EPA used in this analysis, and the degree of change from those used in the FRM, for each
level of applied mass reduction technology. The costs are quoted in 2013 dollars and the
analysis assigns them to the year 2025 for EVs and PHEVs and the year 2017 for HEVs. This
assignment follows the convention used in the 2012 FRM analysis, where FLEV battery costs
were assigned to  the earlier year to reflect considerations such as the relatively larger number of
FLEV batteries that were in production relative to PFLEV and BEV batteries.

   The costs shown are BatPaC output figures minus warranty costs.  The warranty costs
computed by BatPaC are subtracted because the EPA analysis accounts for warranty costs by
means of indirect cost multipliers (ICMs).

   In the wider analysis, EPA uses these cost figures combined with a learning curve that assigns
battery costs for each year over the full time frame of the rule. This curve was developed by first
considering the BatPaC costs as applicable to the 2025 MY for EVs and PFffiVs and to the 2017
MY for FLEVs. EPA then unlearned those costs back to the present year using the curve shown
in Section 5.3.2.1.4. This allows EPA to estimate costs applicable to MYs 2017 through 2025.
                                             5-346

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
The changes in direct manufacturing costs from year-to-year reflect cost changes due to learning
effects.

   As shown in Table 5.115, projected battery costs for many electrified vehicle configurations
have fallen substantially from those projected in the FRM. These changes are the result of many
influences, including changes to cost assumptions and methodology inherent to BatPaC Version
3, changes to EPA sizing assumptions (usable battery capacity, motor power, energy efficiency,
etc.) that have in many cases resulted in reductions to gross battery capacity and power
requirements, and changes to EPA inputs to BatPaC (particularly, use of improved pack and
module topologies).
     Table 5.115  Average Change in Projected Battery Pack DMC from 2012 FRM to 2016 Draft TAR
Electrified
Vehicle Type
EV75
EV100
EV150/200
PHEV40
PHEV20
HEV
Average change
Change in
pack cost
-24.9%
-27.1%
-24.0%
-12.2%
-8.7%
29.6%
Change in cost
per kWh
-13.4%
-15.0%
-18.7%
-1.5%
-3.2%
27.7%
   Costs for EV75 and EV100 have fallen by an average of about 25 percent on a total cost basis
and by about 13 to 15 percent on a cost per kWh basis.  The main influences on this change stem
from improvement to pack topology and cell sizing, reductions in pack capacity and P/E ratio,
etc.

   Although EV200 costs are not directly comparable because the FRM modeled EV150,
projected costs have fallen by about 24 percent relative to EV150 despite the increase in range.

   PHEV40 battery costs have fallen by about 13 percent, having benefited from forces similar
to those that have reduced BEV costs, but not as much, because PHEV target battery power has
increased relative to the FRM.

   PHEV20 battery costs have decreased slightly. The main reason would be the decision to
model PHEV20 as a blended PHEV with half the electric motor power of the previous EREV
configuration.  The reduction due to this is reduced by the increase in PHEV battery power
requirements relative the FRM, which  increases the P/E ratio and accordingly the cost.

   FIEV costs have increased significantly by an average of almost 30 percent.  This appears to
be a result of the change to a mixed NMC and LMO cathode, increased costs projected by
BatPaC Version 3 relating in part to the use of thinner electrodes for a given power requirement,
and the use of BatPaC liquid cooling costs instead of the FEV teardown costs for air cooling that
were used in the FRM.  It should be noted that BatPaC  does not model passively cooled cell
assemblies without significant air flow passages between the cells, which would probably have a
lower cost than a liquid cooled pack. However, as modeled by BatPaC, liquid cooled HEV
packs have a slightly lower cost than the available air cooled options.
                                            5-347

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
          Table 5.116 Estimated Direct Manufacturing Costs in MY2025 for EV75 Battery Packs
EV75*
(450k/yr)
0% CWR
2% CWR
7.5% CWR
10% CWR
20% CWR
FRM (2008 baseline)

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$5,115
$ 6,021
$ 7,724
$ 5,995
$ 7,310
$ 8,332
$/kWh
$224
$215
$232
$203
$195
$193
Pack
$ 5,098
$ 5,965
$ 7,635
$ 5,952
$ 7,237
$ 8,242
$/kWh
$225
$215
$232
$204
$196
$193
Pack
$ 4,996
$ 5,818
$ 7,397
$ 5,843
$ 7,045
$ 8,005
$/kWh
$228
$216
$231
$206
$196
$193
Pack
$ 4,962
$ 5,755
$ 7,295
$ 5,800
$ 6,963
$ 7,883
$/kWh
$229
$216
$231
$207
$196
$194
Pack
$ 4,768
$ 5,509
$ 6,907
$ 5,625
$ 6,610
$ 7,474
$/kWh
$233
$219
$231
$211
$197
$194
Draft TAR

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$3,962
$4,411
$5,807
$4,514
$5,380
$5,856
$/kWh
$203
$184
$192
$177
$164
$157
Pack
$3,940
$4,391
$5,752
$4,489
$5,351
$5,805
$/kWh
$205
$186
$193
$179
$165
$158
Pack
$3,893
$4,331
$5,603
$4,431
$5,278
$5,674
$/kWh
$208
$189
$193
$182
$168
$159
Pack
$3,873
$4,308
$5,538
$4,406
$5,248
$5,614
$/kWh
$210
$190
$194
$183
$169
$159
Pack
$3,788
$4,203
$5,404
$4,301
$5,121
$5,457
$/kWh
$219
$196
$195
$189
$175
$165
Change from FRM

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
-22.5%
-26.7%
-24.8%
-24.7%
-26.4%
-29.7%
$/kWh
-9.4%
-14.2%
-17.2%
-12.7%
-15.9%
-18.5%
Pack
-22.7%
-26.4%
-24.7%
-24.6%
-26.1%
-29.6%
$/kWh
-9.0%
-13.7%
-17.0%
-12.5%
-15.5%
-18.3%
Pack
-22.1%
-25.6%
-24.3%
-24.2%
-25.1%
-29.1%
$/kWh
-8.6%
-12.6%
-16.4%
-11.9%
-14.2%
-17.6%
Pack
-21.9%
-25.2%
-24.1%
-24.0%
-24.6%
-28.8%
$/kWh
-8.4%
-12.1%
-16.1%
-11.7%
-13.7%
-17.6%
Pack
-20.5%
-23.7%
-21.8%
-23.5%
-22.5%
-27.0%
$/kWh
-5.9%
-10.1%
-15.5%
-10.3%
-11.0%
-14.8%
Note:
       CWR = target percent reduction in vehicle curb weight.
       Actual reduction will be less if it would require applying more than 20 percent mass reduction to glider.
       *NMC622 cathode.
                                                 5-348

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
         Table 5.117 Estimated Direct Manufacturing Costs in MY2025 for EV100 Battery Packs
EV100*
(450k/yr)
0% CWR
2% CWR
7.5% CWR
10% CWR
20% CWR
FRM (2008 baseline)

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$ 6,105
$ 7,054
$ 8,630
$ 7,293
$ 8,641
$ 9,962
$/kWh
$201
$189
$195
$186
$173
$173
Pack
$ 6,083
$ 7,001
$ 8,535
$ 7,237
$ 8,571
$ 9,879
$/kWh
$201
$189
$195
$186
$174
$174
Pack
$ 5,950
$ 6,826
$ 8,283
$ 7,096
$ 8,392
$ 9,676
$/kWh
$204
$190
$194
$188
$175
$175
Pack
$ 5,906
$ 6,770
$ 8,175
$ 7,039
$ 8,321
$ 9,554
$/kWh
$205
$191
$194
$189
$176
$176
Pack
$ 5,817
$ 6,662
$ 7,999
$ 6,953
$ 8,215
$ 9,392
$/kWh
$206
$192
$194
$190
$177
$177
Draft TAR

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$4,533
$5,306
$6,476
$5,404
$6,266
$6,266
$/kWh
$175
$166
$161
$159
$144
$135
Pack
$4,511
$5,278
$6,417
$5,374
$6,227
$6,227
$/kWh
$176
$167
$161
$160
$144
$135
Pack
$4,450
$5,207
$6,265
$5,342
$6,139
$6,139
$/kWh
$179
$170
$162
$164
$147
$137
Pack
$4,428
$5,179
$6,197
$5,312
$6,102
$6,102
$/kWh
$180
$171
$162
$165
$148
$138
Pack
$4,345
$5,095
$6,122
$5,223
$5,995
$5,995
$/kWh
$185
$175
$164
$169
$151
$142
Change from FRM

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
-25.8%
-24.8%
-25.0%
-25.9%
-27.5%
-37.1%
$/kWh
-13.1%
-11.9%
-17.3%
-14.1%
-17.2%
-22.3%
Pack
-25.8%
-24.6%
-24.8%
-25.7%
-27.3%
-37.0%
$/kWh
-12.7%
-11.6%
-17.1%
-13.9%
-17.0%
-22.1%
Pack
-25.2%
-23.7%
-24.4%
-24.7%
-26.8%
-36.6%
$/kWh
-12.3%
-10.4%
-16.5%
-12.5%
-16.3%
-21.5%
Pack
-25.0%
-23.5%
-24.2%
-24.5%
-26.7%
-36.1%
$/kWh
-12.0%
-10.1%
-16.3%
-12.3%
-16.0%
-21.4%
Pack
-25.3%
-23.5%
-23.5%
-24.9%
-27.0%
-36.2%
$/kWh
-10.6%
-8.9%
-15.9%
-10.9%
-14.7%
-19.9%
Note:
       CWR = target percent reduction in vehicle curb weight.
       Actual reduction will be less if it would require applying more than 20 percent mass reduction to glider.
       *NMC622 cathode.
                                                 5-349

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
         Table 5.118 Estimated Direct Manufacturing Costs in MY2025 for EV200 Battery Packs
EV200*
(450k/yr)
0% CWR
2% CWR
7.5% CWR
10% CWR
20% CWR
EV150 in FRM (2008 baseline)

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$ 8,080
$ 9,753
$ 11,120
$ 10,109
$ 12,114
$ 13,878
$/kWh
$177
$174
$167
$171
$162
$161
Pack
$ 8,048
$ 9,714
$ 11,073
$ 10,109
$ 12,112
$ 13,818
$/kWh
$178
$174
$167
$171
$162
$161
Pack
$ 8,048
$ 9,714
$ 11,073
$ 10,109
$ 12,112
$ 13,759
$/kWh
$178
$174
$167
$171
$162
$161
Pack
$ 8,048
$ 9,714
$ 11,073
$ 10,109
$ 12,112
$ 13,759
$/kWh
$178
$174
$167
$171
$162
$161
Pack
$ 8,048
$ 9,714
$ 11,073
$ 10,109
$ 12,112
$ 13,759
$/kWh
$178
$174
$167
$171
$162
$161
EV200 in Draft TAR

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$6,712
$7,394
$8,851
$7,734
$9,160
$9,795
$/kWh
$156
$140
$133
$138
$127
$119
Pack
$6,675
$7,351
$8,797
$7,688
$9,101
$9,732
$/kWh
$157
$141
$134
$139
$128
$120
Pack
$6,588
$7,246
$8,743
$7,555
$8,966
$9,579
$/kWh
$160
$143
$134
$141
$130
$122
Pack
$6,572
$7,224
$8,743
$7,555
$8,966
$9,515
$/kWh
$161
$144
$134
$141
$130
$123
Pack
$6,588
$7,224
$8,743
$7,555
$8,966
$9,515
$/kWh
$160
$144
$134
$141
$130
$123
Change from FRM (including change from EV150 to EV200)

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
-16.9%
-24.2%
-20.4%
-23.5%
-24.4%
-29.4%
$/kWh
-11.8%
-19.4%
-20.4%
-19.5%
-21.6%
-25.7%
Pack
-17.1%
-24.3%
-20.6%
-23.9%
-24.9%
-29.6%
$/kWh
-11.4%
-19.0%
-20.1%
-19.0%
-21.2%
-25.4%
Pack
-18.1%
-25.4%
-21.0%
-25.3%
-26.0%
-30.4%
$/kWh
-9.9%
-17.7%
-19.6%
-18.0%
-20.0%
-24.4%
Pack
-18.3%
-25.6%
-21.0%
-25.3%
-26.0%
-30.8%
$/kWh
-9.5%
-17.5%
-19.6%
-18.0%
-20.0%
-24.0%
Pack
-18.1%
-25.6%
-21.0%
-25.3%
-26.0%
-30.8%
$/kWh
-9.9%
-17.5%
-19.6%
-18.0%
-20.0%
-24.0%
Note:
       CWR = target percent reduction in vehicle curb weight.
       Actual reduction will be less if it would require applying more than 20 percent mass reduction to glider.
       *NMC622 cathode.
                                                 5-350

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
        Table 5.119 Estimated Direct Manufacturing Costs in MY2025 for PHEV20 Battery Packs
PHEV20*
(450k/yr)
0% CWR
2% CWR
7.5% CWR
10% CWR
20% CWR
FRM (2008 baseline)

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$2,531
$ 2,962
$ 3,734
$ 2,835
$ 3,424
$ 3,874
$/kWh
$364
$347
$368
$316
$300
$295
Pack
$ 2,517
$ 2,938
$ 3,696
$ 2,813
$ 3,393
$ 3,834
$/kWh
$364
$348
$369
$317
$301
$295
Pack
$ 2,469
$ 2,835
$ 3,592
$ 2,754
$ 3,309
$ 3,732
$/kWh
$370
$345
$369
$319
$302
$295
Pack
$ 2,447
$ 2,808
$ 3,546
$ 2,730
$ 3,274
$ 3,681
$/kWh
$371
$346
$368
$320
$303
$297
Pack
$ 2,431
$ 2,784
$ 3,510
$ 2,703
$ 3,244
$ 3,671
$/kWh
$373
$347
$369
$323
$303
$296
Draft TAR

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$2,463
$2,690
$3,157
$2,737
$3,025
$3,190
$/kWh
$382
$340
$316
$325
$279
$259
Pack
$2,454
$2,678
$3,136
$2,727
$3,008
$3,169
$/kWh
$385
$342
$318
$328
$281
$261
Pack
$2,433
$2,649
$3,080
$2,699
$2,962
$3,115
$/kWh
$394
$349
$321
$335
$285
$264
Pack
$2,424
$2,638
$3,070
$2,688
$2,942
$3,103
$/kWh
$397
$352
$322
$337
$287
$265
Pack
$2,420
$2,638
$3,070
$2,683
$2,937
$3,103
$/kWh
$399
$352
$322
$339
$288
$265
Change from FRM

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
-2.7%
-9.2%
-15.4%
-3.4%
-11.6%
-17.6%
$/kWh
4.9%
-1.9%
-14.2%
3.1%
-7.0%
-12.0%
Pack
-2.5%
-8.9%
-15.1%
-3.1%
-11.4%
-17.3%
$/kWh
5.8%
-1.6%
-13.8%
3.6%
-6.7%
-11.6%
Pack
-1.4%
-6.6%
-14.3%
-2.0%
-10.5%
-16.5%
$/kWh
6.5%
1.1%
-12.8%
4.9%
-5.6%
-10.6%
Pack
-0.9%
-6.0%
-13.4%
-1.5%
-10.1%
-15.7%
$/kWh
7.1%
1.7%
-12.4%
5.5%
-5.2%
-10.8%
Pack
-0.5%
-5.2%
-12.5%
-0.7%
-9.5%
-15.5%
$/kWh
7.0%
1.4%
-12.6%
5.1%
-5.0%
-10.5%
Note:
       CWR = target percent reduction in vehicle curb weight.
       Actual reduction will be less if it would require applying more than 20 percent mass reduction to glider.
       *Blended LMO-NMC cathode.
                                                 5-351

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
        Table 5.120 Estimated Direct Manufacturing Costs in MY2025 for PHEV40 Battery Packs
PHEV40*
(450k/yr)
0% CWR
2% CWR
7.5% CWR
10% CWR
20% CWR
FRM (2008 baseline)

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$ 3,644
$ 4,390
$ 6,006
$ 4,247
$ 5,269
$ 6,122
$/kWh
$262
$257
$296
$236
$231
$233
Pack
$ 3,619
$ 4,343
$5,921
$ 4,207
$ 5,212
$ 6,050
$/kWh
$262
$257
$295
$237
$231
$233
Pack
$ 3,542
$ 4,228
$5,671
$ 4,101
$ 5,065
$ 5,900
$/kWh
$264
$258
$291
$238
$231
$232
Pack
$ 3,542
$ 4,228
$ 5,671
$ 4,100
$ 5,065
$ 5,900
$/kWh
$264
$258
$291
$237
$231
$232
Pack
$ 3,542
$ 4,228
$ 5,671
$ 4,100
$ 5,065
$ 5,900
$/kWh
$264
$258
$291
$237
$231
$232
Draft TAR

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$3,130
$3,705
$5,528
$3,661
$4,620
$5,073
$/kWh
$260
$251
$295
$233
$229
$221
Pack
$3,111
$3,599
$5,550
$3,635
$4,622
$5,026
$/kWh
$262
$246
$296
$234
$231
$221
Pack
$3,077
$3,559
$5,552
$3,579
$4,574
$4,999
$/kWh
$264
$247
$296
$236
$232
$222
Pack
$3,078
$3,559
$5,550
$3,579
$4,574
$4,999
$/kWh
$264
$247
$296
$236
$232
$222
Pack
$3,077
$3,559
$5,552
$3,579
$4,574
$4,999
$/kWh
$264
$247
$296
$236
$232
$222
Change from FRM

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
-14.1%
-15.6%
-8.0%
-13.8%
-12.3%
-17.1%
$/kWh
-0.8%
-2.4%
-0.5%
-1.4%
-1.1%
-5.1%
Pack
-14.0%
-17.1%
-6.3%
-13.6%
-11.3%
-16.9%
$/kWh
-0.1%
-4.1%
0.2%
-1.1%
0.0%
-4.8%
Pack
-13.1%
-15.8%
-2.1%
-12.7%
-9.7%
-15.3%
$/kWh
-0.1%
-4.2%
1.7%
-0.5%
0.2%
-4.5%
Pack
-13.1%
-15.8%
-2.1%
-12.7%
-9.7%
-15.3%
$/kWh
-0.1%
-4.2%
1.7%
-0.5%
0.2%
-4.5%
Pack
-13.1%
-15.8%
-2.1%
-12.7%
-9.7%
-15.3%
$/kWh
-0.1%
-4.2%
1.7%
-0.5%
0.2%
-4.5%
Note:
       CWR = target percent reduction in vehicle curb weight.
       Actual reduction will be less if it would require applying more than 20 percent mass reduction to glider.
       *NMC622 cathode.
                                                 5-352

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
       Table 5.121 Estimated Direct Manufacturing Costs in MY2017 for strong HEV Battery Packs
STRONG HEV*
(450k/yr)
0% CWR
2% CWR
7.5% CWR
10% CWR
20% CWR
FRM (2008 baseline)

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$726
$801
$938
$779
$876
$ 1,010
$/kWh
$896
$804
$809
$747
$682
$676
Pack
$722
$796
$929
$775
$870
$ 1,003
$/kWh
$909
$815
$817
$758
$691
$685
Pack
$712
$783
$909
$762
$853
$983
$/kWh
$950
$849
$848
$790
$718
$711
Pack
$708
$777
$900
$757
$846
$974
$/kWh
$970
$866
$862
$806
$731
$724
Pack
$700
$765
$882
$746
$830
$957
$/kWh
$ 1,008
$901
$894
$839
$760
$747
Draft TAR

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
$984
$ 1,051
$ 1,197
$ 1,033
$ 1,123
$ 1,194
$/kWh
$ 1,216
$ 1,057
$976
$984
$855
$792
Pack
$980
$ 1,046
$ 1,188
$ 1,029
$ 1,117
$ 1,187
$/kWh
$ 1,236
$ 1,074
$988
$ 1,000
$868
$803
Pack
$971
$ 1,033
$ 1,168
$ 1,017
$ 1,100
$ 1,167
$/kWh
$ 1,297
$ 1,123
$ 1,029
$ 1,047
$907
$836
Pack
$966
$ 1,027
$ 1,158
$ 1,011
$ 1,093
$ 1,158
$/kWh
$1,326
$1,148
$1,050
$1,070
$925
$853
Pack
$958
$ 1,016
$ 1,140
$ 1,001
$ 1,078
$ 1,142
$/kWh
$ 1,383
$ 1,198
$ 1,093
$ 1,118
$966
$882
Change from FRM

Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Pack
35.6%
31.2%
27.7%
32.6%
28.2%
18.3%
$/kWh
35.8%
31.5%
20.7%
31.7%
25.5%
17.1%
Pack
35.8%
31.4%
27.9%
32.8%
28.4%
18.4%
$/kWh
36.0%
31.7%
20.9%
31.9%
25.7%
17.2%
Pack
36.3%
32.0%
28.4%
33.4%
29.0%
18.7%
$/kWh
36.5%
32.2%
21.5%
32.5%
26.3%
17.5%
Pack
36.6%
32.3%
28.7%
33.6%
29.2%
18.8%
$/kWh
36.7%
32.5%
21.7%
32.7%
26.5%
17.7%
Pack
37.0%
32.8%
29.2%
34.2%
29.8%
19.3%
$/kWh
37.2%
33.0%
22.2%
33.3%
27.1%
18.2%
Note:
       CWR = target percent reduction in vehicle curb weight.
       Actual reduction will be less if it would require applying more than 20 percent mass reduction to glider.
       *Blended LMO-NMC cathode.
                                                 5-353

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.4.3.7.6    Discussion of Battery Cost Projections

   In Section 5.2.4.4.9 (Evaluation of 2012 FRM Battery Cost Projections),  the agencies
reviewed the 2020-2022 cell-level costs projected by GM for its LG-supplied cells for the Chevy
Bolt EV, and converted them to estimated pack-level costs per gross kWh.  These estimated
costs were shown to appear generally lower than the pack-level costs for EV150 that were
generated by the 2012 FRM analysis. Figure 5.114 extends this comparison to the pack-level
costs for EV200 projected by this Draft TAR analysis. Although these Draft TAR projected costs
are significantly lower than the costs projected in the 2012 FRM analysis, they appear consistent
with and in many cases appear to remain conservative with respect to the trend established by the
GM/LG pack-converted cost estimates.
      400

      350
   1-250
    •
   — 200
   _c
      15:
    n.i
    a
       5C

-8-
-S-
                           8     i
 *  FRM

 O  Draft TAR

-ft—GM/LG low

-B—GM/LG high





       0  l
        2015  2016  2017  2018  2019  2020  2021  2022  2023  2024  2025  2026
                                      Year

 Figure 5.114 Comparison of Estimated Pack-Converted GM/LG Costs to 2012 FRM EV150 and Draft TAR
                                     EV200 Projections
   As discussed in Section 5.2.4.4.9, comparisons of the GM/LG costs to those of the 2012 FRM
and Draft TAR analyses are subject to some uncertainty. However, comparison on this basis to
the 2012 FRM projections suggests that those projections may have been conservative with
respect to trends in battery cost that have occurred since the FRM. This outcome suggests that
EPA's battery costing methodology, with the updates and refinements discussed previously, is an
appropriate basis on which to derive updated projections for this Draft GHG Assessment.  As
suggested throughout this analysis, it should be noted that battery costs have many drivers, and
future cost projections derived by any methodology are subject to significant uncertainties.
                                             5-354

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.4.3.7.7    Battery Pack Costs Used in OMEGA
 Table 5.122 Linear Regressions of Strong Hybrid Battery System Direct Manufacturing Costs vs Net Mass
                              Reduction Applicable in MY2017 (2013$)
Vehicle Class
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Strong HEV
-$176x+$984
-$235x+$l,051
-$379x+$l,196
-$217x+$l,033
-$299x+$l,123
-$365x+$l,194
               Note: "x" in the equations represents the net weight reduction as a percentage.
 Table 5.123 Linear Regressions of Battery Electric Battery System Direct Manufacturing Costs vs Net Mass
                               Reduction Applicable in MY2025 (2013$)
Vehicle Class
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
PHEV20
-$403x+$2,463
-$518x+$2,689
-$l,039x+$3,157
-$502x+$2,737
n/a
n/a
PHEV40
-$891x+$3,130
-$2,607x+$3,685
-$28,870x+$5,337
-$l,293x+$3,661
n/a
n/a
EV75
-$885x+$3,960
-$l,123x+$4,414
-$2,702x+$5,807
-$l,136x+$4,515
n/a
n/a
EV100
-$l,121x+$4,534
-$l,319x+$5,306
-$2,823x+$6,475
-$l,064x+$5,407
n/a
n/a
EV200
-$l,628x+$6,710
-$2,063x+$7,394
-$2,630x+$8,851
-$2,315x+$7,734
n/a
n/a
               Note: "x" in the equations represents the net weight reduction as a percentage.
                   Table 5.124 Costs for MHEV48V Battery (dollar values in 2013$)
Vehicle
Class



All
All
All
Cost
type



DMC
1C
TC
DMC: base
year cost
1C: complexity


$306
Highl

DMC:
learning
curve
1C: near term
thru
31
2024

2017




$306
$172
$478
2018




$277
$170
$447
2019




$258
$169
$427
2020




$244
$168
$413
2021




$234
$168
$401
2022




$225
$167
$392
2023




$218
$167
$384
2024




$212
$166
$378
2025




$206
$102
$309
               Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                                  5-355

-------
                         Technology Cost, Effectiveness, and Lead-Time Assessment
          Table 5.125 Costs for Strong Hybrid Batteries (dollar values in 2013$)
Vehicle
Class
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
WRtech
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
WRnet
5
10
15
5
10
15
5
10
15
5
10
15
6
11
16
6
11
16
5
10
15
5
10
15
5
10
15
5
10
15
6
11
16
6
11
16
5
10
15
5
10
15
5
10
15
5
10
15
6
11
16
6
11
16
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
$975
$966
$957
$1,039
$1,027
$1,015
$1,177
$1,159
$1,140
$1,022
$1,011
$1,000
$1,105
$1,090
$1,075
$1,172
$1,154
$1,136
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl
Highl


















DMC:
learning
curve
1C: near
term
thru
31
31
31
31
31
31
31
31
31
31
31
31
31
31
31
31
31
31
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024


















2017
$975
$966
$957
$1,039
$1,027
$1,015
$1,177
$1,159
$1,140
$1,022
$1,011
$1,000
$1,105
$1,090
$1,075
$1,172
$1,154
$1,136
$549
$545
$540
$586
$579
$572
$664
$653
$642
$576
$570
$564
$623
$614
$606
$661
$650
$640
$1,524
$1,511
$1,497
$1,625
$1,606
$1,588
$1,841
$1,811
$1,782
$1,598
$1,581
$1,564
$1,727
$1,704
$1,681
$1,833
$1,804
$1,776
2018
$883
$875
$867
$941
$930
$919
$1,066
$1,049
$1,032
$925
$916
$906
$1,000
$987
$973
$1,061
$1,045
$1,028
$544
$539
$534
$579
$573
$566
$656
$646
$635
$570
$564
$558
$616
$608
$599
$654
$643
$633
$1,426
$1,413
$1,401
$1,520
$1,503
$1,486
$1,723
$1,695
$1,667
$1,495
$1,479
$1,463
$1,616
$1,594
$1,572
$1,715
$1,688
$1,661
2019
$823
$815
$808
$877
$867
$857
$994
$978
$962
$862
$853
$844
$932
$920
$907
$989
$974
$958
$540
$535
$530
$575
$569
$562
$652
$641
$631
$566
$560
$554
$611
$603
$595
$649
$639
$629
$1,362
$1,350
$1,338
$1,452
$1,435
$1,419
$1,645
$1,619
$1,592
$1,428
$1,413
$1,398
$1,544
$1,523
$1,502
$1,638
$1,612
$1,587
2020
$779
$772
$765
$830
$821
$811
$941
$926
$910
$816
$808
$799
$883
$871
$859
$937
$922
$907
$537
$532
$527
$572
$566
$559
$648
$638
$627
$563
$557
$551
$608
$600
$592
$645
$635
$625
$1,316
$1,304
$1,292
$1,402
$1,386
$1,370
$1,589
$1,563
$1,538
$1,379
$1,365
$1,350
$1,491
$1,471
$1,450
$1,582
$1,557
$1,533
2021
$745
$738
$731
$794
$785
$776
$900
$885
$871
$781
$773
$764
$844
$833
$821
$896
$882
$868
$535
$530
$525
$570
$563
$557
$646
$635
$625
$560
$554
$548
$606
$598
$589
$643
$633
$623
$1,279
$1,268
$1,256
$1,363
$1,348
$1,333
$1,545
$1,520
$1,496
$1,341
$1,327
$1,313
$1,450
$1,430
$1,411
$1,538
$1,514
$1,490
2022
$717
$711
$704
$764
$756
$747
$866
$852
$838
$752
$744
$736
$813
$802
$791
$862
$849
$836
$533
$528
$523
$568
$561
$555
$643
$633
$623
$559
$553
$547
$604
$596
$587
$641
$631
$621
$1,250
$1,239
$1,228
$1,332
$1,317
$1,302
$1,510
$1,485
$1,461
$1,310
$1,296
$1,283
$1,416
$1,397
$1,378
$1,503
$1,480
$1,456
2023
$694
$688
$682
$740
$731
$723
$838
$825
$811
$728
$720
$712
$787
$776
$765
$835
$822
$809
$531
$527
$522
$566
$560
$553
$642
$631
$621
$557
$551
$545
$602
$594
$586
$639
$629
$619
$1,226
$1,214
$1,203
$1,306
$1,291
$1,276
$1,480
$1,456
$1,432
$1,285
$1,271
$1,257
$1,389
$1,370
$1,351
$1,474
$1,451
$1,428
2024
$674
$668
$662
$719
$711
$702
$815
$801
$788
$707
$699
$692
$764
$754
$744
$811
$798
$786
$530
$525
$520
$565
$558
$552
$640
$630
$620
$556
$550
$544
$601
$592
$584
$637
$627
$617
$1,204
$1,194
$1,183
$1,284
$1,269
$1,255
$1,455
$1,431
$1,408
$1,263
$1,249
$1,236
$1,365
$1,346
$1,328
$1,448
$1,426
$1,403
2025
$657
$651
$645
$700
$692
$685
$794
$781
$768
$689
$682
$674
$745
$735
$725
$790
$778
$766
$327
$324
$321
$348
$344
$340
$394
$388
$382
$342
$339
$335
$370
$365
$360
$393
$387
$381
$984
$975
$966
$1,048
$1,037
$1,025
$1,188
$1,169
$1,150
$1,031
$1,020
$1,009
$1,115
$1,100
$1,085
$1,183
$1,165
$1,146
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                         5-356

-------
                         Technology Cost, Effectiveness, and Lead-Time Assessment
      Table 5.126 Costs for 20 Mile Plug-in Hybrid Batteries (dollar values in 2013$)
Vehicle
Class
SmCar
SmCar
StCar
StCar
LgCar
LgCar
SmMPV
SmMPV
LgMPV
LgMPV
Truck
Truck
SmCar
SmCar
StCar
StCar
LgCar
LgCar
SmMPV
SmMPV
LgMPV
LgMPV
Truck
Truck
SmCar
SmCar
StCar
StCar
LgCar
LgCar
SmMPV
SmMPV
LgMPV
LgMPV
Truck
Truck
WRtech
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
15
20
WRnet
6
11
6
11
5
10
6
11
4
9
6
11
6
11
6
11
5
10
6
11
4
9
6
11
6
11
6
11
5
10
6
11
4
9
6
11
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
$2,439
$2,419
$2,658
$2,632
$3,105
$3,053
$2,707
$2,682
$2,991
$2,949
$3,131
$3,082
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2












DMC:
learning
curve
1C: near
term
thru
26
26
26
26
26
26
26
26
26
26
26
26
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024












2017
$3,933
$3,901
$4,287
$4,246
$5,008
$4,924
$4,366
$4,326
$4,825
$4,756
$5,050
$4,971
$1,988
$1,972
$2,167
$2,146
$2,531
$2,489
$2,207
$2,186
$2,438
$2,404
$2,552
$2,512
$5,921
$5,872
$6,454
$6,391
$7,539
$7,413
$6,573
$6,512
$7,263
$7,160
$7,602
$7,483
2018
$3,691
$3,661
$4,023
$3,984
$4,700
$4,621
$4,097
$4,059
$4,527
$4,463
$4,739
$4,665
$1,970
$1,954
$2,147
$2,126
$2,508
$2,466
$2,187
$2,166
$2,416
$2,382
$2,529
$2,490
$5,661
$5,614
$6,171
$6,110
$7,208
$7,088
$6,284
$6,226
$6,944
$6,845
$7,268
$7,154
2019
$3,490
$3,461
$3,804
$3,767
$4,444
$4,369
$3,874
$3,838
$4,281
$4,220
$4,480
$4,410
$1,955
$1,939
$2,131
$2,110
$2,490
$2,448
$2,170
$2,150
$2,398
$2,364
$2,510
$2,471
$5,445
$5,400
$5,935
$5,877
$6,933
$6,817
$6,044
$5,988
$6,679
$6,584
$6,991
$6,881
2020
$3,319
$3,292
$3,618
$3,582
$4,226
$4,155
$3,684
$3,650
$4,071
$4,013
$4,261
$4,194
$1,943
$1,927
$2,117
$2,097
$2,474
$2,432
$2,156
$2,136
$2,383
$2,349
$2,494
$2,455
$5,262
$5,218
$5,735
$5,679
$6,700
$6,588
$5,841
$5,786
$6,454
$6,362
$6,755
$6,649
2021
$3,172
$3,146
$3,457
$3,423
$4,038
$3,971
$3,521
$3,488
$3,890
$3,835
$4,072
$4,008
$1,932
$1,916
$2,106
$2,085
$2,460
$2,419
$2,144
$2,124
$2,370
$2,336
$2,480
$2,441
$5,104
$5,061
$5,563
$5,509
$6,498
$6,389
$5,665
$5,612
$6,260
$6,171
$6,552
$6,449
2022
$3,043
$3,018
$3,317
$3,284
$3,874
$3,809
$3,377
$3,346
$3,732
$3,679
$3,906
$3,845
$1,922
$1,906
$2,095
$2,075
$2,448
$2,407
$2,134
$2,114
$2,358
$2,324
$2,468
$2,429
$4,965
$4,924
$5,412
$5,359
$6,322
$6,216
$5,511
$5,460
$6,090
$6,004
$6,374
$6,274
2023
$2,929
$2,904
$3,192
$3,161
$3,729
$3,667
$3,251
$3,221
$3,592
$3,541
$3,760
$3,701
$1,914
$1,898
$2,086
$2,066
$2,437
$2,396
$2,124
$2,105
$2,348
$2,314
$2,457
$2,419
$4,843
$4,803
$5,278
$5,227
$6,166
$6,063
$5,375
$5,325
$5,940
$5,855
$6,217
$6,120
2024
$2,827
$2,803
$3,081
$3,051
$3,599
$3,539
$3,138
$3,109
$3,467
$3,418
$3,629
$3,572
$1,906
$1,891
$2,078
$2,058
$2,427
$2,387
$2,116
$2,096
$2,338
$2,305
$2,448
$2,409
$4,733
$4,694
$5,159
$5,109
$6,027
$5,926
$5,254
$5,205
$5,806
$5,723
$6,077
$5,981
2025
$2,735
$2,712
$2,981
$2,952
$3,482
$3,424
$3,036
$3,008
$3,355
$3,307
$3,511
$3,456
$1,226
$1,215
$1,336
$1,323
$1,560
$1,534
$1,360
$1,348
$1,503
$1,482
$1,573
$1,549
$3,961
$3,928
$4,317
$4,275
$5,043
$4,958
$4,396
$4,355
$4,858
$4,789
$5,085
$5,005
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                         5-357

-------
                          Technology Cost, Effectiveness, and Lead-Time Assessment
      Table 5.127 Costs for 40 Mile Plug-in Hybrid Batteries (dollar values in 2013$)
Vehicle
Class
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
WR
tech
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
WR
net
6
5
3
7
0
5
6
5
3
7
0
5
6
5
3
7
0
5
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
DMC: base
year cost
1C:
complexity
$3,076
$3,554
$4,471
$3,570
$4,629
$4,960
High2
High2
High2
High2
High2
High2






DMC:
learning
curve
1C: near
term
thru
26
26
26
26
26
26
2024
2024
2024
2024
2024
2024






2017
$4,962
$5,733
$7,211
$5,759
$7,467
$7,999
$2,508
$2,897
$3,645
$2,911
$3,774
$4,043
$7,469
$8,630
$10,856
$8,670
$11,240
$12,042
2018
$4,656
$5,380
$6,767
$5,404
$7,007
$7,507
$2,485
$2,871
$3,612
$2,884
$3,740
$4,007
$7,141
$8,251
$10,379
$8,289
$10,746
$11,513
2019
$4,402
$5,086
$6,398
$5,110
$6,625
$7,097
$2,466
$2,850
$3,585
$2,863
$3,712
$3,976
$6,868
$7,936
$9,983
$7,972
$10,336
$11,074
2020
$4,187
$4,838
$6,085
$4,860
$6,300
$6,750
$2,450
$2,831
$3,562
$2,844
$3,688
$3,951
$6,637
$7,669
$9,647
$7,704
$9,988
$10,701
2021
$4,001
$4,623
$5,815
$4,644
$6,021
$6,450
$2,437
$2,816
$3,542
$2,828
$3,667
$3,929
$6,437
$7,438
$9,357
$7,472
$9,688
$10,379
2022
$3,838
$4,435
$5,578
$4,455
$5,776
$6,188
$2,425
$2,802
$3,524
$2,815
$3,649
$3,909
$6,263
$7,237
$9,103
$7,269
$9,425
$10,097
2023
$3,694
$4,268
$5,369
$4,288
$5,559
$5,956
$2,414
$2,790
$3,509
$2,802
$3,633
$3,892
$6,108
$7,058
$8,878
$7,090
$9,192
$9,848
2024
$3,566
$4,120
$5,182
$4,139
$5,366
$5,749
$2,405
$2,779
$3,495
$2,791
$3,619
$3,877
$5,970
$6,898
$8,678
$6,930
$8,985
$9,626
2025
$3,450
$3,986
$5,014
$4,004
$5,192
$5,562
$1,546
$1,786
$2,247
$1,794
$2,326
$2,492
$4,996
$5,772
$7,261
$5,799
$7,518
$8,054
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
          Table 5.128 Costs for 75 Mile BEV Batteries (dollar values in 2013$)
Vehicle
Class
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
WR
tec
h
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
W
R
net
10
15
20
10
15
20
10
15
20
10
15
20
5
10
15
10
15
20
10
15
20
10
15
20
10
15
20
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
DMC:
base
year
cost
1C:
complex
ity
$3,872
$3,827
$3,783
$4,301
$4,245
$4,189
$5,536
$5,401
$5,266
$4,401
$4,344
$4,288
$5,312
$5,243
$5,174
$5,638
$5,538
$5,437
High2
High2
High2
High2
High2
High2
High2
High2
High2
DMC:
learnin
g curve
1C: near
term
thru
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
2024
2024
2024
2024
2024
2024
2024
2024
2024
2017
$6,245
$6,173
$6,102
$6,938
$6,847
$6,757
$8,930
$8,712
$8,494
$7,099
$7,007
$6,916
$8,568
$8,457
$8,346
$9,094
$8,932
$8,770
$3,156
$3,120
$3,084
$3,506
$3,461
$3,415
$4,513
$4,403
$4,293
2018
$5,860
$5,793
$5,726
$6,511
$6,426
$6,341
$8,380
$8,175
$7,971
$6,661
$6,576
$6,490
$8,040
$7,936
$7,831
$8,534
$8,382
$8,230
$3,128
$3,092
$3,056
$3,475
$3,430
$3,384
$4,473
$4,363
$4,254
2019
$5,541
$5,477
$5,414
$6,156
$6,075
$5,995
$7,923
$7,729
$7,536
$6,298
$6,217
$6,136
$7,602
$7,503
$7,404
$8,069
$7,925
$7,781
$3,104
$3,069
$3,033
$3,449
$3,404
$3,359
$4,439
$4,331
$4,222
2020
$5,270
$5,209
$5,149
$5,854
$5,778
$5,702
$7,535
$7,351
$7,167
$5,990
$5,913
$5,835
$7,230
$7,136
$7,042
$7,674
$7,537
$7,400
$3,084
$3,049
$3,014
$3,427
$3,382
$3,337
$4,410
$4,303
$4,195
2021
$5,036
$4,978
$4,920
$5,594
$5,521
$5,448
$7,200
$7,025
$6,849
$5,724
$5,650
$5,576
$6,908
$6,819
$6,729
$7,333
$7,202
$7,072
$3,067
$3,032
$2,997
$3,407
$3,363
$3,318
$4,386
$4,279
$4,172
2022
$4,831
$4,776
$4,720
$5,367
$5,297
$5,227
$6,908
$6,739
$6,571
$5,491
$5,420
$5,350
$6,628
$6,542
$6,456
$7,035
$6,910
$6,784
$3,052
$3,017
$2,982
$3,391
$3,346
$3,302
$4,364
$4,258
$4,151
2023
$4,650
$4,596
$4,543
$5,166
$5,098
$5,031
$6,649
$6,486
$6,324
$5,285
$5,217
$5,149
$6,379
$6,296
$6,214
$6,771
$6,651
$6,530
$3,039
$3,004
$2,969
$3,376
$3,332
$3,288
$4,345
$4,239
$4,133
2024
$4,488
$4,437
$4,385
$4,986
$4,921
$4,856
$6,417
$6,261
$6,104
$5,101
$5,036
$4,970
$6,157
$6,077
$5,997
$6,536
$6,419
$6,303
$3,027
$2,992
$2,958
$3,363
$3,319
$3,275
$4,328
$4,222
$4,117
2025
$4,342
$4,292
$4,243
$4,824
$4,761
$4,698
$6,209
$6,057
$5,906
$4,936
$4,872
$4,808
$5,957
$5,880
$5,803
$6,323
$6,211
$6,098
$1,946
$1,923
$1,901
$2,162
$2,133
$2,105
$2,782
$2,714
$2,646
                                         5-358

-------
                         Technology Cost, Effectiveness, and Lead-Time Assessment
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
5
10
15
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
5
10
15
10
15
20
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
High2
High2
High2
High2
High2
High2
High2
High2
High2


















2024
2024
2024
2024
2024
2024
2024
2024
2024


















$3,588
$3,541
$3,495
$4,330
$4,274
$4,218
$4,596
$4,514
$4,432
$9,401
$9,293
$9,186
$10,444
$10,308
$10,172
$13,443
$13,115
$12,787
$10,686
$10,549
$10,411
$12,898
$12,731
$12,563
$13,691
$13,447
$13,202
$3,555
$3,510
$3,464
$4,291
$4,236
$4,180
$4,555
$4,474
$4,393
$8,988
$8,885
$8,782
$9,986
$9,855
$9,725
$12,852
$12,539
$12,225
$10,217
$10,085
$9,953
$12,331
$12,171
$12,011
$13,089
$12,856
$12,622
$3,529
$3,483
$3,438
$4,259
$4,204
$4,148
$4,521
$4,440
$4,360
$8,645
$8,546
$8,447
$9,604
$9,479
$9,354
$12,362
$12,060
$11,758
$9,827
$9,700
$9,573
$11,860
$11,707
$11,553
$12,590
$12,365
$12,141
$3,506
$3,461
$3,416
$4,231
$4,177
$4,122
$4,492
$4,412
$4,331
$8,354
$8,258
$8,163
$9,281
$9,160
$9,039
$11,945
$11,654
$11,362
$9,496
$9,374
$9,251
$11,461
$11,312
$11,164
$12,166
$11,949
$11,732
$3,486
$3,441
$3,396
$4,208
$4,153
$4,099
$4,467
$4,387
$4,307
$8,103
$8,010
$7,917
$9,002
$8,884
$8,767
$11,586
$11,303
$11,021
$9,210
$9,092
$8,973
$11,116
$10,972
$10,828
$11,800
$11,589
$11,379
$3,469
$3,425
$3,380
$4,187
$4,133
$4,079
$4,445
$4,365
$4,286
$7,883
$7,793
$7,703
$8,758
$8,643
$8,529
$11,272
$10,997
$10,722
$8,961
$8,845
$8,729
$10,815
$10,675
$10,534
$11,480
$11,275
$11,070
$3,454
$3,410
$3,365
$4,169
$4,115
$4,061
$4,425
$4,346
$4,267
$7,688
$7,600
$7,512
$8,542
$8,430
$8,319
$10,994
$10,725
$10,457
$8,740
$8,627
$8,514
$10,548
$10,411
$10,275
$11,196
$10,997
$10,797
$3,441
$3,396
$3,352
$4,153
$4,099
$4,045
$4,408
$4,329
$4,251
$7,515
$7,429
$7,343
$8,349
$8,240
$8,131
$10,745
$10,483
$10,221
$8,542
$8,432
$8,322
$10,310
$10,176
$10,042
$10,943
$10,748
$10,553
$2,212
$2,183
$2,155
$2,669
$2,635
$2,600
$2,834
$2,783
$2,733
$6,288
$6,216
$6,144
$6,986
$6,895
$6,803
$8,991
$8,772
$8,552
$7,148
$7,055
$6,963
$8,627
$8,515
$8,403
$9,157
$8,994
$8,830
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
          Table 5.129 Costs for 100 Mile BEV Batteries (dollar values in 2013$)
Vehicle
Class
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
WR
tec
h
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
W
R
net
8
13
18
7
12
17
8
13
18
7
12
17
3
8
13
7
12
17
8
13
18
7
12
17
8
13
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
DMC:
base
year cost
1C:
complexi
ty
$4,445
$4,389
$4,332
$5,214
$5,148
$5,082
$6,249
$6,108
$5,967
$5,332
$5,279
$5,226
$6,214
$6,131
$6,047
$6,540
$6,443
$6,346
High2
High2
High2
High2
High2
High2
High2
High2
DMC:
learnin
g curve
1C: near
term
thru
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
26
2024
2024
2024
2024
2024
2024
2024
2024
2017
$7,169
$7,078
$6,988
$8,410
$8,303
$8,197
$10,080
$9,852
$9,625
$8,601
$8,515
$8,429
$10,024
$9,888
$9,753
$10,548
$10,392
$10,235
$3,623
$3,577
$3,532
$4,250
$4,196
$4,143
$5,094
$4,979
2018
$6,727
$6,642
$6,557
$7,892
$7,792
$7,692
$9,459
$9,245
$9,032
$8,071
$7,990
$7,910
$9,406
$9,279
$9,152
$9,899
$9,752
$9,605
$3,591
$3,545
$3,500
$4,212
$4,159
$4,105
$5,049
$4,935
2019
$6,360
$6,280
$6,200
$7,461
$7,367
$7,273
$8,943
$8,741
$8,539
$7,631
$7,555
$7,478
$8,893
$8,773
$8,653
$9,359
$9,220
$9,081
$3,564
$3,519
$3,474
$4,180
$4,127
$4,075
$5,011
$4,897
2020
$6,049
$5,973
$5,897
$7,096
$7,006
$6,917
$8,506
$8,314
$8,121
$7,257
$7,185
$7,112
$8,458
$8,344
$8,230
$8,901
$8,769
$8,637
$3,541
$3,496
$3,451
$4,153
$4,101
$4,048
$4,978
$4,866
2021
$5,781
$5,708
$5,635
$6,781
$6,695
$6,610
$8,128
$7,944
$7,761
$6,935
$6,866
$6,797
$8,082
$7,973
$7,864
$8,506
$8,379
$8,253
$3,521
$3,476
$3,432
$4,130
$4,078
$4,026
$4,951
$4,839
2022
$5,546
$5,476
$5,406
$6,505
$6,423
$6,341
$7,798
$7,621
$7,445
$6,653
$6,587
$6,520
$7,754
$7,649
$7,544
$8,160
$8,039
$7,918
$3,504
$3,459
$3,415
$4,110
$4,058
$4,006
$4,926
$4,815
2023
$5,338
$5,270
$5,203
$6,261
$6,182
$6,103
$7,505
$7,336
$7,166
$6,404
$6,340
$6,276
$7,463
$7,362
$7,261
$7,854
$7,737
$7,621
$3,488
$3,444
$3,400
$4,092
$4,040
$3,988
$4,905
$4,794
2024
$5,152
$5,087
$5,022
$6,044
$5,967
$5,891
$7,244
$7,080
$6,917
$6,181
$6,119
$6,057
$7,203
$7,106
$7,009
$7,580
$7,468
$7,356
$3,475
$3,431
$3,387
$4,076
$4,024
$3,973
$4,886
$4,775
2025
$4,985
$4,922
$4,859
$5,847
$5,773
$5,699
$7,009
$6,850
$6,692
$5,980
$5,920
$5,861
$6,969
$6,875
$6,781
$7,334
$7,226
$7,117
$2,234
$2,205
$2,177
$2,620
$2,587
$2,554
$3,141
$3,070
                                        5-359

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
18
7
12
17
3
8
13
7
12
17
8
13
18
7
12
17
8
13
18
7
12
17
3
8
13
7
12
17
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
High2
High2
High2
High2
High2
High2
High2
High2
High2
High2


















2024
2024
2024
2024
2024
2024
2024
2024
2024
2024


















$4,864
$4,347
$4,303
$4,260
$5,066
$4,997
$4,929
$5,331
$5,252
$5,173
$10,792
$10,656
$10,520
$12,660
$12,500
$12,339
$15,174
$14,832
$14,489
$12,947
$12,818
$12,689
$15,089
$14,886
$14,682
$15,879
$15,644
$15,408
$4,821
$4,308
$4,265
$4,222
$5,020
$4,953
$4,885
$5,283
$5,205
$5,126
$10,318
$10,188
$10,057
$12,104
$11,951
$11,797
$14,508
$14,180
$13,852
$12,378
$12,255
$12,131
$14,426
$14,232
$14,037
$15,182
$14,957
$14,731
$4,784
$4,275
$4,233
$4,190
$4,983
$4,915
$4,848
$5,243
$5,166
$5,088
$9,924
$9,799
$9,674
$11,642
$11,494
$11,347
$13,954
$13,639
$13,324
$11,906
$11,787
$11,668
$13,876
$13,688
$13,501
$14,602
$14,386
$14,169
$4,754
$4,248
$4,205
$4,163
$4,951
$4,884
$4,817
$5,210
$5,132
$5,055
$9,590
$9,469
$9,348
$11,250
$11,107
$10,965
$13,484
$13,180
$12,875
$11,505
$11,390
$11,275
$13,409
$13,227
$13,046
$14,111
$13,901
$13,692
$4,727
$4,224
$4,182
$4,140
$4,923
$4,856
$4,790
$5,181
$5,104
$5,027
$9,301
$9,184
$9,067
$10,911
$10,773
$10,635
$13,079
$12,783
$12,488
$11,159
$11,048
$10,936
$13,005
$12,830
$12,654
$13,686
$13,483
$13,280
$4,704
$4,203
$4,161
$4,119
$4,899
$4,833
$4,766
$5,155
$5,079
$5,002
$9,049
$8,935
$8,821
$10,615
$10,481
$10,347
$12,724
$12,436
$12,149
$10,856
$10,748
$10,640
$12,653
$12,482
$12,311
$13,315
$13,117
$12,920
$4,683
$4,185
$4,143
$4,101
$4,877
$4,811
$4,746
$5,133
$5,056
$4,980
$8,826
$8,714
$8,603
$10,353
$10,222
$10,091
$12,410
$12,130
$11,849
$10,588
$10,483
$10,377
$12,340
$12,174
$12,007
$12,986
$12,794
$12,601
$4,665
$4,168
$4,127
$4,085
$4,858
$4,793
$4,727
$5,112
$5,037
$4,961
$8,626
$8,518
$8,409
$10,119
$9,991
$9,863
$12,129
$11,855
$11,581
$10,349
$10,246
$10,143
$12,061
$11,899
$11,736
$12,693
$12,505
$12,316
$2,999
$2,680
$2,653
$2,626
$3,123
$3,081
$3,039
$3,287
$3,238
$3,189
$7,218
$7,127
$7,036
$8,468
$8,360
$8,253
$10,149
$9,920
$9,691
$8,660
$8,573
$8,487
$10,092
$9,956
$9,820
$10,621
$10,463
$10,306
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                 Table 5.130 Costs for 200 Mile BEV Batteries (dollar values in 2013$)
Vehicle
Class
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
WR
tec
h
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
20
W
R
net
8
8
10
8
4
8
8
8
10
8
4
8
8
8
10
8
4
8
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
DMC:
base
year cost
1C:
complexi
ty
$6,580
$7,229
$8,588
$7,549
$9,057
$9,564
High2
High2
High2
High2
High2
High2






DMC:
learnin
g curve
1C: near
term
thru
26
26
26
26
26
26
2024
2024
2024
2024
2024
2024






2017
$10,613
$11,660
$13,852
$12,176
$14,608
$15,426
$5,364
$5,893
$7,000
$6,153
$7,383
$7,796
$15,977
$17,553
$20,852
$18,329
$21,991
$23,222
2018
$9,959
$10,942
$12,998
$11,426
$13,709
$14,476
$5,316
$5,840
$6,938
$6,098
$7,317
$7,726
$15,275
$16,781
$19,936
$17,524
$21,025
$22,202
2019
$9,416
$10,345
$12,290
$10,803
$12,961
$13,686
$5,276
$5,796
$6,885
$6,052
$7,262
$7,668
$14,692
$16,141
$19,175
$16,855
$20,223
$21,354
2020
$8,955
$9,839
$11,688
$10,274
$12,327
$13,017
$5,242
$5,759
$6,841
$6,013
$7,215
$7,619
$14,197
$15,597
$18,529
$16,287
$19,542
$20,635
2021
$8,558
$9,402
$11,169
$9,818
$11,779
$12,438
$5,212
$5,726
$6,803
$5,980
$7,175
$7,576
$13,770
$15,128
$17,972
$15,797
$18,954
$20,015
2022
$8,210
$9,020
$10,715
$9,419
$11,301
$11,933
$5,187
$5,698
$6,770
$5,950
$7,139
$7,539
$13,397
$14,718
$17,485
$15,369
$18,440
$19,472
2023
$7,902
$8,681
$10,313
$9,065
$10,877
$11,485
$5,164
$5,673
$6,740
$5,924
$7,108
$7,506
$13,066
$14,355
$17,053
$14,990
$17,985
$18,991
2024
$7,627
$8,379
$9,954
$8,750
$10,498
$11,086
$5,144
$5,651
$6,714
$5,901
$7,080
$7,476
$12,771
$14,030
$16,668
$14,651
$17,579
$18,562
2025
$7,379
$8,107
$9,631
$8,466
$10,157
$10,726
$3,307
$3,633
$4,316
$3,794
$4,552
$4,806
$10,686
$11,740
$13,947
$12,259
$14,709
$15,532
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                                 5-360

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
5.3.4.3.7.8    Electrified Vehicle Costs Used In OMEGA (Battery + Non-battery Items)

   Costs presented in the tables that follow sum the battery, non-battery and, where applicable,
the in-home charger related costs for mild, strong and plug-in hybrids and full battery electric
vehicles.
                     Table 5.131  Full System Costs for 48V Mild Hybrids (2013$)
Vehicle Class
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
WRtech
5
5
5
5
5
5
WRnet
1.5
2
2.5
2.5
2.5
3
Cost type
TC
TC
TC
TC
TC
TC
2017
$1,045
$1,045
$1,045
$1,045
$1,045
$1,045
2018
$1,007
$1,007
$1,007
$1,007
$1,007
$1,007
2019
$939
$939
$939
$939
$939
$939
2020
$919
$919
$919
$919
$919
$919
2021
$902
$902
$902
$902
$902
$902
2022
$888
$888
$888
$888
$888
$888
2023
$876
$876
$876
$876
$876
$876
2024
$865
$865
$865
$865
$865
$865
2025
$792
$792
$792
$792
$792
$792
              Note: TC=total costs.
                      Table 5.132  Full System Costs for Strong Hybrids (2013$)
Vehicle Class
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
WRtech
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
WRnet
5
10
15
5
10
15
5
10
15
5
10
15
6
11
16
6
11
16
Cost type
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
2017
$4,088
$4,054
$4,020
$4,459
$4,410
$4,362
$5,142
$5,059
$4,975
$4,314
$4,271
$4,229
$4,846
$4,784
$4,721
$5,119
$5,040
$4,962
2018
$3,963
$3,930
$3,897
$4,325
$4,278
$4,231
$4,990
$4,909
$4,827
$4,183
$4,142
$4,101
$4,703
$4,642
$4,582
$4,967
$4,891
$4,815
2019
$3,499
$3,470
$3,441
$3,815
$3,773
$3,732
$4,397
$4,326
$4,255
$3,692
$3,656
$3,619
$4,144
$4,090
$4,037
$4,377
$4,310
$4,243
2020
$3,431
$3,402
$3,373
$3,741
$3,700
$3,659
$4,313
$4,243
$4,173
$3,620
$3,584
$3,549
$4,064
$4,012
$3,959
$4,293
$4,227
$4,161
2021
$3,374
$3,346
$3,318
$3,679
$3,639
$3,599
$4,242
$4,174
$4,105
$3,560
$3,525
$3,490
$3,998
$3,947
$3,895
$4,223
$4,158
$4,094
2022
$3,325
$3,297
$3,270
$3,627
$3,587
$3,548
$4,182
$4,114
$4,046
$3,509
$3,474
$3,440
$3,941
$3,891
$3,840
$4,163
$4,099
$4,035
2023
$3,283
$3,255
$3,228
$3,580
$3,541
$3,502
$4,129
$4,062
$3,995
$3,464
$3,430
$3,396
$3,891
$3,841
$3,791
$4,110
$4,047
$3,984
2024
$3,244
$3,217
$3,190
$3,539
$3,501
$3,462
$4,082
$4,015
$3,949
$3,424
$3,390
$3,356
$3,847
$3,797
$3,748
$4,063
$4,001
$3,938
2025
$3,008
$2,983
$2,958
$3,286
$3,250
$3,215
$3,794
$3,733
$3,671
$3,176
$3,144
$3,113
$3,577
$3,531
$3,485
$3,777
$3,719
$3,661
       Note: TC=total costs.
  Table 5.133  Full System Costs for 20 Mile Plug-in Hybrids, Including Charger & Charger Labor (2013$)
Vehicle
Class
SmCar
SmCar
StCar
StCar
LgCar
LgCar
SmMPV
SmMPV
LgMPV
LgMPV
Truck
Truck
WRtech
15
20
15
20
15
20
15
20
15
20
15
20
WRnet
6
11
6
11
5
10
6
11
4
9
6
11
Cost
type
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
2017
$10,136
$10,055
$11,115
$11,003
$13,287
$13,059
$11,095
$10,990
$12,502
$12,329
$13,054
$12,849
2018
$9,840
$9,763
$10,792
$10,683
$12,905
$12,683
$10,768
$10,666
$12,137
$11,970
$12,672
$12,473
2019
$9,143
$9,072
$10,006
$9,907
$11,909
$11,708
$9,999
$9,906
$11,231
$11,078
$11,720
$11,539
2020
$8,931
$8,862
$9,773
$9,677
$11,634
$11,438
$9,764
$9,673
$10,968
$10,819
$11,445
$11,269
2021
$8,746
$8,678
$9,570
$9,476
$11,393
$11,201
$9,559
$9,471
$10,739
$10,594
$11,206
$11,033
2022
$8,582
$8,516
$9,391
$9,299
$11,180
$10,992
$9,378
$9,291
$10,537
$10,394
$10,994
$10,824
2023
$8,437
$8,372
$9,231
$9,140
$10,990
$10,805
$9,217
$9,132
$10,356
$10,216
$10,804
$10,638
2024
$8,305
$8,241
$9,087
$8,998
$10,819
$10,637
$9,071
$8,988
$10,193
$10,055
$10,634
$10,470
2025
$7,505
$7,448
$8,214
$8,133
$9,798
$9,633
$8,184
$8,108
$9,211
$9,087
$9,606
$9,459
       Note: TC=total costs.
                                                5-361

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.134 Full System Costs for 40 Mile Plug-in Hybrids, Including Charger & Charger Labor (2013$)
Vehicle
Class
SmCar
StCar
LgCar
SmMPV
LgMPV
Truck
WRtech
20
20
20
20
20
20
WRnet
6
5
3
7
0
5
Cost
type
TC
TC
TC
TC
TC
TC
2017
$12,644
$14,689
$19,170
$14,482
$18,425
$19,641
2018
$12,259
$14,242
$18,600
$14,034
$17,850
$19,026
2019
$11,391
$13,200
$17,130
$13,035
$16,543
$17,625
2020
$11,115
$12,879
$16,720
$12,713
$16,130
$17,184
2021
$10,875
$12,600
$16,362
$12,434
$15,771
$16,800
2022
$10,662
$12,352
$16,045
$12,187
$15,453
$16,460
2023
$10,473
$12,132
$15,762
$11,967
$15,170
$16,158
2024
$10,303
$11,933
$15,506
$11,768
$14,915
$15,885
2025
$9,260
$10,727
$13,991
$10,554
$13,357
$14,220
     Note: TC=total costs.
    Table 5.135 Full System Costs for 75 Mile BEVs, Including Charger & Charger Labor (2013$)
Vehicle
Class
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
WRtech
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
WRnet
10
15
20
10
15
20
10
15
20
10
15
20
5
10
15
10
15
20
Cost
type
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
2017
$10,905
$10,787
$10,669
$12,645
$12,374
$12,103
$16,820
$16,214
$15,607
$12,102
$11,950
$11,797
$15,076
$14,718
$14,361
$15,106
$14,834
$14,562
2018
$10,473
$10,361
$10,249
$12,155
$11,892
$11,629
$16,184
$15,595
$15,007
$11,622
$11,477
$11,332
$14,479
$14,131
$13,783
$14,495
$14,236
$13,977
2019
$10,115
$10,008
$9,901
$11,748
$11,491
$11,234
$15,653
$15,080
$14,507
$11,225
$11,086
$10,947
$13,982
$13,643
$13,303
$13,987
$13,740
$13,492
2020
$9,812
$9,709
$9,606
$11,401
$11,150
$10,899
$15,202
$14,642
$14,081
$10,888
$10,755
$10,621
$13,560
$13,228
$12,895
$13,558
$13,320
$13,082
2021
$9,550
$9,451
$9,351
$11,102
$10,856
$10,610
$14,812
$14,263
$13,714
$10,598
$10,469
$10,341
$13,196
$12,870
$12,543
$13,187
$12,958
$12,728
2022
$9,321
$9,225
$9,128
$10,840
$10,598
$10,357
$14,469
$13,930
$13,392
$10,345
$10,220
$10,095
$12,877
$12,556
$12,235
$12,864
$12,642
$12,420
2023
$9,119
$9,025
$8,932
$10,608
$10,370
$10,133
$14,166
$13,636
$13,106
$10,121
$10,000
$9,879
$12,594
$12,278
$11,963
$12,578
$12,362
$12,147
2024
$8,938
$8,847
$8,756
$10,401
$10,166
$9,932
$13,894
$13,372
$12,851
$9,921
$9,804
$9,686
$12,341
$12,030
$11,719
$12,323
$12,113
$11,904
2025
$7,595
$7,505
$7,416
$8,877
$8,682
$8,487
$11,788
$11,355
$10,921
$8,301
$8,183
$8,066
$10,501
$10,242
$9,984
$10,310
$10,098
$9,886
     Note: TC=total costs.
    Table 5.136  Full System Costs for 100 Mile BEVs, Including Charger & Charger Labor (2013$)
Vehicle
Class
SmCar
SmCar
SmCar
StCar
StCar
StCar
LgCar
LgCar
LgCar
SmMPV
SmMPV
SmMPV
LgMPV
LgMPV
LgMPV
Truck
Truck
Truck
WRtech
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
10
15
20
WRnet
8
13
18
7
12
17
8
13
18
7
12
17
3
8
13
7
12
17
Cost
type
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
2017
$12,300
$12,153
$12,007
$14,941
$14,646
$14,352
$18,663
$18,042
$17,421
$14,371
$14,228
$14,084
$17,344
$16,950
$16,555
$17,312
$17,049
$16,785
2018
$11,807
$11,667
$11,528
$14,353
$14,067
$13,781
$17,949
$17,346
$16,744
$13,792
$13,655
$13,518
$16,649
$16,266
$15,884
$16,603
$16,352
$16,101
2019
$11,398
$11,264
$11,131
$13,864
$13,585
$13,306
$17,354
$16,767
$16,181
$13,311
$13,180
$13,049
$16,072
$15,699
$15,325
$16,014
$15,774
$15,535
2020
$11,051
$10,922
$10,794
$13,448
$13,176
$12,904
$16,848
$16,275
$15,702
$12,903
$12,778
$12,652
$15,581
$15,216
$14,851
$15,515
$15,285
$15,055
2021
$10,752
$10,627
$10,503
$13,089
$12,822
$12,556
$16,410
$15,849
$15,287
$12,552
$12,431
$12,310
$15,158
$14,800
$14,442
$15,085
$14,863
$14,641
2022
$10,490
$10,370
$10,249
$12,774
$12,513
$12,251
$16,027
$15,476
$14,924
$12,246
$12,128
$12,011
$14,787
$14,435
$14,084
$14,709
$14,495
$14,280
2023
$10,259
$10,142
$10,025
$12,496
$12,239
$11,981
$15,686
$15,145
$14,603
$11,974
$11,861
$11,747
$14,458
$14,112
$13,767
$14,377
$14,169
$13,961
2024
$10,052
$9,938
$9,824
$12,247
$11,993
$11,740
$15,382
$14,848
$14,315
$11,733
$11,622
$11,511
$14,164
$13,824
$13,483
$14,081
$13,878
$13,676
2025
$8,532
$8,424
$8,315
$10,421
$10,210
$9,999
$13,032
$12,589
$12,145
$9,828
$9,716
$9,605
$12,025
$11,742
$11,459
$11,804
$11,597
$11,391
     Note: TC=total costs.
                                             5-362

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
     Table 5.137 Full System Costs for 200 Mile BEVs, Including Charger & Charger Labor (2013$)
Vehicle
Class
SmCar
StCar
L^Car
SmMPV
LgMPV
Truck
WRtech
20
20
20
20
20
20
WRnet
8
8
10
8
4
8
Cost
type
TC
TC
TC
TC
TC
TC
2017
$17,485
$19,808
$24,232
$19,750
$24,207
$24,649
2018
$16,764
$19,006
$23,270
$18,934
$23,210
$23,618
2019
$16,166
$18,338
$22,469
$18,257
$22,381
$22,762
2020
$15,658
$17,771
$21,788
$17,684
$21,677
$22,036
2021
$15,221
$17,282
$21,200
$17,189
$21,069
$21,410
2022
$14,838
$16,853
$20,684
$16,757
$20,537
$20,863
2023
$14,499
$16,473
$20,227
$16,374
$20,066
$20,379
2024
$14,197
$16,134
$19,819
$16,033
$19,645
$19,947
2025
$12,001
$13,674
$16,746
$13,422
$16,612
$16,706
       Note: TC=total costs.
5.3.4.4 Aerodynamics: Data and Assumptions for this Assessment

   In Section 5.2.5 (Aerodynamics: State of Technology), the agencies reviewed the assumptions
associated with two levels of aerodynamic drag reduction technology, Aero 1 and Aero 2.  These
represented applications of drag reduction technology resulting in a 10 percent and 20 percent
reduction in aerodynamic drag, respectively.

   That Section also reviewed the findings of several studies including: (a) the 2015 NAS
Report; (b) a joint aerodynamics test program between EPA, Transport Canada,  and other
organizations; a CARB study performed by Control-Tec; and an informal survey of aerodynamic
technologies at the 2015 North American International Auto Show (NAIAS).

   These studies were seen to generally support the assumptions for cost and effectiveness of
Aero 1 and Aero 2 as defined in the 2012 FRM.  The findings of the NAS report generally
supported the assumptions for Aero 1 and Aero 2 as being applicable to the 2020-2025 time
frame. The findings of the Joint Aerodynamics Assessment Program and the Control-Tec
analysis also were shown to lend support to the feasibility of the 10 percent and  20 percent
effectiveness levels assumed for Aero 1 and Aero 2.  The penetration of passive and active
aerodynamic technologies as surveyed at the 2015 NAIAS was also shown to demonstrate that
manufacturers are already implementing many passive and active aerodynamic technologies in
MY2015 vehicles, with significant opportunity remaining to further apply these  technologies in a
more optimized fashion as vehicles enter redesign cycles in the future.

   At this time, EPA is therefore continuing to use  the FRM cost and effectiveness assumptions
for passive and active aerodynamic technology as a basis for OMEGA runs for this  Draft TAR
analysis. In Section 5.2, some tradeoffs and interactions  among specific aerodynamic
technologies were identified that suggest there could be value in refining the specific
combinations of technologies that are assumed to make up the Aerol and Aero2 packages that
are applied to vehicles in OMEGA. However, because EPA has not changed the costs associated
with specific aerodynamic technologies from those used in the FRM, EPA has not chosen at this
time to make such adjustments to the aerodynamic  packages. EPA intends to continue analyzing
costs and package combinations prior to the draft determination.

   Costs associated with  aero treatments and technologies are equivalent to those used in the
2012 FRM except for updates to 2013 dollars and use of a new learning curve (curve 24). The
aero costs are shown below.
                                            5-363

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                  Table 5.138 Costs for Aero Technologies (dollar values in 2013$)
Tech
Passive aero
Passive aero
Passive aero
Active aero
Active aero
Active aero
Passive+ Active
Cost type
DMC
1C
TC
DMC
1C
TC
TC
DMC: base year cost
1C: complexity
$43
Low2

$128
Med2


DMC: learning curve
1C: near term thru
24
2018

24
2024


2017
$41
$10
$51
$123
$49
$172
$223
2018
$40
$10
$50
$120
$49
$170
$220
2019
$39
$8
$48
$118
$49
$167
$215
2020
$39
$8
$47
$116
$49
$165
$212
2021
$38
$8
$46
$114
$49
$163
$210
2022
$38
$8
$46
$113
$49
$162
$207
2023
$37
$8
$45
$111
$49
$160
$205
2024
$37
$8
$45
$110
$49
$159
$203
2025
$36
$8
$44
$108
$37
$145
$189
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.5 Tires: Data and Assumptions for this Assessment

   In Section 5.2.6 (Tires: State of Technology), the agencies reviewed the assumptions
associated with two levels of low rolling resistance tire technology, LRRT1 and LRRT2. These
represented applications of rolling resistance reduction technology corresponding to a 10 percent
and 20 percent reduction in rolling resistance, respectively.

   That Section reviewed the findings of the 2015 NAS Report, which examined the agencies'
2012 FRM assumptions for feasibility, cost, and effectiveness for LRRT1 and LRRT2.  The
report concluded that the feasibility and effectiveness projected by the agencies for a 20 percent
reduction in rolling resistance in the 2020-2025 time frame appears to be reasonable. With
regard to costs, the Committee  substantially agreed with the costs projected by the agencies,
while noting that the problem of maintaining tread wear and traction requirements while
reducing rolling resistance continues to present engineering challenges that could affect tire
costs.

   The Section also reviewed EPA's activity in following industry developments and trends in
application of low rolling resistance technologies to light-duty vehicles,  and a project to track
trends in rolling resistance of OEM tires through the Control-Tec project. It also reviewed an
ongoing joint research program with Transport Canada and other agencies to study the rolling
resistance and traction characteristics of low-rolling resistance tires.

   At this time, these efforts have suggested that the 2012 FRM estimates of cost and
effectiveness for LRRT1 and LRRT2 remain reasonable for the time frame of the rule. EPA is
therefore continuing to use the FRM cost and effectiveness assumptions for LRRT1 and LRRT2
as a basis for OMEGA runs for this Draft TAR GHG Assessment.

   In the FRM and this Draft TAR GHG Assessment, LRRT1 remains defined as a 10 percent
reduction in rolling resistance from a base tire, and is estimated to result in a 1.9 percent
effectiveness improvement for all vehicle classes.

   Similarly, LRRT2 remains defined as a 20 percent reduction in rolling resistance from a base
tire,  and is estimated to result in a 3.9 percent effectiveness improvement.

   Costs associated with lower rolling resistance tires are equivalent to those used in the 2012
FRM except for updates to 2013 dollars and use of a new learning curve (curve 32 for LRRT2).
The LRRT costs are shown below.
                                             5-364

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
             Table 5.139 Costs for Lower Rolling Resistance Tires (dollar values in 2013$)
Tech
LRRT1
LRRT1
LRRT1
LRRT2
LRRT2
LRRT2
Cost type
DMC
1C
TC
DMC
1C
TC
DMC: base year cost
1C: complexity
$6
Low2

$43
Low2

DMC: learning curve
1C: near term thru
1
2018

32
2024

2017
$6
$1
$7
$56
$10
$66
2018
$6
$1
$7
$53
$10
$64
2019
$6
$1
$7
$51
$10
$62
2020
$6
$1
$7
$49
$10
$60
2021
$6
$1
$7
$48
$10
$58
2022
$6
$1
$7
$47
$10
$57
2023
$6
$1
$7
$45
$10
$56
2024
$6
$1
$7
$44
$10
$55
2025
$6
$1
$7
$43
$8
$52
              Note: DMC=direct manufacturing costs; IC=indirect costs;
              TC=total costs; both levels of lower rolling resistance are incremental to today's baseline tires.
5.3.4.6 Mass Reduction: Data and Assumptions for this Assessment

   This section describes the specific assumptions for mass reduction cost and effectiveness that
are used in this Draft TAR assessment of the GHG standards. These assumptions are based
largely on the information presented in Section 5.2, and the agencies' joint assessment of the
state of mass reduction technology which highlighted notable applications of mass reduction in
production vehicles since the FRM, and the significant amount of research and development into
lightweight materials and designs as shown in information in the Appendix.

   Section 5.3.4.6.1 describes the mass reduction costs and the cost curve development
methodology that are used in the analysis. Two separate  cost curves were developed from the
studies described in Section 5.2; one that is applied to cars and cross-over utility vehicles that
typically have  a unibody construction, and another that is applied to light duty trucks that
typically have  a body-on-frame construction.

   Section 5.3.4.6.2 details the methodology for determining how much mass reduction is
already present in the MY2014 baseline fleet. This information is then used to assign the
appropriate costs for additional mass reduction beyond what has been applied to each vehicle in
the baseline.

   Section 5.3.4.6.3 describes the assumptions used in the GHG analysis for the effectiveness of
mass reduction for reducing emissions.

   Section 5.3.4.6.4 contains sample tables of the direct manufacturing cost (DMC), indirect
costs (1C), and total costs (TC) for cars (unibody) and trucks (body  on frame) over 2017-2025
with learning applied given example baseline percent mass reduction. The analysis utilizes
baseline costs adjusted for every 0.5 percent mass reduction

   The treatment of mass reduction in the fleet safety analyses is explained in Chapter 8.

5.3.4.6.1      Cost Curves

   The Direct Manufacturing Cost (DMC) curve utilized in the 2012 FRM was a linear cost
curve starting at  $0 for no mass reduction with costs increasing at a constant rate of $4.36/lb for
each percent mass reduction (e.g. 10 percent mass reduction = $0.436/lb (or  $0.96/kg)), see
Figure 5.115.  The cost curve was applied to all vehicles uniformly, with the assumption that all
vehicles in the MY2008 baseline were starting from a level of zero percent mass reduction.
                                             5-365

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
$1.00
^$0.80
£
^$0.60
tt
<3 $0.40
.•s
D $0.20
$0.00
0
Mass Reduction Cost




^




X^



X



_x
x^

s

X


ope =
.X
X'


= 4.36
X^



x
^














% 5% 10% 15% 20% 25%
Percent of Mass Reduction
           Figure 5.115 2012 FRM Mass Reduction Direct Manufacturing Cost Curve ($/lb)

   These FRM costs were based primarily on a number of different component, subsystem and
vehicle -level studies.  While these studies represented the best information available on mass
reduction costs at the time, EPA continues to believe that tear-down and analysis of actual
vehicle components are the best way to derive technology cost estimates. For mass reduction in
particular, the most cost-effective management of vehicle mass will likely involve a holistic
approach to vehicle design which takes into consideration not only the primary opportunities for
mass reduction, but also all secondary mass reduction opportunities. CAE analyses to evaluate
crash, NVH and dynamic factors are also very important in determining material, grade, gauge
and geometry of BIW and other components. For this draft TAR, the FRM costs have been
updated in two important ways. First, the costs used in in this assessment were directly informed
by several of holistic full vehicle tear-down studies described in  Section 5.2. Second, this
assessment uses one cost curve for cars and cross-over utility vehicles (CUVs), and a different
cost curve for light-duty trucks, as appropriate for vehicles with fundamentally different design
and usage characteristics. Within EPA's application of technology packages, the Car/CUV curve
is applied to Vehicle Types 1-7 and 13 which are defined as the non-towing vehicles and
typically unibody construction. The light duty truck curve is applied to Vehicle Types 8-12  and
14-19 which are defined as towing vehicles,  and are typically body-on-frame vehicles. An
explanation of the Vehicle Types can be found in Chapter 12.

   The baseline model year in the FRM was MY2008 and the vast majority of vehicles at that
time were developed without the significant incentives for mass reduction, or the metals and
approaches that have been  created as a result of, the current GHG and CAFE standards, and
therefore form a reasonable basis from which to measure future mass reduction. The vehicle and
component designs typical of MY2008-2010 era vehicles are assumed to represent the "null"
technology for mass reduction, consistent with the definition of a null technology definition  for
powertrain, aero, tire, etc. as described in Section 5.3.1.1.

   The GHG analysis for the Draft TAR uses a MY2014 baseline fleet, so that when determining
the cost of mass reduction in this draft TAR, EPA recognizes it is important to account for any
mass reduction that has been applied beyond the "null" mass reduction level typical of MY2008
era vehicles. Since the emissions reducing benefits of mass reduction aren't realized unless the
overall curb weight decreases, mass reduction technology for this Draft TAR is defined to be
                                             5-366

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
equivalent to a reduction in curb weight, with some additional adjustments described in 5.3.4.6.2.
The mass reduction application to electrified vehicles accounts for the additional weight of
electrical components as described in 5.3.4.3.7.1.  This methodology for accounting for mass
reduction in electrified vehicles is the same as that used for the FRM.

   The car and CUV costs for this Draft TAR were developed as described in the following
section based on the tear-down studies of the MY2011 Honda Accord, and the MY2010 Toyota
Venza, conducted by NHTSA and EPA respectively.  These two vehicles represent designs
which have primarily steel structures, and component design and materials typical of MY2008-
2010 era cars and CUVs.

   The truck costs for this Draft TAR are also described in the following section and are based
on two light duty pickup truck studies.  Both studies focused on the Chevrolet Silverado 1500,
which is a steel intensive truck design.  One study utilized a MY2011 model which was
introduced for MY2007, and is representative of MY2008-2010 vintage light-duty trucks. The
second study utilized a MY2014 model which was redesigned and slightly lighter than the
MY2011 version. The methodology described here considered both LDT study results to
develop a MY2008 applicable cost curve.  Light-duty trucks, and pickup-trucks in particular,
have a number of unique characteristics which influence the potential solutions for achieving
mass reduction. The use of a body-on-frame design in which the bed and cab are separately
mounted to a frame that provides the main load bearing structure for towing, hauling, and crash
performance. Because of these unique load requirements, the opportunity to achieve secondary
mass reduction may be less than other passenger vehicles since the overall vehicle and subsystem
designs will still need to meet vehicle functional objectives under these unique load conditions.

   Overall, EPA believes these new Car/CUV and LDT cost curves are more representative than
the FRM's linear curve of direct manufacturing costs for applying mass. The holistic vehicle
studies provide a more comprehensive evaluation  of the opportunity for mass reduction as they
take into account all vehicle systems (e.g. body, interior, suspension, engine, drivetrain) as well
as the potential for secondary mass reduction  (e.g. decrease the size of powertrain and
suspension components as loads are reduced) In addition, vehicle functional  objectives are also
considered through CAE modeling and simulation (material grade and gauge, NVH
characteristics, vehicle acceleration, crashworthiness). In addition, the results in each study
show that while high levels of mass reduction may increase costs, there are also opportunities for
cost savings, especially at lower levels of mass reduction. These findings are consistent with
statements from industry, including suppliers  and  several OEMs.  An article  released by the
Center for Automotive Research (CAR) in February 2016 illustrates this point as it states  "The
figure below [Figure 5] illustrates a generic cost curve for lightweighting that is broadly
supported."571
                                             5-367

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
         Figure 7: General Auto Manufacturer Cost Curve to Lightweight Vehicles
                       COST
                      INCREASE
                        JO





                       COil
                      SAVINGS
                                                      Marginal
                                 Cost Savings
                                                               Expensive
                                           % Mass Reduction
         ©Center for Automotive Research
Page |  12
     Figure 5.116 CAR Figure for "General Auto Manufacturer Cost Curve to Lightweight Vehicles"

   It should be noted that while the costs used in this assessment are applied broadly across the
fleet, EPA recognizes that each particular vehicle will present specific opportunities for mass
reduction that are in some cases are more cost-effective, and in other cases less cost-effective
than were available on the vehicles selected for the tear-down studies. However, it is important
to note that the cost curves are intended to be representative of mass reduction applied to typical
MY2008-2010 vehicles, with subsequent adjustments for any additional mass reduction present
in MY2014 baseline. Also note that the cost curves represent component and system mass
reductions that are entirely applied towards a reduction in vehicle curb weight, as opposed to
offsetting mass increases from the addition of content and features.

5.3.4.6.1.1    Cost Curve for Cars and CUVs

    The cost curve for Cars/CUVs developed for this Draft TAR described below represents an
estimate of the Direct Manufacturing Cost (DMC) for mass reduction technologies that are
expected to be broadly available in 2020. Total Costs, which are made up of both DMCs and
Indirect Costs (ICs), are also presented in this section for completeness, although the details of
calculating ICs are provided separately. Additionally, learning is applied to mass reduction
consistent with the other technologies in this assessment to account for changes in costs over
time. More detail on the methods for calculating indirect costs and learning is provided in
Section 5.3.4.6.4.

   Car/CUV DMC Curve Generation

   The Car/CUV direct manufacturing cost curve is based on EPA's midsize CUV study based
on the Venza, and NHTSA's passenger car study based on the Accord.  This section describes the
development of the Car/CUV DMC  curve. Four related topics for the resultant passenger
car/CUV cost curve are also discussed. First is a discussion of the potential concerns for the cost
savings in the cost curve from a 2008 era vehicle.  Second, a cost curve adjustment methodology
is described such that vehicles with an acknowledged baseline mass reduction percentage will
                                              5-368

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
have higher costs for additional mass reduction.  Third, this section addresses additional
technology points (such as aluminum BIW) and whether extension of the current cost curve can
represent these points.  Lastly, the complete cost curve is shown with DMC, Indirect cost (1C)
and resultant Total Cost (TC).

   Development of the Car/CUV DMC curve for use in EPA modeling is completed in the
following steps outlined in Table 5.140.
                            Table 5.140 Car/CUV DMC Development
 STEP
TASK
         Begin with the cost curves for the Passenger Car and the Midsize CUV lightweighting studies (both of
                                    which are of the 2008 design era).
              Update the individual curves given OEM, peer review feedback and other considerations.
         Translate the Passenger Car DMC cost curve to use a similar methodology as the Midsize CUV curve.
            Average the new Passenger Car and Midsize CUV curves using the best fit line for each curve.
   STEP 1:  The 2012 NHTSA Passenger CarLLL and EPA Midsize CUV572 study cost curves are
shown in Figure 5.117.
           ff.	•	—rvn-              Jiii

                   Mass Reduction [% of curb vehicle weight)

        Figure -158: L\YY Mas* Savings versus Incremental Costs (wilh Powertraink Tun
                                                            Vehicle Level Cost Curve
                                                                            w/ Compounding


                                                                            w/o Compounding
                                                                            —Optimized Vehtcle Solution
                                                                            [-50.47/kg, IS.26%)
          S Vehide Mass-Reduction
    Figure 5.117 2012 NHTSA Passenger Car and EPA Midsize CUV Lightweighting Study Cost Curves

   STEP 2:  Both cost curves in Figure 5.117 were updated since 2012. The cost curve from the
2012 Midsize CUV study572 was adjusted based on peer review, OEM feedback and other
considerations specific to the report. The resultant cost curve for a MY2008-2010 era midsize
CUV is shown in Figure 5.118. The final $/kg and percent mass reduction results for the whole
vehicle direct manufacturing cost for the HSS BIW and aluminum closure point is $0.50/kg and
17.6 percent mass reduction for the Midsize CUV.
                                               5-369

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                $4 -
               -$12
                    Engineered Solution (HSS BIW)
                    ($0.50/kg @ 17.6% Mass Reduction)
                                                                      18%   20%
                                         y = 3968.3x3 - 1282.6x- + 160.78x - 9.9319
                                        % Vehicle Mass Reduction
         Figure 5.118 EPA Updated Midsize CUV Direct Manufacturing Cost Curve from Midsize CUV
                                           Study

   NHTSA's 2012 passenger car cost curve Engineered Solution point has been updated two
times.  Figure 5.119 shows the point LWV1.1 which is the updated engineered solution point
achieved through analysis of Honda's comments573.  The point LWV1.2 is the updated
engineered solution point and includes the updates from NHTSA's analysis of Honda's comments
as well as re-analyzed BIW for the IIHS small overlap crash. Design changes to the BIW result
in some additional mass and cost.  The analysis and NHTSA's updated  cost curves are presented
in their report published in February 2016.574
                                    10 Crtt       ISAM.       iomt
                                 Mass Reduction |% of Curb Vehicle Weight)
   Figure 5.119 NHTSA Updated Passenger Car Direct Manufacturing Cost Curve from Passenger Car
                                        StudyEEE'574

   STEP 3: For this analysis, the NHTSA passenger car cost curve was translated using a similar
methodology to the EPA midsize CUV cost curve so that the two curves represent the same basis
and can be averaged. Review of Figure 5.118 (EPA) and Figure 5.119 (NHTSA) reveals that
EEE LWV 1.2 contains the IIHS small overlap design solution and mass add of 6.9kg and $26.88. LWV 1.1 only
  addresses NHTSA's responses to Honda's comments.
                                              5-370

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
two different methodologies were utilized to create these curves. Figure 5.118 is a cost curve
represented of one vehicle solution which includes one set of mass reduction technology ideas
focused on using an AHSS BIW and a number of aluminum components.  Figure 5.119 is based
on several whole vehicle solutions including 1) AHSS BIW and closures, 2) AHSS BIW and
Aluminum closures and chassis frame, 3) Aluminum BIW, closures and chassis frame, 4)
Composite BIW and Mag/Al closures. The detailed work in the study is based on the AHSS
BIW and Aluminum Closures and Chassis frame point and the remaining are estimates.

   Details on the differences in the study approaches and methodologies used to generate the
original 2012 passenger car and CUV curves are presented in Table 5.141.
                                           5-371

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
    Table 5.141 Methodology Differences Between Original 2012 Car and CUV Mass Reduction Studies
                                  NHTSA
                           (Passenger Car Study)
                                                          EPA
                                            (Midsize CUV: Phase 2 Low Dev Study)
    COSTS and
   TIMEFRAME
     Costs are representative of 2017.
    Vehicle of MY2011 (similar to 2008)
                                              Costs are representative of 2020.
                                             Vehicle of MY2010 (similar to 2008)
  TECHNOLOGY
 IDENTIFICATION
 -Study examined a number of components,
    material choices and mfg techniques.
  -Several components using new materials
            were redesigned.
   -Powertrain mass reduction confined to
    downsizing from Civic to avoid engine
     efficiency technologies (powertrain
      component MR beyond scope).
                                         -Study examined each component within the
                                           vehicle for mass reduction possibilities.
                                          -A few non-established technologies were
                                         adopted with the expectation research would
                                         make them, or similar technologies, available
                                               in the 2020-2025 timeframe.
                                           -Reference some technologies, such as
                                         wheels, utilized in Phase 1 study on Midsize
                                                         CUV.575
      BIW
Material replacement, computer optimization
 of load paths and material grade and gauge.
                                           Material replacement, grade and gauge
                                                      optimization
   BIW SAFETY
   CRASH/NVH
      inCAE
-Include NVH, FMVSS and other crash tests, in
     design, grade and gauge decisions.
 -Mass add due to IIHS small overlap included
in updated study with BIW changes applied in
          vehicle solution point.
                                         -Include NVH, FMVSS and other crash tests in
                                                grade and gauge decisions.
                                              - No IIHS small overlap in design
                                         -Utilized NHTSA mass add and cost findings in
                                            baseline safety credits, not cost curve.
  ORDERING OF
   PRIMARY MR
 TECHNOLOGIES
 for COST CURVE
 -Some technologies and related MR grouped
             into two points.
    -Glider technologies used as primary
        (glider=vehicle-powertrain)
                                              -Ordered in lowest $/kg and then
                                           cumulatively added for $/kg over %MR.
                                             -Only grouped ideas as required for
                                                implementation feasibility.
 INCORPORATION
  OF SECONDARY
  MASS SAVINGS
      (SMS)
 -SMS for two intermediate points (for body,
chassis and powertrain MR) determined using
            factors to primary.
 -Full SMS applied to only individual vehicle
             solution points.
-SMS is inherent in the powertrain downsizing
 (Civic components adopted into the solution
 for several system components, ex: engine).
                                            -Study examined a number of major
                                         components that could be made smaller due
                                         to a lighter vehicle at the main solution point.
                                            -SMS was ratio'd at each level of mass
                                          reduction from 100% SMS at solution point
                                          back toward zero percent mass reduction.
                                           -SMS based in downsizing components.
   COST CURVE
   EXPRESSION
  -Curve is a connection of vehicle solution
   estimates for several material focused
solutions. Rigorous analysis (CAE analysis etc.)
 performed for AHSS BIW + Al intensive point
                 only.
 -DMC curve included two points for grouped
 glider technologies (non-structure and non-
   structure with aluminum closures) with
  system technologies and SMS included in
         whole vehicle solution.
                                           -The cost curve was created through the
                                          cumulative addition of best value primary
                                           mass reduction components, up through
                                            aluminum closures, and resulted in a
                                           continuous curve for the AHSS BIW and
                                               aluminum intensive solution.
                                           -Compounded curve includes primary +
                                            secondary percent mass red and $/kg.
   To create a similar cost curve as shown in Figure 5.118, several steps must be taken.  This is
achieved through an understanding of the methodology for Figure 5.118, referencing Figure
5.119 and using the descriptions and data in the 2012 NHTSA lightweighting report.

   a) Evaluate the data used to create the Engineered Solution, at the AHSS BIW and Aluminum
Closures and Chassis Frame point, in Figure 5.119 and separate into primary and secondary
technologies.  See Table 5.142. The mass save and costs were adjusted for the NHTSA response
                                                  5-372

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
to Honda's comments. (The IIHS mass add was not included in this curve for the mass add in
EPA's analysis is used in the MY2014 baseline calculations for safety credit and applied to all
passenger cars and the applicable CUV's/SUV's).

   b) Create the primary cost curve. The primary non-compounded cost curve, such as that
shown in the green/top curve of Figure 5.118, represents cumulatively added mass reduction and
costs for individual primary mass reduction technologies for the AHSS BIW and Aluminum
intensive vehicle solution.  These primary technologies can be adopted without concern of the
resultant mass of the vehicle.

   c) Determine the secondary mass and cost savings at the Engineered Solution point.
Secondary mass and cost reduction technologies are determined and applied for a number of
components, including the chassis and powertrain that could be redesigned to reflect the
reduction in load associated with a reduced vehicle curb weight.

   d) Create the compounded cost curve.  Ratio the secondary mass and cost savings from the
Engineered Solution point across the percent mass reduction from the primary cost curve. In
Figure 5.118 this is shown as the purple/bottom curve.  This  effectively shifts the cost curve
down and to the right. The translated curve in Figure 5.120 shows that NHTSA's approach for
only applying secondary mass at points greater than 10 percent was maintained.

   e) Create best fit curves to the data to be used in averaging.
    Table 5.142 Designation of Primary and Secondary Mass reduction for 2012 NHTSA Accord-based
                                   Passenger Car Study
System
Body
Hood
Radiator
Front Bumper
Rear Bumper
Deck lid
Fenders
Front Door Frame
Rear Door Frame
Front Suspension
Interior Systems
HVAC
MR Technology/List of System
Components
AHSS
Aluminum
Radiator
Hoses
Radiator Support
Fan system
Expansion Bottle
AHSS
AHSS
Stamped Al
Stamped Al
Stamped Al
Stamped Al
Lower Control Arm (AHSS)
Steering Knuckles (Al)
Stabilizer Bar (AHSS)
Engine Cradle (Al)
Other
Material Change and Downsize
Trim (Mucell)
Front Seat(mg base)
Rear Seat (composite back)
Instrument Panel (mag)
Mucell
Primary
*
*
*
*
*
*
*
*
*
*
*
*
Secondary
*

*






*

*
                                             5-373

-------
Technology Cost, Effectiveness, and Lead-Time Assessment

Electrical
Rear Suspension
Wheels & Tires
Engine
Transmission
Drive Shafts
Fuel System
Exhaust
Brake System
Steering
Battery
Fuel
Insulation - NVH
Downsize to Civic
Wiring and wiring harness (Al/copper)
Headlamps
Tail Lamps
K Frame
Suspension Arms
Bearing Hub
Stabilizer Bar
Other
Wheels (AHSS)
Tires
Spare Wheel and Tire
Car Jack
Downsize (Civic)
Downsize
Downsize from Civic
Downsize Fuel Tank
Exhaust on Body
Exhaust on Engine
Heat Shields
Downsize from Civic
Front Calipers
Rear Calipers
Pads (Front)
Pads (Rear)
Brake Discs (front)
Brake Discs (rear)
ABS system
Vacuum Pump
Emergency Brake
Steering shaft assembly
Steering rack
Power steering
Downsize to Civic
Downsize
Less fuel used with smaller tank
Add in 3.2kg at $10 - 3M Thinsulate,
Quietblend

*
*
*

*



*






*
*
*
*
*
*
*
*
*
*
*

             5-374

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                               Car DMC Curve, MY2008
                     Car Data in CUV Methodology, SMS at 10%+
                           y=1.4602ln(x) +4.4672
                               R'= 0.994
                                        i Mass Reduction
  Figure 5.120 Car DMC Curve from Car Data shown in CUV Methodology ($/kg vs %MR), Engineered
                  Solution AHSS BIW & Aluminum Closures and Chassis Frames

   One notable difference between Figure 5.118 and Figure 5.120 is the place of the cost save
mass reduction and the amount of cost increase mass reduction offset by these savings. The CUV
curve, Figure 5.118, has notable mass savings in the primary (green/top) curve due to the number
and cost estimates for the primary technologies. The Car curve, Figure 5.120, has notable mass
savings  in the secondary mass (as noted by the negative slope to the Engineered Solution point)
due to the notable number of downsized technologies. Regardless of the specific technologies
used to make the DMC curve, the curves reflect two different ways that mass reduction may be
implemented on 2008 era vehicles.

   STEP 4: To combine the two 2008 era DMC curves (Car and CUV) into a single Car/CUV
curve, the best fit equations for the cost curves in Figure 5.118 and Figure 5.120 were
determined and averaged together. The result is shown in Figure 5.121. This specific curve is
utilized  in the application of mass reduction for vehicles with a MY2014 baseline percent mass
reduction of zero. Vehicles with a MY2014 baseline percent mass reduction above zero will
have an adjusted cost curve applied.  Calculation of the MY2014 baseline percent mass reduction
is discussed in Section 5.3.4.6.2.  Adjustment of the curve for vehicles with MY2014 baseline
percent  mass reduction is discussed in Section 5.3.4.6.2 as well as in "Cost Curve Adjusted for
Baseline Mass Reduction Percent" further on in this discussion. The FRM linear cost curve is
also included in the figure and it is noted that the new cost curve lies below the FRM cost curve.

   Another factor in regards to costs for mass reduction is the improvement in fuel  efficiency.
EPA's analysis includes an increase in fuel efficiency of 5.2 percent for all vehicles in the 2020-
2025 timeframe.  Others have estimated the improvement as being 6 percent to 8 percent and the
recent presentation by IBIS Associates, Inc. at the 2016 DOE Annual Merit Review576 listed that
up to 7 percent improvement in fuel efficiency for every 10 percent reduction in vehicle mass. If
EPA's estimate is too low then other technologies will have to be adopted  in order to make up
this difference and in essence raising EPA's cost estimate for overall compliance with the
standards.
                                             5-375

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                                 Car/CUV DMC Curve
                                  MY2008 Baseline
                                               y = 1.9496ln(x) + 3.5376
                                                   R2 - 0.9869
                                                                       25%
                                       % Mass Reduction
                    cuv
                              Car
                                        Average
                                                    FRM
Log. (Average)
      Figure 5.121 DMC Curve for 2008 Era Car/CUV (2013$/kg v %MR) - HSS BIW, Al intensive
                             Car/CUV - MY2008 Baseline
                                   DMC ($/vehicle)
                                                                        . ::
                                       % Mass Reduction
 Figure 5.122 DMC Curve for 2008 Era Car/CUV (2013$/vehicle for a 3000 pound vehicle) -AHSS BIW, Al
                                        Intensive

   The DMC curve for Car/CUV, Figure 5.121, shows cost savings when starting from a 2008
era vehicle design. Other resources acknowledge mass reduction at cost savings including the
diagram by CAR in Figure 5.116. Several other documents also acknowledge some cost neutral
mass reduction. These include the presentation by IBIS Associates, Inc. at the 2016 DOE
                                            5-376

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
Annual Merit Review576 , and the 2015 NAS study on the costs for mass reduction for passenger
cars and light duty trucks.492>FFF Answers to several questions on this topic are listed below.

   1.  What types of technologies/approaches result in a mass reduction and a cost savings?

        Table 5.143 Technologies/Approaches that Result in a Mass Reduction and a DMC Savings
Approach
DMC Savings and Mass
Reduction
Secondary Mass
Technology
Design Part Integration and
Optimization
Material and Component Design
Optimization
Material Processing
Design and Processing
Identification and Modification of
components for Secondary Mass
Savings
Supporting Notes
Enhanced by improvements in CAE tools
(Ex: airbag housing)
Redesign a component for less mass of an
existing material
(Ex: scalloped edges in BIW)
Ex: Mucell
Ex: Hollow Tube
Use less of a material due to less load
stresses
(Ex: downsized engine, brakes)
   2. Have OEM's expressed the ability to achieve mass reduction at a cost savings?

   Information from an October 2015 GM investor presentation addressed the issue of cost
savings for adopting mass reduction in 2016 vehicle releases. The Malibu and Camaro are on
their second redesign since 2008 and the Cruze on its first since 2008. Using 2008 and 2016
curb weight values, from edmunds.com and A2Macl, mass reductions of 4.4 to 12.1 percent577
over 2008 MY are estimated, as shown in Table 5.144.
   VEHICLES WITH
   MORE EFFICIENCY
   AT BETTER MARGINS
      VOLT
c Nearly 2SO pounds lighter
o S3-mile battery range
-. Improved range by 40%
c Variable Profit Improvement
 -S3,500/unit
    MALIBU
-' Neatly 300 pounds lighter
o 48-mpg estimated
 for hybrid
o Improved FE by 8%
o Variable Profit Improvement
 -Sl.SOO/unit
     CRUZE
c Nearly 250 pounds tighter
o 40-mpg estimated
c Improved FE by 1296
o Variable Profit Improvement
 -$l,500/unit
    CAMARO
o Almost 400 pounds lighter
o 32-mpg estimated
o Improved mpq by 7%
o Variable Profit Improvement
 -Sl.OOO/unit
                                    REMOVING $2B in Material Cost to Fund the Future
     Figure 5.123 GM Investment Conference Call "Vehicles with More Efficiency at Better Margins"
1FF The 2015 NAS study includes a 'miri curve for the DMC for passenger car which reflects 6.5 percent mass
  reduction at a cost neutral, and includes 40 percent secondary mass for passenger cars at $1.00/kg cost save and
  25 percent secondary mass for light duty trucks at $1.00/kg cost save, at points of 10 percent primary mass
  reduction or more.
                                                  5-377

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
        Table 5.144 Estimate of Percent Change in Mass Reduction Compared to 2008 Estimates
Vehicle


Malibu
Cruze
Camaro
Curb
Weight
2008
3415-3649
3000
3780-3860
Curb
Weight
2014
3393-3660
3118
3719-3820
2014% MR
Change
(CWonly)
-0.6% to +0.3%
+4%
-1.6% to -1.0%
2016 Curb
Weight Dec
Est
300
250
400
Est CW 2016


3093-3388
2868
3319-3420
Est%MR
2016 over
2008
-9.4 to -7.9
-4.4
-12.1 to -11.4
       *source of information: edmunds.com/A2Macl
   Cost Curve Adjusted for Baseline Mass Reduction Percent

   Since the 2012 FRM, some manufacturers have reduced the curb weight of some of their
vehicles, modified the design to allow compliance with new FMVSS and IIHS safety
requirements and increased vehicle footprint.  The Draft TAR uses a MY2014 baseline for which
a baseline percent mass reduction per vehicle is calculated. The percent mass reduction is based
on a change in curb weight in MY2014 from MY2008 (along with an allowance for safety
compliance and vehicle footprint increase), and not the amount of mass reduction technology
applied. The reason for this is that the mass reduction technologies are not always evident by the
eye in the vehicle  and the benefits of mass reduction are not achieved unless the overall vehicle
is lighter. The detailed methodology for estimating the amount of mass reduction already present
in a MY2014 vehicle is presented in Section 5.3.4.6.2.

   The methodology for identifying and assigning baseline mass reduction is reflected in the
calculations for the higher future cost for mass reduction and a potential decrease in the total
additional mass reduction that can be applied to any given vehicle. Figure 5.124 and Figure
5.125 show the results of the DMC curve of a MY2014 vehicle with 5 percent mass reduction
applied since MY2008. The original maximum DMC save of $200 (for a vehicle with zero
percent MY2014 baseline mass reduction) is removed and the net  zero cost mass reduction at 16
percent is also eliminated. The overall mass reduction potential, given the AHSS BIW and
Aluminum intensive solution, has been reduced from 20 percent to 15 percent.
                       Passenger Car/CUV DMC Curve  Change
                         W/ 5% MR Baseline MY2014 ($/kg)
                   2.00
                                                                     20%
                 |
                 "
               % Mass Reduction

-5%MR Baseline 2014MY    —•—0%MR Baseline 2014MY
Figure 5.124 DMC Curve Adjusted for Car/CUV with 5 Percent Baseline Mass Reduction for MY2014 ($/kg)
                                             5-378

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                       Passenger Car/CUV DMC Curve Change
                       W/ 5% MR Baseline MY2014 ($/vehicle)
                                                                    20%
                                       % Mass Reduction
                         •5%MR Baseline 2014MY
-0%MR Baseline 2014MY
   Figure 5.125 DMC Curve Adjusted for Car/CUV with 5 percent Baseline Mass Reduction for MY2014
                                        ($/veh)
   Total Costs for Car/CUV Mass Reduction

   As described in Section 5.3.2.2.2, this assessment adopts a methodology for estimating the
indirect costs of mass reduction based on separating direct manufacturing costs according to
whether the components are purchased supplier parts, or OEM-produced. The OEM's markup on
supplier produced components is expected to be less than the markup on an OEM produced
component since the supplier markups are included in the OEM piece price to the supplier.

   Figure 5.126 and Figure 5.127 show the resultant DMC cost curve, ICM curve and Total Cost
curve for those vehicles designated to be assigned the passenger car cost curve for mass
reduction based on MY2008.  This curve is based on a vehicle with no baseline mass reduction
differences noted between MY2008 and MY2014. If a vehicle were to have a baseline mass
reduction noted for MY2014 then the Total Costs would increase due to the Direct
Manufacturing Cost increases.
                                            5-379

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                          Passenger Car Cost Curve ($/lb v %MR)
                                         % Mass Reduction
                            •DMC
                                       •1C
                                                 TC
Log. (DMC)
         Figure 5.126 Resultant Passenger Car Cost Curve (2013$/lb, 3000 pound vehicle shown)

   Note: DMC, 1C using ICMs and Total Cost, MY2008 with 0 percent Baseline MR, applicable
in MY2020 with learning effects determined by learning curve 30 (see Section 5.3.2.1.4)).
                              Passenger Car Cost Curve ($/vehicle vs %MR)
                       600
                      -300
                                                % Mass Reduction
                                            •DMC
                                                     •1C
                                                            TC
       Figure 5.127 Resultant Passenger Car Cost Curve (2013$/vehicle, 3000 pound vehicle shown)

   Note: DMC, 1C using ICMs and Total Cost, MY2008 with 0 percent Baseline MR, applicable
in MY2020 with learning effects determined by learning curve 30 (see Section 5.3.2.1.4)).
                                             5-380

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Comparison of Data for Lightweight Car/CUV with Aluminum BIW

   In order to assess the opportunity to reduce the mass of passenger cars and CUV's beyond
what was considered in the cost curves discussed, the EPA reviewed alternatives for all
aluminum body-in-white. The alternatives presented here are not reflected in the draft TAR cost
curves, but are included to recognize that EPA does not expect a significant inflection upward in
cost with mass reduction beyond what has been considered in the draft TAR analysis of 2008 era
vehicles. The solution points from the lightweight studies for the TAR contain AHSS BIW and
aluminum intensive components correspond to mass reduction levels of 17.6 percent and just
over 20 percent for the CUV and passenger car holistic vehicle studies respectively. In addition
to the Aluminum BIW discussed below, the feasibility of achieving higher levels of mass
reduction was shown in the work by DOE/Ford/Magna in which 23.5 percent mass reduction
was achieved relative to a MY2013 Fusion000 for the Mach 1 design, as described in Section
5.2. The overall BIW design was multi-material with 64 percent aluminum, 29 percent steel and
7 percent hot stamping. A number of vehicles were built and crashed, including IIHS ODB, with
acceptable results and several notes for further improvement in the BIW design  to CAE
predictive correlation were noted.  Costing was not a part of this project, however the SAE paper
states "multi-material automotive bodies can achieve weight reduction with cost effective
performance." 578

   Several additional design solutions at higher levels of mass reduction with all aluminum BIW
were developed using the Venza and Accord-based studies as starting points, as shown in Figure
5.128, along with an extrapolation of the best fit Car/CUV cost curve from Figure 5.121.
                   Simulation of Avg Pass Car/Midsize  CUV, Al BIW
                   Cost Curve Extended Through 30%- Base 2008
                                      NHTSA w/AI BIW
                                      ($2.83/kg, 23.2%)
                              Al Assoc, Midsize CUV, EPA
                              CAE, w/AI BIW
ARB Data point
w/AI BIW on CUV
 -0.64%/kg, 31%)
                                       % Mass Reduction
             Figure 5.128 Car/CUV DMC Curve Extended to Points with Aluminum BIW
GGG The MY2013 Fusion was one redesign beyond the 2008 era Fusion. The base vehicle is approximately 250 Ibs
  heavier and the top trim is approximately 100 Ibs heavier in 2013 compared to 2008. The 2013 Fusion is
  approximately 2.80sq ft larger in footprint compared to the 2008 era Fusion and slightly taller and wider overall.
  Several safety features were also included. (https://en.wikipedia.org/wiki/Ford_Fusion_(Americas))
                                             5-381

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Figure 5.129 shows two points for the CUV aluminum intensive solution. One point is from
the ARB-sponsored study by Lotus Engineering579 and one point is from the Aluminum
Association study through EDAG.580 The ARB full vehicle data point with optimized BIW
design and reduction of BIW components is 531kg (31 percent) mass reduction at -$0.64/kg.
The Aluminum Association study of an all-aluminum BIW, based on material replacement into
the CAE model from the original U.S. EPA Midsize CUV study, resulted in a total vehicle
solution of $1.12/kg at a total of 476kg (27.8 percent) mass reduced. NHTSA studied the
aluminum intensive vehicle design for the passenger car (based on the MY2011 Accord) and the
result is a point at $2.83/kg for 23.2 percent.

   Table 5.145 shows the detailed results of the studies. The cost/kg estimate for the NHTSA
estimated point are likely overestimated given the recent reduction in the commodity price for
aluminum and comments in the 2001 JOM source document used for the cost estimate indicates
that costs have very likely decreased since this work was completed.HHH581 Similarities are seen
in the mass reduction results between the two aluminum intensive projects for the Midsize CUV
(Lotus Engineering and EDAG) and these include the total BIW/closure/bumper total mass
reduction which is only 6kg apart (201.7kg v 207.7kg respectively). The differences between the
two projects include the BIW designs used and the resultant estimated costs. The EDAG study
used the existing BIW design and the materials of aluminum alloy sheet, extrusion and casting.
The Lotus Engineering solution also utilized the different aluminum components while
optimizing component aggregation as only 169 components were used in the BIW compared to
the original 419 and significant savings with the new manufacturing processes were assumed.
       Table 5.145 Three Aluminum Intensive Vehicle Design Summary - DMC ($), %MR and $/kg
Aluminum BIW,
Closures, Chassis
BIW
Closures/Fenders
Bumpers
TOTAL
Total Vehicle
$/kg
2012 ARB/Lotus
(midsize CUV- 1711kg)
Mass save
(kg)
140.7
59
2
201.7
530
Cost
($)
239
-381
9
-133
-342
-$0.64/kg
2012 Al Assoc/EDAG
(midsize CUV -1711kg)
Mass save
(kg)
162.2
43.2
2.3
207.7
464*
Cost
($)
780
106
8.6
894.6
+520*
$1.12/kg
2012 NHTSA/Electricore/
EDAG (Pass Car- 1480kg)
Mass save
(kg)
113
44
-
157
343.6
Cost
($)
782
153.7
-
935.7
971.9
$2.83/kg
Note: *adjusted for changes in the EPA baseline Midsize CUV cost curve into which the aluminum BIW was placed


   Future Work for Proposed Determination

   EPA recognizes that mass reduction technology will play an important role in meeting the
2022-2025 MY standards.  The agency will continue to monitor and research developments in
material development, material substitution approaches, design optimization and manufacturing.
HHH Investigation into the supporting documentation for the analysis revealed that the information was taken from a
  2001 article in the Journal of Minerals, Metals and Materials Society. The article states "In fact, design
  developments by Audi already have resulted in significant cost reductions between its first- and second-
  generation vehicles. These have come about through parts consolidation, process substitutions, and part
  simplification."
                                             5-382

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
EPA plans to revisit the assessment of overall mass reduction costs and the evaluation of mass
reduction in the baseline fleet for the proposed determination.

5.3.4.6.1.2    Cost Curve for Light Duty Trucks

   The cost curve for light-duty trucks developed for this assessment as described below
represents an estimate of the Direct Manufacturing Cost (DMC) for mass reduction technologies
that are expected to be broadly available in 2020. Total costs, which are made up of both DMCs
and Indirect Costs (ICs), are also presented herein. More detail on the methods for calculating
indirect costs and learning for mass reduction are provided in Section 5.3.2.

   Light Duty Truck DMC Curve Generation

   The LDT direct manufacturing cost curve was created through combining the results of the
EPA MY2011 base LDT and NHTSA MY2014111 base LDT lightweighting studies which are
outlined in Section 5.2. Development of the LDT DMC curve for use in EPA modeling is
completed in the following steps outlined in Table 5.146. This section also includes discussion
of the complete LDT cost curve with Direct Manufacturing Cost (DMC), Indirect cost (1C) and
resultant Total Cost (TC).
                          Table 5.146  LDT DMC Curve Development
STEP
1
2
3
TASK
Begin with the MY2011 and MY2014 Light Duty Truck cost curves
(the vehicles are of different design eras)
Translate the NHTSA LDT DMC cost curve to use a similar methodology as the EPA LDT DMC
Average the two LDT curves using the best fit line for each curve. Account for difference
between the two studies.
cost curve
in eras
   STEP 1: The cost curve for the Car/CUV was based on two 2008 era vehicles and hence
represents the technology of lightweighting on 2008 era vehicles. The MY2011 Silverado 1500
design and the MY2014 Silverado 1500 are from two different design eras.  The MY2011
Silverado 1500 is a 2008 design era vehicle. The MY2014 Silverado 1500 is the next redesign
and has been redesigned with safety complianceJJJ, some lightweighting and slightly larger size.
All of these features will come into play later on in the LDT cost curve development process
described herein.  The curb weight difference between the MY2011 and MY2014 light duty
truck study vehicles is 22kg as shown in Table 5.147.
   Table 5.147  Comparison of MY2011 and MY2014 Crew Cab Silverado 1500
                                                                          582

Cabin Design
2x4, 4x4
Truck Bed Length
MY2011 Silverado 1500
Crew Cab
4x4
5.8ft
My2014 Silverado 1500A
Crew Cab
4x4
5.8ft
111 The final report for the MY2014LDT was not available in time for this Draft TAR analysis. The Proposed
  Determination will contain an updated analyses given the final mass reduction and cost information from the final
  MY2014 LDT lightweighting study.
111 The safety design features in the MY2014 Silverado include higher compliance to the IIHS small overlap crash
  test as well as compliance with FMVSS crash tests that came in during the 2008 and 2014 timeframe.
                                             5-383

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
Engine
Transmission
Wheelbase
Track Width
Curb Weight*
5.3LV8FFV(315hp)
4 speed
143.5 inches
68 in front, 67 in rear
2454kg
5.3LV6(355hp)
6 speed
143.5 inches
68.7-68.9 in front, 67.6-67.9 in rear
2432kg
*The curb weights were from the EPA and NHTSA light duty truck lightweighting studies. The mass decrease for
these two trucks is 22kg.
AThe 2014 LOT incorporates materials to address the safety standards that came into effect between 2012 and
2014. The 2014 LOT is also slightly larger
   The two light duty truck DMC curves are reviewed to assure the two cost curves utilize a
similar methodology.

   The EPA light duty truck direct manufacturing cost curve is shown in Figure 5.129 and is
based on EPA's light duty truck light-weighting study (MY2011 Silverado 1500). The
lightweight design is aluminum intensive combined with an AHSS frame. The DMC curve was
created using the similar cost curve methodology in the EPA Midsize CUV study.  EPA's
methodology is to a) cumulatively add all of the primary mass reduction technologies (not
dependent on vehicle mass for optimization) and costs (green/top curve), b) add an NVH
allotment component by ratio across all mass save steps,  and c) determine secondary mass at
primary solution point and ratio secondary mass savings  across the primary mass curve to create
the compounded curve (purple/bottom curve).  The original engineered solution point to the
study was 20.8 percent mass reduction at $4.35/kg. The  cost curve on the MY2011 LDT result
was modified with a re-evaluation of the NVH allotment to 15kg from 50kg, both at $3/kg, based
on new NVH technology.  The resultant cost curve is shown in Figure 5.129 with an engineered
solution point of 22  percent mass reduction at $3.92/kg direct manufacturing cost.  The MY2011
LDT was the same design cycle as the MY2008 LDT582.
                          US EPA Light Duty Truck Cost Curve, 2008 Basis
                                 3E«O6«> • 971919»* • 1S2207«>• 12534»' «SSg 1»» • 12 302
                                        If =09975
                                      % Vrhidc M.W Kcdurtioo
                                                                                   583
  Figure 5.129 U.S. EPA Light Duty Pickup Truck Direct Manufacturing Cost Curve, MY2008 Design

   The NHTSA light duty truck direct manufacturing cost curve is shown in Figure 5.130 and is
calculated by NHTSA based on the data from the May/June pre-peer review version light duty
truck light-weighting study (MY2014 Silverado 1500).KKK NHTSA's cost curve methodology
   The NHTSA LDT final report was not available in time for this Draft TAR analysis.
                                             5-384

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
changed from that used in development of the cost curve for the passenger car.  The initial curve
was created by a) cumulatively add mass and cost of primary technologies (from glider) and then
b) apply downsized components (non-glider) at the main vehicle solution point (Aluminum
Intensive and AHSS frame (noted as 'AHSS+A1 Solution (LWV)')) and c) connect the end of the
cumulatively added technologies to the vehicle  solution point with a straight. The cumulative
add of the primary  (glider) technology ends at approximately 12.5 percent. The solution point,
located  at 17.5 percent, includes non-glider technologies (engine, transmission, exhaust, etc.) and
any other secondary components optimized for  maximum mass reduction  for the solution point.
The costs are assumed to increase linearly from the 12.5 percent point to the 17.5 percent point.
Additional technologies from the Aluminum Solution and the CFRP Solution are applied to
achieve a total of 20 percent mass reduction from the MY2014 LDT design. Unlike the
passenger car cost curve, the cumulative cost curve does take into account some of the secondary
mass reduction opportunities to account for different powertrain options and platform sharing.
Any changes to NHTSA's interpretation of the data/cost curve in the final  light duty truck
lightweighting study will be incorporated for the Proposed Determination.
                              NHTSA Light Truck Mass Reduction Cost Curve
                           y - 172698x5 - 40530x* - 2587.5x3» 1022*' - 25.202x * 0.1934
                                      K' = 0.9965
                                                        'AHSS-AL
                                                         Solution (UW)
                            4%   6%   8%   10%  12%  14%  16%  18%  20%  22%  24%  26%
                                             ?. MR
                                                                                LLL
      Figure 5.130 NHTSA Light Duty Truck Direct Manufacturing Cost Curves, MY2014 Design

   STEP 2: Translate the NHTSA LDT DMC curve to use a similar methodology as the EPA
LDT DMC curve.  The cost curves from the EPA MY2011 LDT study and NHTSA interpreted
curve from the MY2014 LDT study data were similar in methodology however differences still
remain. Table 5.148 contains a comparison of the two cost curve methodologies. For
determination of the final cost curve for the light duty truck, the cost curve for the MY2014
based study is recalculated using the MY2011 cost study methodology.
   See Section 5.2 for NHTSA light duty truck lightweighting study.
                                             5-385

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
             Table 5.148  Light Duty Truck Study Cost Curve Methodology Comparison
Topic
COSTS and
TIMEFRAME
TECHNOLOGY
IDENTIFICATION
BIW/FRAME
BIW/FRAME
SAFETY
CRASH/NVH in
CAE
Base Vehicle
Comply with IIHS
Small Overlap
ORDERING OF
PRIMARY MR
TECHNOLOGIES
FOR COST CURVE
NVH (noise)
INCORPORATION
OF SECONDARY
MASS SAVINGS
(SMS)
Cost Curve
Expression
EPA
(Peer Reviewed MY2011 LOT Study)
Costs are representative of 2020
LOT of MY2011 Design (similar to 2008)
A large number of individual vehicle
components including engine, trans.
Material replacement, material grade and
gauge optimization.
-Include NVH, FMVSS, etc. crash results
-IIHS small overlap evaluated in study by
Transport Canada in which base LOT crash
used for CAE development and solution
developed in CAE for mass add determination
Poor rating likely - review of crash results of
MY2013 Silverado 1500
-All technologies with primary mass reduction
ordered lowest to highest $/kg
-Only grouped ideas as required for
implementation feasibility.
Originally 50kg at $150*
(adjusted 15kg and $45)
-Study examined a number of major
components that could be made smaller due
to a lighter vehicle at the main solution point.
-SMS based on downsizing
-SMS ratio'd at each level of mass reduction
from 100% SMS at solution point back toward
zero percent mass reduction.
- Cumulative addition of best value primary
mass reduction components, up through
aluminum closures, and resulted in a
continuous curve for the AHSS BIW and
aluminum intensive solution.
-Compounded curve includes primary +
secondary percent mass red and $/kg.
NHTSA
(NHTSA Curve from Data of the Pre-Peer Review
MY2014 LOT study)
Costs are representative of 2017
LOT of MY2014 Design (new design)
A number of components and adoption of
systems from lighter vehicles (engine, etc.).
Material replacement, computer optimization of
load paths, grade and gauge optimization for
AHSS frame with Al intensive design only
-Include NVH, FMVSS, etc. crash results
-IIHS small overlap based on observation of F150
IIHS crash results, applied to Silverado 1500
Marginal according to IIHS website
-Glider technologies for AHSS frame/AI intensive
solution used as primary tech
-Technologies ordered in lowest to highest $/kg
order and cumulatively summed
-Whole vehicle solution and technologies for
other materials plotted and used to achieve 20
percent
Incorporated in vehicle load path design
Additional ~3kg at $10.69
-Applied at solution points only
-Inferred in line connecting end of primary
cumulative curve and vehicle solution for AHSS
Frame/AI Intensive solution point
-Inferred in points up to 20 percent mass
reduction
-Cumulative add glider technologies for AHSS
frame with Al intensive solution
- SMS at solution point
-Additional aluminum and CFRP technologies used
to reach 20 percent mass reduction
Note:
* Learned in 2015 through the DOE/Ford/Magna cosponsored Machl/Mach2 SAE papers
   In order to combine cost curves, it is important that the two cost curves are considered in the
same methodology since the exact same technologies were not evaluated. The NHTSA cost
curve shows additional technologies beyond the main solution point, of AHSS+A1 (LWV)
Solution, to achieve 20 percent mass reduction from the MY2014 basis. The following analysis
will show how these technologies are not required for the combined cost curve to achieve 20
                                            5-386

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
percent mass reduction on a 2008 era basis. The following steps were performed to translate the
NTHSA resultant study findings for the AHSS+A1 Solution (LWV) into the EPA LOT cost
curve methodology.

   a) Evaluate the data used to create the AHSS+A1 Solution (LWV) point and assure all
technologies are separated into their primary and secondary components, see Table 5.149.MMM

   b) Create the primary cost curve.  The primary non-compounded cost curve, such as that
shown in the green/top curve of Figure 5.129, represents cumulatively added mass reduction and
costs for individual primary mass reduction technologies for the LWV solution. These primary
technologies can be adopted without concern of the resultant mass of the vehicle.

   c) Determine the secondary mass and cost savings at the AHSS+A1 Solution (LWV) point.
Secondary mass and cost reduction technologies are determined and applied for a number of
components, including the chassis and powertrain that could be redesigned to reflect the
reduction in load associated with a reduced vehicle curb weight.

   d) Create the compounded cost curve. Ratio the secondary mass and cost savings from the
AHSS+A1 Solution (LWV) point across the percent mass reduction from the primary cost curve.
In Figure 5.129 this is shown as the purple/bottom curve.  The translated curve in Figure 5.131,
although not evident, does contain the NHTSA's approach for only applying secondary mass at
points greater than 10 percent was maintained. The secondary mass and cost savings were offset
by the mass  reduction technology at that point. Figure 5.132 is an expression of the curve in
$/vehicle v.  percent mass reduction.

   e) Create best fit curves to the data for use in averaging.
MMM Figure 5.130 became available in the May/June 2016 timeframe and will be updated when the final report
  becomes available. The cost curve for the EPA mass reduction modeling was completed prior to the availability
  of this final curve from the May/June timeframe.
                                             5-387

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
 Table 5.149 Re-Designation of Secondary Technologies Listed in NHTSA Light Duty Truck Lightweighting
                                             Report
System
Cab
FESM (per vehicle)
Radiator Support Structure
Front Bumper
Rear Bumper
Chassis Frame
Towing Hitch
Front Suspension
Rear Suspension
Wheels &TiresNNN
Engine
Transmission
Drive Shafts
Fuel System
Exhaust
Brake System000
Water Cooling
Battery
Fuel
Technology
Aluminum
Aluminum
Al & Cast Magnesium
AHSS
AHSS
AHSS
AHSS
Lower Control Arm (Al to AHSS)
Others downsized
Leaf spring: 1 steel +2FGRP
Others downsized
eVOLVE Rims
Downsize
Rear Diff Housing to Al
Other Downsize
Downsize
Downsize fuel tank
Downsize
Keep Iron Discs
- Master Cyl DS
- front discs DS
-Front calipers (to Al) and DS
-Front Pads DS
- Rear Discs DS
- Rear Calipers (to Al and DS)
-Rear pads DS
-Park Brake to EPB
-Caliper Supports DS
Downsize
Downsize
Less fuel used with smaller tank
Primary
*
*
*




*
*
*

*



*



Secondary



*
*
*
*
*
*

*
*
*
*
*
*
*
*
*
1411 The material for the wheels were changed, but the size remained the same - hence primary mass reduction
  change.
000 The brakes contained several technology changes. Some changes were material/design and these are primary
  changes, downsizing of components are secondary and since there was some of both in the work then the mass
  reduction for brakes falls into both categories.
                                                 5-388

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                    Whole Vehicle NHTSA* Light Duty Truck Cost
                    Curve  ($/kg vs %MR, 2014 basis), SMS at 10%
                   5.00  y = 6.2543ln(x)+ 27.254
                            R2 = 0.9065  	—,
                            2%    4%   6%    8%   10%   12%   14%   16%   18%


:

!2 = 0.9815








                                          % Mass Reduction
  Figure 5.131 NHTSA Light Duty Truck (MY2014) Data Points in EPA Cost Curve Methodology ($/kg v
                        %MR) for Aluminum Intensive with AHSS Frame
\M
Ci
v 120°
T5
ii 1000
>
£ 800
4J
M
Direct Manufacturing C<
($/vehicle
M Nl *» C
g o 8 8 S
0 -
tole Vehilce NHTSA* Light Duty Truck Cost
jrve ($ vs %MR, 2014 Basis), SMS at 10%+










S




/
-*



^




^-»
^




^




^*
t*^




^




s>






!»_-•*--
2% 4% 6% 8% 10% 12% 14% 16% 18%
% Mass Reduction
Figure 5.132 NHTSA Light Duty Truck (MY2014) Data Points in EPA Cost Curve Methodology ($/vehicle v
                                         %MR)

   Comparing the best fit curve calculated results from Figure 5.131 with NHTSA's cost curve
presented in Figure 5.130, it is observed that the calculated $/kg at the 17.5 percent mass
reduction point is $3.02/kg with EPA's analysis of the LDT data while the NTHSA cost curve
best fit curve result is $3.55/kg. The differences are likely due to the offset of the best fit curve
offset at this point in Figure 5.130 and partially due to the preliminary nature of the data
available from NHTSA's light duty truck study at the time of EPA's calculation of the translated
                                             5-389

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
cost curve.  Both curves show the same amount of mass reduction at a cost save which is
approximately two percent.  A direct comparison between the EPA and NHTSA results cannot
yet be made due to the fact that the curves represent different vehicle design years. Additional
cost curve manipulation must occur before the cost curves can be averaged.

   STEP 3: To average the two LDT DMC curves using the best fit line for each curve, the
differences  in eras between the  two studies must first be addressed. As done in EPA modeling,
the 2008 era cost curve is adjusted for differences in curb weight (with factors for safety and
footprint) to match the 2014 era vehicle.  The difference in curb weight is found in Table 5.147
and is noted as 22kg. Adjustments for safety, such as mass add to comply with new 2010-2014
FMVSS standards and IIHS small overlap are credited to the MY2014 vehicle, along with
adjustment  for larger footprint and are performed per steps outlined in 5.3.4.6.2.1. The total
mass difference between the MY2011 LDT and MY2014 LDT is 22kg (curb weight)+l 1.6kg
(FMVSS safety allowance) + 22kg (IIHS small overlap allowance) 7.9kg (footprint calculation)
which equals 63.5kg or 2.6 percent of the MY2011 LDT. All of the mass reduction ideas in the
cost curve within the 2.6 percent are cost save ideas and as a result the resultant cost curve
increases in $/kg for these ideas are not available to offset cost increase ideas.  The resultant EPA
LDT cost curve to represent a MY2014 vehicle is illustrated in Figure 5.133.
            I   58
                   US EPA LDT CURVE ADJUSTED FOR BASELINE MASS (-2.6%)
                                                       y=-36,661x- + 33.138x- 0.221
                                                             R1 = 0.9854
                                              1	1	1	1	1-
                             4%   6%    8%    10%    12%   14%    16%   18%  20%
                                      % Vehicle Mass Reduction
          Figure 5.133 EPA Adjusted MY2011 LDT Cost Curve for 2014 LDT Design (-2.6%)

   The two MY2014 based direct manufacturing cost curves for the Aluminum with AHSS
frame solution, represented by the best fit line for each cost curve, as shown in Figure 5.131
(NHTSA*=data by EPA) and Figure 5.133 (EPA), are then averaged together. The result cost
curve is shown in Figure 5.134.  The dip in the NHTSA curve is due to the application of
secondary mass savings at points of 10 percent mass reduction and greater.
                                             5-390

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                  Combine EPA & NHTSA* MY2014 LOT Cost Curves
                                                          = 1.8513ln(x) +6.9611
                                                             R2 = 0.9558
                                                           15%
                                 20%
              :
             5
                        EPA LOT a dj
                                        % Mass Reduction
NHTSA*
Log. (Avg)
      Figure 5.134 Combined Direct Manufacturing Cost Curve using EPA LDT and NHTSA LDT

   The average curve shown in Figure 5.134 is specific to the example for the MY 2014 LDT.
To create a cost curve which can be applicable to all vehicles with various MY2014 percent
baseline mass reduction, the cost curve in Figure 5.134 must be brought back to a 2008 era base.
EPA uses 2008 era base cost curves in its passenger car/midsize CUV curve and making the
LDT curve a 2008 era base will be consistent.  The MY2014 base DMC curve is converted back
to MY2008PPP by adding in the removed points from the EPA LDT cost curve (0-2.6 percent and
all cost save) and the resultant curve is  shown in Figure 5.135 with costs per vehicle shown in
Figure 5.136. Note that the overall cost is reduced due to the initial points being all cost save
items.QQQ

   The DMC cost curve shown in Figure 5.135 is applied as-is for vehicles with no mass
reduction identified in their MY2014 baseline. For LDT with a MY2014 baseline mass
reduction noted,  such as the 2.6 percent noted on the MY2014 Silverado 1500, the cost curve
will be adjusted with the same methodology as used to form the EPA MY2011 LDT curve to a
MY2014 LDT curve, previously described. This methodology results in further mass reduction
technologies being more expensive on vehicles that incorporate mass reduction technologies that
result in a change in curb weight from their previous design. This methodology also results in a
reduction in the maximum mass reduction percentage that can be applied as noted if comparing
Figure 5.135 and Figure 5.134.
ppp The MY2011 Silverado 1500 is of the same design cycle as the MY2008 Silverado 1500.
QQQ for EPA's analysis, the LDT DMC cost curve is being applied to all vehicles designated as a truck and this
  include some SUV's and CUV's which meet the truck definition and may be unibody in design.
                                             5-391

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                        MY2008 Light Duty Truck DMC Curve
                                    ($/kgvs%MR)
             ,3  -20%
                -4
10%
15%
20%
^r
f


y = 3.4888ln(x) +8.3412
R2 = 0.9932









                                       % Mass Reduction
                        • Combined LOT
                                          •FRM
         Log. (Combined LOT)
    Figure 5.135 Direct Manufacturing Cost Curve for 2008 Era Light Duty Trucks (2013$/kg vs %MR)
                               MY2008 Light Duty Truck
                               DMC($)/Vehicle vs%MR
                                       10.0%       15.0%

                                       % Mass Reduction
                          25.0%
  Figure 5.136 MY2008 Light Duty Truck DMC (2013$/Vehicle for a 6000 pound truck) vs Mass Reduction
   Total Costs for Light Duty Truck Mass Reduction

   As described in Section 5.3.2.2.2, this assessment adopts a methodology for estimating the
indirect costs of mass reduction based on separating direct manufacturing costs according to
whether the components are purchased supplier parts, or OEM-produced.  The OEM's markup on
supplier produced components is expected to be less than the markup on an OEM produced
component since the supplier markups are included in the OEM piece price to the supplier.

   Figure 5.137 and Figure 5.138 show the resultant DMC cost curve, ICM curve and Total Cost
curve for those vehicles designated to be assigned the light duty truck cost curve for mass
                                             5-392

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
reduction based on MY2008. These curves are based on a vehicle with no baseline mass
reduction differences noted between MY2008 and MY2014. If a vehicle were to have a baseline
mass reduction noted for MY2014 then the Total Costs would increase due to the Direct
Manufacturing Cost increases.
                                         MY2020; Truck
                  1.00
                o -1.00
                  -2.00
                  -4.00
                  -5.00
                                        -DMC 	1C 	TC
        Figure 5.137 Resultant Light duty Truck Cost Curve (2013$/lb, 6000 pound vehicle shown)

   Note: DMC, 1C using ICMs and Total Cost, MY2008 with 0 percent Baseline MR, applicable
in MY2020 with learning effects determined by learning curve 30 (see Section 5.3.2.1.4))
     Figure 5.138 Resultant Light Duty Truck Cost Curve (2013$/vehicle, 6000 pound vehicle shown)
                                              5-393

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Note: DMC, 1C using ICMs and Total Cost, MY2008 with 0 percent Baseline MR, applicable
in MY2020 with learning effects determined by learning curve 30 (see Sections.3.2.1.4))

   Future Work for Proposed Determination

   EPA recognizes that mass reduction technology will play an important role in meeting the
2022-2025 MY standards.  The agency will continue to monitor and research developments in
material development, material substitution approaches, design optimization and manufacturing.
EPA plans to incorporate NHTSA's final LDT DMC curve technology points as well  as revisit
the assessment of overall mass reduction costs and the evaluation of mass reduction in the
baseline fleet for the proposed determination.
5.3.4.6.2
Mass Reduction in the Baseline MY2014 Fleet
   The baseline fleet methodology for this Draft TAR has been updated from the FRM for model
year and for starting percent mass reduction as shown in Table 5.150. For the FRM, the
MY2008 fleet was the baseline fleet and it was assumed that each vehicle in the baseline had
zero mass reductionRRR irrespective of any differences in vehicle type, the use of lightweight
materials, or overall vehicle design and implementation.  Each vehicle was also assumed to have
the same maximum potential for additional mass reduction.

   For the Draft TAR, mass reduction continues to be defined as a decrease in curb weight.  This
definition provides a direct relationship between the level of mass reduction, the cost, and the
benefits achieved. As shown in Table 5.150, the Draft TAR is updated to a MY2014 baseline
and is adjusting the incremental costs and the maximum mass reduction potential on the percent
mass reduction that is calculated in the MY2014 baseline. This updated approach has important
implications for cost estimation since mass reduction becomes increasingly more expensive  at
higher levels.
                    Table 5.150 Draft TAR Mass Reduction Baseline Revisions
TOPIC
Baseline MY
Percent Mass
Reduction
Potential Maximum
Mass Reduction
Mass Reduction
Cost
FRM
2008
0 percent- all
vehicles
Same for all vehicles
Same for all vehicle
Draft TAR
2014
-Vehicle specific: MY2014-MY2008 curb weight difference
plus MY2014 footprint and safety mass adjustments.
-OR Vehicle Specific: OEM lightweighting trend from
current vehicles with MY2008/MY2014 models applied to
new 2014 models.
-Calculations use 0.5 increments percent mass reduction
Differs depending on MY2014 baseline calculated percent
mass reduction.
(Max= 20 percent - MY2014 baseline percent)
Cost curve costs are modified depending on MY2014
baseline calculated percent mass reduction.
   After evaluating a variety of alternatives, EPA estimated mass reduction for each vehicle in
the MY2014 baseline fleet relative to the corresponding MY2008 vehicle. If a vehicle did not
have a MY2008 counterpart then the sales weighted average percent mass reduction over the
111111 In terms of dollars per kilogram curb weight reduction.
                                             5-394

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
OEM's nameplate product line is used to represent the expectation of the amount of mass
reduction technology within the vehicle. This consumer-oriented "lineage" approach is similar to
model-level comparison, although with additional consideration for models newly introduced or
renamed after MY2008.  The following sections describe the calculations in more detail.

   The methodology considers AWD/4WD v 2WD differences as well as 2014 mass increases
due to new safety requirements and changes in footprint over 2008. Limitations to this analysis
include 1) no adjustment for engine size differences between trim levels, 2) no adjustment for
hybrid or EV trim levels (typically smaller volume and high mpg), and 3) mass additions due to
future potential safety regulations.

5.3.4.6.2.1    Vehicles with MY2008 andMY2014 Production

   Vehicle baseline percent mass  reduction was determined by subtracting the MY2014 curb
weight (with adjustments) from the MY2008 curb weight (with adjustments).  The base curb
weight data for MY2008 was taken directly from the data used in the Light Duty GHG 2017-
2025 FRM. The MY2014 curb weight data was adopted from information in the ARB sponsored
study Control Tec study584, which assembled the baseline from EPA test vehicle weight data and
other sources.

   The following paragraphs describe the methodology utilized in the creation of the MY2014
baseline database.

   1. Sales weight the 2008 models and related trim levels - per vehicle

   When sales weighting the curb weight of several trim levels within a vehicle model in two
different years (2008  and 2014), in order to gain a more accurate picture of change in curb
weight due to mass reduction technology, one needs to ensure that unique vehicle characteristics
do not influence the overall vehicle sales weighted mass in either year.  One vehicle attribute that
would influence trim  level mass is 4WD/AWD v 2WD. A mass  allotment is added to 2WD
vehicles and then the  trim levels are sales weighted within the respective years.

   a. Adjust the 2008 curb weight for 4WD/AWD v 2WD variations.

   A report funded by Transport Canada with Pilot Systems included the evaluation of mass
differences in AWD v 2WD on three different vehicles. The mass amount was determined
through a review of three different AWD  systems - Jeep Cherokee, Ford Fusion and VW Tiguan.
The mass differences were  135kg, 72kg, and 78kg respectively for an average of 95kg or 2091bs.
A value of 2001bs was used to provide an adjustment to minimize the influence of this vehicle
characterization difference  in the baseline sales weighted curb weight.585

   b. Sales weight the 2008 vehicle trim levels per vehicle.

   2. Sales weight the 2014 models and related trim levels - per vehicle adjusting for
footprint and safety

   The same AWD/4WD v 2WD  adjustment is made on the 2014 vehicle trim levels. Vehicle
trim levels are then sales weighted. Prior to calculating the final  MY2014 baseline mass
reduction allotments,  adjustments to the MY2014 curb weight were made to account for
footprint, which is a change in vehicle characteristics that influence CAFE and GHG target
levels,  and MY2008-MY2014 increased FMVSS and IfflS crash requirements.
                                             5-395

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   a.   Adjust the 2014 curb weight for 4WD/AWD v 2WD (as with 2008). Same mass
difference is utilized as for 2008 models.

   b.   Sales weight the 2014 vehicle trim levels per vehicle.

   c.   Adjust the 2014 curb weight data for footprint increase

   Footprint is allowed to increase from 2008 to 2014 without penalty and as a result a
kg/square foot credit was applied to footprint differences between the 2008 and 2014 vehicles.
The main idea behind this action is that if vehicles remain at a similar curb weight but increased
in size then they did incorporate mass reduction technology to offset the increased footprint.

   The methodology used to determine the footprint mass credit (mass/sqft) is as follows:

   1) Identify the portions of the vehicle that would be affected by an increase in footprint area
(passenger compartment back seat leg room).

   2) Gather mass data from a number of vehicles, using the A2Macl database (mass) for BIW,
glass, and interior masses.   Choose vehicles which span the 6 vehicles classes (small car,
standard car, large car, small SUV, large SUV and truck).

   3) Gather footprint data on the same vehicles.

   4) Determine the mass/sqft by dividing the total mass of these components per vehicle by the
total vehicle footprint.  The resultant average mass/area per vehicle class is shown in Table
5.151.
                 Table 5.151 Footprint Density per Vehicle Class (Ib/sqft and kg/sqft)
AvgFP
Density
Ib/sqft
kg/sqft
Small
18.56
8.43
Midsize
20.07
9.12
Large
21.13
9.60
Pickups
11.88
5.40
Small MPV
20.72
9.42
Large
MPV/Truck
23.56
10.71
   The averages in Table 5.151 were applied to all respective vehicles for which it was
determined there was an increase in footprint in 2014 compared to 2008. Table 5.152 shows the
application of the average kg/sqft mass credit to the Acura MDX and RDX, which are designated
Large MPV/Truck.

   1) Determine if the vehicle footprint did increase from 2008 to 2014. The vehicle footprint
and footprint of related trim levels, if applicable, were sales weighted for both 2008 and 2014
and the 2008 model footprint average was subtracted  from the 2014 model footprint average. A
positive number meant an increase in overall footprint ("Delta FP"). (Note: if vehicles changed
names in 2014 compared to 2008 then this was noted  and the vehicles still compared with each
other)

   2) Determine the mass increase by multiplying the change in footprint by the footprint
adjustment in the appropriate vehicle class from Table 5.151.

   3) Add the mass credit to the original Delta CW for a new change in mass reduction.

   4) Recalculate the adjusted curb weight percentage
                                             5-396

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   For example, Table 5.152 shows that the adjustment factor for the Acura MDX and RDX, is
10.71 kg/sqft, as from , for they are both considered Large MPV/Truck.  Based on the change in
square feet and the footprint density factor for large MPV's, the credit for mass reduction in the
MDX and RDX are 3.2kg and 13.9kg respectively and the overall % curb weight changes 0.2
percent and 0.8 percent respectively.

                Table 5.152 Examples of Mass Footprint Adjustment (single vehicle)
Make


Acura
Acura
Acura
Acura
Model


MDX
MDX
RDX
RDX
LineagelD


2
2
3
3
MY


2008
2014
2008
2014
Delta
CW
(kg)

-238

-94.1
Delta
CW%


-11.5%

-5.3%
Delta
FP
(sqft)

0.3

1.3
FP
Density
(kg/sqft)

10.71

10.71
AdjFP
(kg)


3.2

13.9
AdjCW
(kg)


-241

-108
Adj CW%



-11.7%

-6.1%
   Footprint changes were not accounted for between trim variants. Examples include light duty
trucks with different cab designs and box lengths,

   d) Adjust the 2014 curb weight data for mass credit for safety (FMVSS and IfflS)

   Several NHTSA safety regulations have come into effect between 2008 and 2014.  Table
5.153 lists the specific FMVSS test as well as the estimated car and light truck mass increase.
The amount of mass increase for the NHTSA/FMVSS safety regulations was determined from
information from NHTSA's 'Corporate Average Fuel Economy for MY2017-2025 Passenger
Cars and Light Trucks' Final Regulatory Impact Analysis document586.

   One IIHS Top Safety Pick requirement, the Small Overlap, was published in 2012 and came
into full effect with the MY2014. Table 5.153 lists the mass  credit estimates for the IIHS small
overlap which EPA also applied to all 2014 MY vehicles for simplicity of analysis reflecting the
assumption that each vehicle will be redesigned to achieve this goal before 2021MY.  The mass
credit for the IIHS small overlap test on 20 percent lightweight vehicles was determined by two
agency studies for which a good/acceptable rating was the goal. One study was funded by
NHTSA, the updated light weight passenger car study587' and a second by Transport Canada588 as
a follow-up study to EPA's light duty pickup light-weighting study along  with one peer review
comment to the study.  The light weight passenger car credit was found to be a range from 6.3-
9.6kg for subcompact to minivans respectively. The lightweight light duty pickup truck
(aluminum intensive) mass reduction mass increase was determined to be 22kg.

   Each vehicle's MY2014 baseline curb weight is credited with the total  safety mass estimate.
                                            5-397

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                    Table 5.153 Additional Safety Mass Added for 2014 Vehicles
ESTIMATED VEHICLE WEIGHT IMPACT OF FMVSS SAFETY REGULATIONS
and IIHS Small Overlap (kg)
Final Rules by FMVSS No.
214 Side Pole
216 Roof Crush
226 Ejection Mitigation
Final Rules Subtotal
IIHS small overlap
On ~20% lightweight vehicle
Total Mass Increase Est*
Passenger Cars Added
Weight (kg)
5.64
5.28
0.91
11.83kg
6.9kg
18.73kg
Light Trucks Added
Weight (kg)
5.25
5.28
1.07
11.6kg
22kg
33.6kg
Compliance Dates
Sept 2009-2012
Sept 2012-2015
Sept 2013-2017

2012/2014 for Top
Safety

Note: All pass cars and some SUV's fall into the passenger car category.  Some SUV's fall into the light duty truck
classification.  It is also understood that some of the IIHS small overlap mass add may be duplicated in the roof
crush NHTSA design adjustments.


   A reality check on these mass increases for light duty pickup trucks can be seen in the
comparison of the MY2011 Silverado 1500 cabin mass (207.2 kg) compared to the MY2014
Silverado 1500 cabin mass (242.6 kg) which were measured in the EPA and NHTSA respective
light duty pickup truck light-weighting studies.  This is a difference of 35.4kg and is the result of
the addition of a door ring and other improvements which the AHSS components provide.
Although this evaluation is on an AHSS cabin design from a mild steel cabin design, the mass
increase supports the overall mass increase, in Table 5.153, which is based on the optimized
solution for an aluminum truck design. The F150 was redesigned for MY2015, however the
mass increase for the IIHS small overlap was not known since it was incorporated into the
overall vehicle redesign.

   For the passenger car, there are some vehicle designs which currently meet the IIHS small
overlap and are not yet designed to meet the IIHS small overlap.

   3.  Calculate the Resultant Curb Weight Difference between MY2008 and MY2014

   With mass credits for change in footprint and safety determined for the MY2014 vehicles,
then adjusted weight reduction amounts can be calculated as shown in Table 5.154. For
example, the Acura MDX which had a curb weight difference of 11.5  percent is noted to now
have a 13.3 percent difference in curb weight given credits for increased footprint and safety.

   This amount of percent mass reduction will then be applied as a baseline mass reduction for
the particular vehicle and if additional mass reduction technology is to be applied to the vehicle
then the mass reduction cost curve will be recalculated prior to mass reduction technology
application.  In this way, the EPA has attempted to quantify the amount of mass reduction a
manufacturer may have already implemented in the baseline fleet and  the associated cost of
increasing the amount of mass reduction form the baseline.  This will be covered in more detail
further in this section.  The vehicles that were found to have an increase in curb weight or had no
change in curb weight from 2008 to 2014, after being adjusted for footprint and safety mass
increase, had no adjustment to the mass reduction cost curve.
                                             5-398

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
  Table 5.154 Examples of Safety Mass Reduction Allotted and Weight Reduction Change (single vehicle)
Vehicle Make



Acura MDX

GM Cadillac
CIS
Land Rover
Range Rover
Chevy Cruze
(Cobalt 2008)
Footprint
Category/Safety
Category

Large SUV/Truck

Med Car/Pass Car

Large SUV/AI
intensive Truck
Compact/Small

Model
Year


2014

2014

2014

2014

Weight
Reduction based
on change in
curb weight (kg)
238

112

354

98

Weight
Reduction
(%)

11.5%

6.4%

14%

3.0%

Change in
footprint


0.3

0.275

2.63

2.67

Add Mass Savings from
Footprint Increase/ Safety


(238+0.3*10.71+11.6+22)7
2067
(112+0.275*9.12+11.83+6.9)7
1755
(354+2.63*10.71+11.6+22)7
2500
(98+8.43*2.67+11.83+6.9)7
1417
Weight
Reductio
n (%)

13.3%

7.6%

16.6%

9.8%

Note: The numbers in the table are for example only.


5.3.4.6.2.2    MY2014 Vehicles without MY2008 Counterparts

   A review of the MY2014 baseline vehicle models in the MY2008 baseline database reveal
that about half of the models did not have a match in MY2008 by which to determine a mass
reduction change (percent change in curb weight). For these vehicles, a methodology was
determined to create estimates of MY2014 baseline percent mass reduction. The adjustment for
safety as listed in the previous section was applied.

   For each vehicle and respective OEM nameplate group, the percent mass reductions from the
group of OEM nameplate vehicles with MY2008-MY2014 comparisons would be sales weighted
together to  obtain a general mass reduction trend for that nameplate. This average sales
weighted value would then be applied to the new MY2014 vehicles that did not have MY2008
comparisons. It was observed that the majority of vehicles that fall under this category did not
incorporate significant mass reduction and so applying the OEM trend towards mass reduction
was an appropriate approximate.

   Additional work will be to review the applicability of this methodology  as well as apply
baseline percent mass reduction for those vehicles which incorporated lightweight technologies
in MY2008 and MY2014 and as a result may not have the correct percent mass reduction
assigned to them to represent their current  state of technology adoption. Such vehicles include
those which were known to be aluminum intensive or carbon fiber intensive in 2008 and 2014
and the OEM did not have other vehicles on which to determine an appropriate sales weighted
evaluation,  such as Lotus, Tesla, and BMWi3/i8. The majority of these vehicles are low volume
and/or already far exceed the 2025 standards.

5.3.4.6.2.3    MY2014 Cost Curve Adjustments Due to Vehicle Baseline MY2014-MY2008 Curb
Weight Differences

   The NAS committee noted in the 2015 report that "It is generally acknowledged that the cost
to reduce mass increases for each additional unit of mass eliminated on a vehicle."492 EPA
agrees that this is the case, however also notes that in order that the benefits of mass reduction be
achieved, the actual curb weight of the vehicle must actually decrease. As  a result the
calculation  for the MY2014 baseline percent mass reduction (compared to 2008MY) is
                                             5-399

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
calculated through comparison of the curb weight and application of several mass reduction
credits.sss

   The MY2014 vehicles found to be heavier, or the same, than their MY2008 counterparts will
start in the cost curve as-is with zero percent mass reduction. Modifications to the cost curve are
made for those vehicles with resultant curb weight decreases for the MY2014, compared to
MY2008 counterparts,  and hence assumes that mass reduction technologies have been adopted to
achieve reduced curb weight. The removal of mass reduction technology starts with the cost
saving technologies and as a result the remaining points on the cost curve increase from their
original position. While the percent baseline  mass reduction is determined on a vehicle specific
basis (in 0.5%MR increments), the amount of cost curve adjustment  ($/vehicle) used in EPA
modeling is based on a vehicle type basis.  Specifically, each vehicle type (1-19) has a set sales
weighted curb weight for all vehicles within  that type based on the vehicle curb weight and sales
information within the type.

   Figure 5.139 and Figure 5.140 illustrate the change in the EPA passenger car cost curve and
the overall Direct Manufacturing Cost estimates for a MY2014 baseline vehicle (vehicle type 5
(1916kg)) that has 5  percent lighter mass (curb weight plus MR credits) than the MY2008.  The
$/kg results are the same across all vehicle types to which the Car/CUV DMC curve is applied.
The overall $/vehicle vary depending on vehicle type and related curb weight.
                           Passenger Car/CUV DMC Curve Change
                                (for Vehicle Type 5 (1916kg))
                             W/ 5% MR Baseline MY2014 ($/kg)
                                                                          20%
                -6.00
                                         % Mass Reduction
                          •5%MR Baseline 2014MY
•0%MR Baseline 2014MY
  Figure 5.139 Car/CUV DMC ($/kg) Curve for MY2014 Vehicle with 5 Percent Lower Curb Weight Than
                                  MY2008 (Vehicle Type 5)
sss This section has described certain credits given to MY2014 vehicles for increased footprint and safety mass
  increases that lowers the curb weight of the MY2014 vehicle in these calculations.
                                              5-400

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                          Passenger Car/CUV DMC Curve Change
                               (for Vehicle Type 5 (1916kg))
                          W/5% MR Baseline MY2014 ($/vehicle)
                                       % Mass Reduction

                         •5%MR Baseline 2014MY   —•— 0%MR Baseline 2014MY
   Figure 5.140 Total Car/CUV DMC ($/vehicle) Curve for MY2014 Vehicle with 5 Percent Lower Curb
                        Weight Than MY2008 (Vehicle Type 5 of 1916kg)

   Table 5.155 shows the calculations for calculating the new $/kg for adding 5 percent
additional mass reduction on top of a passenger vehicle (EPA vehicle type 5) that already has 5
percent baseline mass reduction.  In Table 5.155, the example is based on a vehicle type 5 with a
sales weighted curb weight of 1916kg (42151b). Results are that the additional 5 percent mass
reduction costs $0.40/kg or an increase of $38.32 (DMC, 2013$ in 2020). As noted previously,
this increase ($/vehicle) is applied across all vehicle type 5 that happen to have 5 percent
baseline mass reduction, regardless of the specific curb weight of the particular vehicle.
   Table 5.155 Example of Calculations for Adjusting Car/CUV DMC Curve for 5 Percent Baseline Mass
                                        Reduction
Vehicle has 5%MR and Applying Additional 10%. Vehicle type 5 curb weight is 1916kg.
$/kg points on Original DMC Curve: 10%= -0.95/kg, 5%= -2.3/kg
Calculation Step
Point of max mass reduction
(ex: 10%)
Point of allotted curb weight mass
reduction (ex: 5%)
Subtract the original (5%) from the
total (10%)
Calculate the new $/kg for the
additional 5%MR
$/vehicle for additional 5%
Mass Reduced (kg)
=%MR*vehicle mass
.10*1916
=192 kg
.05*1916
=95.8 kg
192-95.8=95.8kg
$ Difference
=$/kg*mass reduced
-$0.95*192=
-$182.40
-$2.3/kg *95.8=
-$220.34
(-$182.40)-
(-$220.34)
=$37.94
$37.94/95.8kg=+$0.40/kg
0.40*(1916*0.05)=$38.32
   The EPA modeling does not apply mass reduction to passenger cars with individual curb
weights of 31971bs or below.  The maximum amount of mass reduction is also limited to not
allow a passenger vehicle to go below 31971bs and so this limits a large number of vehicles (in
vehicle types 1-7 and 13) to either 2, 5, or 10 percent maximum mass reduction. The maximum
amount of 20 percent mass reduction is allowed on passenger car vehicle type 5 (large car)
                                             5-401

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
whose sales weighted curb weight is large enough such that the 20 percent mass reduction would
not go below 31971bs.TTT'uuu The light duty truck DMC curve is applied to vehicle types 8-12
and 14-19 (towing vehicles) and the maximum percent mass reduction (20 percent) is allowed to
be applied without any lower bound cutoff.  These vehicle types cover midsize to large SUV's
and all size light duty pickup trucks.

5.3.4.6.2.4    Safety Regulation Mass Increase Estimate Post MY2014

   For the Proposed Determination analyses, a consideration of future potential and final
regulation mass reduction offsets will be considered for the 2022-2025 standards. Table 5.156
shows that a range of 7.08-9.51 kg mass increase from potential/future final rules for passenger
cars and light duty trucks is estimated. Due to the timeline required for NHTSA  to progress from
study to full implementation, it is estimated that the NHTSA oblique test may be incorporated
sometime on or after 2022 so the mass increase is a consideration for the 2022-2025 mass
reduction feasibility.
            Table 5.156 Future Safety Regulation Reference. Mass Increase Expectationswv
POTENTIAL RULES
Pedestrian Protection
Forward Collision Warning (with Dynamic
Brake Support and Crash Imminent Braking),
Lane Departure Warning
Oblique
Part 563 EDR
V2V
Final Potential Rules Subtotal
Final Rule: 111 Rear Cameras
May 2016-2018
TOTAL
Automatic Emergency Braking by 2022/2025
Announced 3/17/2016
Passenger Cars Added Weight (kg)
Min
Max
?
0.29
2.72
5.0
0.04
1.56
6.89
9.32
0.19
7.08
?
9.51
?
Light Trucks Added Weight (kg)
Min
Max
p
0.29
2.72
5.0
0.04
1.56
6.89
9.32
0.15
7.08
?
9.51
?
TTT If there was not a mass cutoff for application of mass reduction then the results from the baseline calculations for
  the MY2008 v MY2014 vehicle data show that approximately 50 percent of the more than 1400 passenger car
  vehicle listing in the modeling, representing 54 percent of the volume within the passenger car modeling, has a
  lighter curb weight in MY2014. Within the 50 percent of passenger car vehicle listings in the modeling, the
  majority are within the 0-5 percent range and a few span the 5-20 percent range.  The remaining vehicles are
  either the same or heavier than the 2008 era design vehicles.
uuu When all passenger car and light duty truck vehicles are weighted together the overall mass change is 0.4
  percent or near neutral. This result is in line with the overall mass pattern within the 2014 Trends report111111
  which shows a near neutral change in regards to vehicle mass for 2014/2015 model years. When the vehicles are
  sales weighted average all together, and those with curb weight increases are set to zero, the overall mass
  reduction decrease is 1.9 percent.
vvv Based on "Estimated Vehicle Weight Impact of Safety Regulations - Potential Rulemakings" (reference: SAE
  Government Industry Conference January 2015). Lane departure warning included in previous table on safety
  mass increase.
                                                 5-402

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The estimate of 5kg mass increase for the potential NHTSA oblique test is increased based on
the estimate that the vehicles currently comply with IIHS small overlap and that there will be a
small additional mass increase due to the uniqueness of the oblique testwww.  It is also
understood that restraint modifications to the seat belt and air bag timing may likely be required.
NHTSA is evaluating this at the time of the writing of the Draft TAR and should have a decision
by the time of the Proposed Determination.

5.3.4.6.3     Effectiveness of Mass Reduction

   In the FRM EPA estimated mass reduction related fuel economy improvement to be 5.1
percent for every  10 percent reduction in mass.  This included application of secondary mass
reduction (which considered downsizing of the engine, brake, transmission, suspension, etc.) at
every percent mass reduction.xxx  This methodology recognizes that a manufacturer does not
have a single threshold which results in right-sizing the engine, but rather designs the  vehicle as
a system.

   For the Draft TAR, EPA performed effectiveness  analyses for the standard car class using the
ALPHA model and engine maps representing MY2014 and newer engines. Results showed the
effectiveness for mass reduction is a linear equation based on the engine baseline out CCh
emissions.  As a result an effectiveness of 5.2 percent is utilized for both cars and trucks. For
Discussion of the  Alpha model see Section 5.3.3.2.2.

5.3.4.6.4     Mass Reduction Costs used in OMEGA

   The tables below show an excerpt of the mass reduction costs used in OMEGA. There are 8
tables that follow, with the first four showing mass reduction costs at 5 percent, then 10 percent,
then 15 percent then 20 percent mass reduction for the 8 vehicle types that use the car cost curve.
The next four tables show mass reduction costs at 5 percent, then 10 percent, then 15 percent
then 20 percent mass reduction for the 11 vehicle types that use the truck cost curve. The direct
manufacturing costs (DMC), indirect costs (1C, using ICMs) and the total costs (TC) are shown
along with the sales weighted average curb weight of all vehicles mapped into the indicated
vehicle types, the  complexity levels used for indirect costs and the learning curve factor used as
discussed in Section 5.3.2.

   The cost for additional mass reduction increases with increasing MY2014 baseline mass
reduction. The MY2014 baseline percent mass reduction is determined for each vehicle model
(sales weighted for trim with adjustments for AWD/2WD, adjustments for safety and  footprint
changes) and noted on a 0.5 percent mass reduction increment basis. Since the cost curves are
developed with the greatest cost save/kg mass reduction ideas listed first, which are then
cumulatively added, the calculations for removing the baseline mass reduction percentage is
performed beginning with the lowest cost save portion of the curve.  As a result the additional
mass reduction technology costs increase with increasing MY2014 baseline mass reduction.
    The mass increase for the IIHS small overlap crash test was accounted for in the MY2014 baseline curb
  weight.
xxx This is assumed to be the outcome in 2025 and not necessarily in the transition years. EPA has observed that in
  2016 some OEM's have engine models with 0.1L or 0.2L difference between them and so OEM's are able to be
  successful in their engine downsize-vehicle matching.


                                              5-403

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.157 Costs for 5 Percent Mass Reduction for Non-towing (Car curve) Vehicle Types (2013$)
Vehicle
Type
1
2
3
4
5
6
7
13
1
2
3
4
5
6
7
13
1
2
3
4
5
6
7
13
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
DMC:
CurbWt
1C:
complexity
2629
3131
3557
3495
4215
3967
3494
3767
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2








DMC:
learning
curve
1C: near
term
thru
30
30
30
30
30
30
30
30
2024
2024
2024
2024
2024
2024
2024
2024








2017
-$150
-$178
-$202
-$199
-$240
-$226
-$199
-$214
$26
$31
$35
$35
$42
$39
$35
$37
-$124
-$147
-$167
-$164
-$198
-$186
-$164
-$177
2018
-$145
-$172
-$196
-$193
-$232
-$219
-$193
-$208
$26
$31
$35
$35
$42
$39
$35
$37
-$119
-$141
-$161
-$158
-$190
-$179
-$158
-$170
2019
-$141
-$168
-$190
-$187
-$226
-$212
-$187
-$202
$26
$31
$35
$35
$42
$39
$35
$37
-$115
-$136
-$155
-$152
-$184
-$173
-$152
-$164
2020
-$137
-$163
-$186
-$182
-$220
-$207
-$182
-$196
$26
$31
$35
$35
$42
$39
$35
$37
-$111
-$132
-$150
-$148
-$178
-$168
-$148
-$159
2021
-$134
-$160
-$181
-$178
-$215
-$202
-$178
-$192
$26
$31
$35
$35
$42
$39
$35
$37
-$108
-$128
-$146
-$143
-$173
-$163
-$143
-$155
2022
-$131
-$156
-$177
-$174
-$210
-$198
-$174
-$188
$26
$31
$35
$35
$42
$39
$35
$37
-$105
-$125
-$142
-$140
-$168
-$159
-$140
-$151
2023
-$129
-$153
-$174
-$171
-$206
-$194
-$171
-$184
$26
$31
$35
$35
$42
$39
$35
$37
-$102
-$122
-$139
-$136
-$164
-$155
-$136
-$147
2024
-$126
-$150
-$171
-$168
-$202
-$191
-$168
-$181
$26
$31
$35
$35
$42
$39
$35
$37
-$100
-$119
-$136
-$133
-$161
-$151
-$133
-$144
2025
-$124
-$148
-$168
-$165
-$199
-$187
-$165
-$178
$21
$25
$28
$28
$34
$32
$28
$30
-$103
-$123
-$139
-$137
-$165
-$156
-$137
-$148
   Note: DMC=direct manufacturing cost; IC=indirect cost; TC=total cost.
                                           5-404

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.158 Costs for 10 Percent Mass Reduction for Non-towing (Car curve) Vehicle Types (2013$)
Vehicle
Type
1
2
3
4
5
6
7
13
1
2
3
4
5
6
7
13
1
2
3
4
5
6
7
13
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
DMC:
CurbWt
1C:
complexity
2629
3131
3557
3495
4215
3967
3494
3767
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2








DMC:
learning
curve
1C: near
term
thru
30
30
30
30
30
30
30
30
2024
2024
2024
2024
2024
2024
2024
2024








2017
-$123
-$147
-$167
-$164
-$198
-$186
-$164
-$177
$104
$124
$141
$139
$167
$158
$139
$150
-$19
-$23
-$26
-$25
-$31
-$29
-$25
-$27
2018
-$119
-$142
-$162
-$159
-$192
-$180
-$159
-$171
$104
$124
$141
$139
$167
$158
$139
$150
-$15
-$18
-$20
-$20
-$24
-$23
-$20
-$22
2019
-$116
-$138
-$157
-$154
-$186
-$175
-$154
-$166
$104
$124
$141
$139
$167
$158
$139
$150
-$12
-$14
-$16
-$16
-$19
-$18
-$16
-$17
2020
-$113
-$135
-$153
-$150
-$181
-$171
-$150
-$162
$104
$124
$141
$139
$167
$158
$139
$150
-$9
-$10
-$12
-$12
-$14
-$13
-$12
-$13
2021
-$111
-$132
-$150
-$147
-$177
-$167
-$147
-$158
$104
$124
$141
$139
$167
$158
$139
$150
-$6
-$7
-$8
-$8
-$10
-$9
-$8
-$9
2022
-$108
-$129
-$146
-$144
-$173
-$163
-$144
-$155
$104
$124
$141
$139
$167
$158
$139
$150
-$4
-$5
-$5
-$5
-$6
-$6
-$5
-$5
2023
-$106
-$126
-$144
-$141
-$170
-$160
-$141
-$152
$104
$124
$141
$139
$167
$158
$139
$150
-$2
-$2
-$2
-$2
-$3
-$3
-$2
-$2
2024
-$104
-$124
-$141
-$139
-$167
-$157
-$138
-$149
$104
$124
$141
$139
$167
$158
$139
$150
$0
$0
$0
$0
$0
$0
$0
$0
2025
-$102
-$122
-$139
-$136
-$164
-$155
-$136
-$147
$84
$100
$114
$112
$135
$127
$112
$121
-$18
-$22
-$25
-$24
-$29
-$27
-$24
-$26
   Note: DMC=direct manufacturing cost; IC=indirect cost; TC=total cost.
                                           5-405

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.159 Costs for 15 Percent Mass Reduction for Non-towing (Car curve) Vehicle Types (2013$)
Vehicle
Type
1
2
3
4
5
6
7
13
1
2
3
4
5
6
7
13
1
2
3
4
5
6
7
13
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
DMC:
CurbWt
1C:
complexity
2629
3131
3557
3495
4215
3967
3494
3767
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2








DMC:
learning
curve
1C: near
term
thru
30
30
30
30
30
30
30
30
2024
2024
2024
2024
2024
2024
2024
2024








2017
-$31
-$37
-$42
-$41
-$50
-$47
-$41
-$44
$235
$280
$318
$312
$377
$354
$312
$337
$204
$243
$276
$271
$327
$308
$271
$292
2018
-$30
-$36
-$40
-$40
-$48
-$45
-$40
-$43
$235
$280
$318
$312
$377
$354
$312
$337
$205
$244
$277
$273
$329
$309
$272
$294
2019
-$29
-$35
-$39
-$39
-$47
-$44
-$39
-$42
$235
$280
$318
$312
$377
$354
$312
$337
$206
$245
$278
$274
$330
$311
$274
$295
2020
-$28
-$34
-$38
-$38
-$45
-$43
-$38
-$41
$235
$280
$318
$312
$377
$354
$312
$337
$207
$246
$279
$275
$331
$312
$275
$296
2021
-$28
-$33
-$37
-$37
-$44
-$42
-$37
-$40
$235
$280
$318
$312
$377
$354
$312
$337
$207
$247
$280
$276
$332
$313
$275
$297
2022
-$27
-$32
-$37
-$36
-$43
-$41
-$36
-$39
$235
$280
$318
$312
$377
$354
$312
$337
$208
$247
$281
$276
$333
$314
$276
$298
2023
-$27
-$32
-$36
-$35
-$43
-$40
-$35
-$38
$235
$280
$318
$312
$377
$354
$312
$337
$208
$248
$282
$277
$334
$314
$277
$299
2024
-$26
-$31
-$35
-$35
-$42
-$39
-$35
-$37
$235
$280
$318
$312
$377
$354
$312
$337
$209
$249
$283
$278
$335
$315
$278
$299
2025
-$26
-$31
-$35
-$34
-$41
-$39
-$34
-$37
$189
$226
$256
$252
$304
$286
$252
$271
$164
$195
$222
$218
$263
$247
$218
$235
          Note: DMC=direct manufacturing cost; IC=indirect cost; TC=total cost.
                                           5-406

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.160 Costs for 20 Percent Mass Reduction for Non-towing (Car curve) Vehicle Types (2013$)
Vehicle
Type
1
2
3
4
5
6
7
13
1
2
3
4
5
6
7
13
1
2
3
4
5
6
7
13
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
DMC:
CurbWt
1C:
complexity
2629
3131
3557
3495
4215
3967
3494
3767
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2








DMC:
learning
curve
1C: near
term
thru
30
30
30
30
30
30
30
30
2024
2024
2024
2024
2024
2024
2024
2024








2017
$105
$125
$142
$139
$168
$158
$139
$150
$418
$497
$565
$555
$670
$630
$555
$598
$522
$622
$707
$694
$838
$788
$694
$748
2018
$101
$121
$137
$135
$163
$153
$135
$145
$418
$497
$565
$555
$670
$630
$555
$598
$519
$618
$702
$690
$832
$783
$690
$744
2019
$99
$117
$133
$131
$158
$149
$131
$141
$418
$497
$565
$555
$670
$630
$555
$598
$516
$615
$698
$686
$827
$779
$686
$740
2020
$96
$114
$130
$128
$154
$145
$128
$138
$418
$497
$565
$555
$670
$630
$555
$598
$514
$612
$695
$683
$823
$775
$683
$736
2021
$94
$112
$127
$125
$150
$142
$125
$134
$418
$497
$565
$555
$670
$630
$555
$598
$511
$609
$692
$680
$820
$772
$680
$733
2022
$92
$109
$124
$122
$147
$139
$122
$132
$418
$497
$565
$555
$670
$630
$555
$598
$509
$607
$689
$677
$817
$769
$677
$730
2023
$90
$107
$122
$120
$144
$136
$120
$129
$418
$497
$565
$555
$670
$630
$555
$598
$508
$604
$687
$675
$814
$766
$675
$727
2024
$88
$105
$120
$118
$142
$133
$118
$127
$418
$497
$565
$555
$670
$630
$555
$598
$506
$603
$685
$673
$811
$764
$673
$725
2025
$87
$104
$118
$116
$139
$131
$116
$125
$337
$401
$456
$448
$540
$508
$448
$483
$424
$505
$573
$563
$679
$639
$563
$607
   Note: DMC=direct manufacturing cost; IC=indirect cost; TC=total cost.
                                           5-407

-------
                           Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.161 Costs for 5 Percent Mass Reduction for Towing (Truck curve) Vehicle Types (2013$)
Vehicle
Type
8
9
10
11
12
14
15
16
17
18
19
8
9
10
11
12
14
15
16
17
18
19
8
9
10
11
12
14
15
16
17
18
19
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC:
CurbWt
1C:
complexity
4306
4272
4918
5158
5518
4575
4848
5507
6071
5975
5145
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2











DMC:
learning
curve
1C: near
term
thru
30
30
30
30
30
30
30
30
30
30
30
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024











2017
-$225
-$223
-$257
-$269
-$288
-$239
-$253
-$288
-$317
-$312
-$269
$65
$64
$74
$77
$83
$69
$73
$83
$91
$90
$77
-$160
-$159
-$183
-$192
-$205
-$170
-$180
-$205
-$226
-$222
-$192
2018
-$218
-$216
-$249
-$261
-$279
-$231
-$245
-$278
-$307
-$302
-$260
$65
$64
$74
$77
$83
$69
$73
$83
$91
$90
$77
-$153
-$152
-$175
-$183
-$196
-$163
-$172
-$196
-$216
-$212
-$183
2019
-$211
-$210
-$242
-$253
-$271
-$225
-$238
-$270
-$298
-$293
-$253
$65
$64
$74
$77
$83
$69
$73
$83
$91
$90
$77
-$147
-$146
-$168
-$176
-$188
-$156
-$165
-$188
-$207
-$204
-$175
2020
-$206
-$204
-$235
-$247
-$264
-$219
-$232
-$264
-$291
-$286
-$246
$65
$64
$74
$77
$83
$69
$73
$83
$91
$90
$77
-$141
-$140
-$162
-$169
-$181
-$150
-$159
-$181
-$199
-$196
-$169
2021
-$201
-$200
-$230
-$241
-$258
-$214
-$227
-$257
-$284
-$279
-$241
$65
$64
$74
$77
$83
$69
$73
$83
$91
$90
$77
-$137
-$136
-$156
-$164
-$175
-$145
-$154
-$175
-$193
-$190
-$163
2022
-$197
-$196
-$225
-$236
-$253
-$209
-$222
-$252
-$278
-$273
-$235
$65
$64
$74
$77
$83
$69
$73
$83
$91
$90
$77
-$132
-$131
-$151
-$159
-$170
-$141
-$149
-$169
-$187
-$184
-$158
2023
-$193
-$192
-$221
-$231
-$248
-$205
-$218
-$247
-$272
-$268
-$231
$65
$64
$74
$77
$83
$69
$73
$83
$91
$90
$77
-$129
-$128
-$147
-$154
-$165
-$137
-$145
-$165
-$181
-$178
-$154
2024
-$190
-$188
-$217
-$227
-$243
-$202
-$214
-$243
-$268
-$263
-$227
$65
$64
$74
$77
$83
$69
$73
$83
$91
$90
$77
-$125
-$124
-$143
-$150
-$160
-$133
-$141
-$160
-$176
-$174
-$150
2025
-$187
-$185
-$213
-$224
-$239
-$198
-$210
-$239
-$263
-$259
-$223
$52
$52
$60
$63
$67
$56
$59
$67
$74
$73
$63
-$134
-$133
-$153
-$161
-$172
-$143
-$151
-$172
-$189
-$186
-$160
  Note: DMC=direct manufacturing cost; IC=indirect cost; TC=total cost.
                                          5-408

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.162 Costs for 10 Percent Mass Reduction for Towing (Truck curve) Vehicle Types (2013$)
Vehicle
Type
8
9
10
11
12
14
15
16
17
18
19
8
9
10
11
12
14
15
16
17
18
19
8
9
10
11
12
14
15
16
17
18
19
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC:
CurbWt
1C:
complexity
4306
4272
4918
5158
5518
4575
4848
5507
6071
5975
5145
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2
Low2











DMC:
learning
curve
1C: near
term
thru
30
30
30
30
30
30
30
30
30
30
30
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024











2017
$66
$65
$75
$79
$84
$70
$74
$84
$93
$91
$78
$258
$256
$295
$310
$331
$275
$291
$331
$364
$359
$309
$324
$322
$370
$388
$415
$344
$365
$414
$457
$450
$387
2018
$64
$63
$73
$76
$81
$67
$72
$81
$90
$88
$76
$258
$256
$295
$310
$331
$275
$291
$331
$364
$359
$309
$322
$319
$368
$386
$413
$342
$363
$412
$454
$447
$385
2019
$62
$61
$70
$74
$79
$66
$69
$79
$87
$86
$74
$258
$256
$295
$310
$331
$275
$291
$331
$364
$359
$309
$320
$318
$366
$383
$410
$340
$360
$409
$451
$444
$383
2020
$60
$60
$69
$72
$77
$64
$68
$77
$85
$83
$72
$258
$256
$295
$310
$331
$275
$291
$331
$364
$359
$309
$319
$316
$364
$382
$408
$338
$359
$407
$449
$442
$381
2021
$59
$58
$67
$70
$75
$62
$66
$75
$83
$82
$70
$258
$256
$295
$310
$331
$275
$291
$331
$364
$359
$309
$317
$315
$362
$380
$406
$337
$357
$406
$447
$440
$379
2022
$58
$57
$66
$69
$74
$61
$65
$74
$81
$80
$69
$258
$256
$295
$310
$331
$275
$291
$331
$364
$359
$309
$316
$313
$361
$378
$405
$336
$356
$404
$445
$438
$377
2023
$56
$56
$64
$68
$72
$60
$64
$72
$80
$78
$67
$258
$256
$295
$310
$331
$275
$291
$331
$364
$359
$309
$315
$312
$360
$377
$403
$334
$355
$403
$444
$437
$376
2024
$55
$55
$63
$66
$71
$59
$62
$71
$78
$77
$66
$258
$256
$295
$310
$331
$275
$291
$331
$364
$359
$309
$314
$311
$358
$376
$402
$333
$353
$401
$442
$435
$375
2025
$54
$54
$62
$65
$70
$58
$61
$70
$77
$76
$65
$210
$208
$239
$251
$269
$223
$236
$268
$296
$291
$250
$264
$262
$302
$316
$338
$281
$297
$338
$372
$366
$316
         Note: DMC=direct manufacturing cost; IC=indirect cost; TC=total cost.
                                           5-409

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.163 Costs for 15 Percent Mass Reduction for Towing (Truck curve) Vehicle Types (2013$)
Vehicle
Type
8
9
10
11
12
14
15
16
17
18
19
8
9
10
11
12
14
15
16
17
18
19
8
9
10
11
12
14
15
16
17
18
19
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC:
CurbWt
1C:
complexity
4306
4272
4918
5158
5518
4575
4848
5507
6071
5975
5145
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2











DMC:
learning
curve
1C: near
term
thru
30
30
30
30
30
30
30
30
30
30
30
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024











2017
$551
$546
$629
$660
$706
$585
$620
$704
$777
$764
$658
$582
$577
$664
$696
$745
$618
$655
$744
$820
$807
$695
$1,132
$1,123
$1,293
$1,356
$1,451
$1,203
$1,275
$1,448
$1,596
$1,571
$1,353
2018
$533
$529
$609
$638
$683
$566
$600
$682
$752
$740
$637
$582
$577
$664
$696
$745
$618
$655
$744
$820
$807
$695
$1,115
$1,106
$1,273
$1,335
$1,428
$1,184
$1,255
$1,425
$1,571
$1,546
$1,332
2019
$518
$514
$591
$620
$664
$550
$583
$662
$730
$719
$619
$582
$577
$664
$696
$745
$618
$655
$744
$820
$807
$695
$1,099
$1,091
$1,256
$1,317
$1,409
$1,168
$1,238
$1,406
$1,550
$1,525
$1,313
2020
$505
$501
$576
$604
$647
$536
$568
$645
$712
$700
$603
$582
$577
$664
$696
$745
$618
$655
$744
$820
$807
$695
$1,086
$1,078
$1,241
$1,301
$1,392
$1,154
$1,223
$1,389
$1,531
$1,507
$1,298
2021
$493
$489
$563
$591
$632
$524
$555
$630
$695
$684
$589
$582
$577
$664
$696
$745
$618
$655
$744
$820
$807
$695
$1,075
$1,066
$1,227
$1,287
$1,377
$1,142
$1,210
$1,374
$1,515
$1,491
$1,284
2022
$483
$479
$551
$578
$618
$513
$543
$617
$680
$670
$577
$582
$577
$664
$696
$745
$618
$655
$744
$820
$807
$695
$1,064
$1,056
$1,215
$1,275
$1,364
$1,130
$1,198
$1,361
$1,500
$1,476
$1,271
2023
$473
$469
$541
$567
$606
$503
$533
$605
$667
$657
$565
$582
$577
$664
$696
$745
$618
$655
$744
$820
$807
$695
$1,055
$1,046
$1,205
$1,263
$1,352
$1,121
$1,188
$1,349
$1,487
$1,463
$1,260
2024
$465
$461
$531
$557
$596
$494
$523
$594
$655
$645
$555
$582
$577
$664
$696
$745
$618
$655
$744
$820
$807
$695
$1,046
$1,038
$1,195
$1,253
$1,341
$1,111
$1,178
$1,338
$1,475
$1,452
$1,250
2025
$457
$453
$522
$547
$586
$485
$514
$584
$644
$634
$546
$472
$468
$539
$565
$604
$501
$531
$603
$665
$655
$564
$929
$921
$1,061
$1,112
$1,190
$987
$1,046
$1,188
$1,309
$1,289
$1,110
  Note: DMC=direct manufacturing cost; IC=indirect cost; TC=total cost.
                                           5-410

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
     Table 5.164 Costs for 20 Percent Mass Reduction for Towing (Truck curve) Vehicle Types (2013$)
Vehicle
Type
8
9
10
11
12
14
15
16
17
18
19
8
9
10
11
12
14
15
16
17
18
19
8
9
10
11
12
14
15
16
17
18
19
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
DMC:
CurbWt
1C:
complexity
4306
4272
4918
5158
5518
4575
4848
5507
6071
5975
5145
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2
Med2











DMC:
learning
curve
1C: near
term
thru
30
30
30
30
30
30
30
30
30
30
30
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024











2017
$1,162
$1,153
$1,327
$1,392
$1,489
$1,235
$1,309
$1,486
$1,639
$1,613
$1,389
$1,034
$1,026
$1,181
$1,238
$1,325
$1,098
$1,164
$1,322
$1,458
$1,434
$1,235
$2,196
$2,179
$2,508
$2,630
$2,814
$2,333
$2,473
$2,808
$3,096
$3,047
$2,624
2018
$1,125
$1,116
$1,285
$1,347
$1,441
$1,195
$1,266
$1,438
$1,586
$1,561
$1,344
$1,034
$1,026
$1,181
$1,238
$1,325
$1,098
$1,164
$1,322
$1,458
$1,434
$1,235
$2,159
$2,141
$2,465
$2,585
$2,766
$2,293
$2,430
$2,760
$3,043
$2,995
$2,579
2019
$1,093
$1,084
$1,248
$1,309
$1,400
$1,161
$1,230
$1,397
$1,541
$1,516
$1,306
$1,034
$1,026
$1,181
$1,238
$1,325
$1,098
$1,164
$1,322
$1,458
$1,434
$1,235
$2,127
$2,110
$2,429
$2,547
$2,725
$2,259
$2,394
$2,720
$2,998
$2,951
$2,541
2020
$1,065
$1,057
$1,216
$1,276
$1,365
$1,131
$1,199
$1,362
$1,502
$1,478
$1,272
$1,034
$1,026
$1,181
$1,238
$1,325
$1,098
$1,164
$1,322
$1,458
$1,434
$1,235
$2,099
$2,082
$2,397
$2,514
$2,689
$2,230
$2,363
$2,684
$2,959
$2,912
$2,507
2021
$1,040
$1,032
$1,188
$1,246
$1,333
$1,105
$1,171
$1,330
$1,467
$1,444
$1,243
$1,034
$1,026
$1,181
$1,238
$1,325
$1,098
$1,164
$1,322
$1,458
$1,434
$1,235
$2,074
$2,058
$2,369
$2,484
$2,658
$2,203
$2,335
$2,652
$2,924
$2,878
$2,478
2022
$1,018
$1,010
$1,163
$1,220
$1,305
$1,082
$1,147
$1,302
$1,436
$1,413
$1,217
$1,034
$1,026
$1,181
$1,238
$1,325
$1,098
$1,164
$1,322
$1,458
$1,434
$1,235
$2,052
$2,036
$2,344
$2,458
$2,630
$2,180
$2,311
$2,624
$2,894
$2,847
$2,452
2023
$999
$991
$1,141
$1,196
$1,280
$1,061
$1,124
$1,277
$1,408
$1,386
$1,193
$1,034
$1,026
$1,181
$1,238
$1,325
$1,098
$1,164
$1,322
$1,458
$1,434
$1,235
$2,032
$2,016
$2,321
$2,434
$2,604
$2,159
$2,288
$2,599
$2,866
$2,820
$2,428
2024
$981
$973
$1,120
$1,175
$1,257
$1,042
$1,104
$1,254
$1,383
$1,361
$1,172
$1,034
$1,026
$1,181
$1,238
$1,325
$1,098
$1,164
$1,322
$1,458
$1,434
$1,235
$2,015
$1,998
$2,301
$2,413
$2,581
$2,140
$2,268
$2,576
$2,840
$2,795
$2,407
2025
$964
$957
$1,101
$1,155
$1,236
$1,024
$1,086
$1,233
$1,360
$1,338
$1,152
$839
$832
$958
$1,004
$1,075
$891
$944
$1,072
$1,182
$1,164
$1,002
$1,803
$1,789
$2,059
$2,159
$2,310
$1,915
$2,030
$2,306
$2,542
$2,501
$2,154
       Note: DMC=direct manufacturing cost; IC=indirect cost; TC=total cost.
5.3.4.7 Other Vehicle Technologies
5.3.4.7.1
Electrified Power Steering: Data and Assumptions for this Assessment
   For the 2017-2025 final rule, EPA and NHTSA estimated a 1 to 2 percent effectiveness for
electrified power steering in light duty vehicles, based on the 2015 NAS report, Sierra Research
Report and confidential OEM data. The 2010 Ricardo study also confirmed this estimate.
NHTSA and EPA reviewed these effectiveness estimates and found them to be accurate, thus
they have been retained for this Draft TAR.

   Costs associated with electric power steering are equivalent to those used in the 2012 FRM
except for updates to 2013 dollars and use of new learning curves (curve 24). The electric power
steering costs incremental to hydraulic power steering are shown below.
                                             5-411

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                Table 5.165 Costs for Electric Power Steering (dollar values in 2013$)
Cost type
DMC
1C
TC
DMC: base year cost
1C: complexity
$96
Low2

DMC: learning curve
1C: near term thru
24
2018

2017
$92
$23
$115
2018
$90
$23
$113
2019
$88
$18
$106
2020
$87
$18
$105
2021
$85
$18
$104
2022
$84
$18
$102
2023
$83
$18
$101
2024
$82
$18
$100
2025
$81
$18
$99
              Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.7.2      Improved Accessories: Data and Assumptions for this Assessment

   In MYs 2017-2025 final rule, the agencies used an effectiveness value in the range of 1 to 2
percent.

   For this Draft TAR GHG assessment, EPA considered two levels of improved accessories.
Level 1 of this technology (IACC1) incorporates a high efficiency alternator (70 percent
efficiency).  The second level of improved accessories (IACC2) adds the higher efficiency
alternator and incorporates a mild regenerative alternator strategy, as well as intelligent cooling.
NHTSA and EPA jointly reviewed the estimates of 1 to 2 percent effectiveness estimates used in
the 2017-2025 final rule for level IACC1. EPA used effectiveness values in the 1.2 to 1.8 percent
range, varying with vehicle subclass.

   Costs associated with improved accessories are equivalent to those used in the 2012 FRM
except for updates to 2013 dollars and use of new learning curves (curve 24). The improved
accessory costs (levels  1 and 2) are shown below. Cost is higher for improved accessories level 2
due to the inclusion of a higher efficiency alternator and a mild level of regeneration, hence the
$40 to  $50 higher cost. Both improved accessory costs are incremental to the baseline.
              Table 5.166 Costs for Improved Accessories Level 1 (dollar values in 2013$)
Cost type
DMC
1C
TC
DMC: base year cost
1C: complexity
$78
Low2

DMC: learning curve
1C: near term thru
24
2018

2017
$74
$19
$93
2018
$73
$19
$92
2019
$72
$15
$87
2020
$70
$15
$85
2021
$69
$15
$84
2022
$68
$15
$83
2023
$67
$15
$82
2024
$66
$15
$81
2025
$66
$15
$81
              Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
              Table 5.167 Costs for Improved Accessories Level 2 (dollar values in 2013$)
Cost type
DMC
1C
TC
DMC: base year cost
1C: complexity
$126
Low2

DMC: learning curve
1C: near term thru
24
2018

2017
$120
$30
$151
2018
$118
$30
$148
2019
$116
$24
$140
2020
$114
$24
$138
2021
$112
$24
$136
2022
$111
$24
$135
2023
$109
$24
$133
2024
$108
$24
$132
2025
$106
$24
$130
5.3.4.7.3
Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
Secondary Axle Disconnect: Data and Assumptions for this Assessment
   The 2017-2025 final rule estimated an effectiveness improvement of 1.0 to 1.5 percent for
axle disconnect. Based on the 2011 Ricardo report, NHTSA and EPA refined this range to 1.2 to
1.4 percent. EPA retains these figures for this Draft TAR GHG assessment.

   The cost associated with secondary axle disconnect is equivalent to that used in the 2012
FRM except for updates to 2013 dollars and use of new learning curves (curve 24). The costs are
shown below.
                                              5-412

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
               Table 5.168 Costs for Secondary Axle Disconnect (dollar values in 2013$)
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$85
Low2

DMC: learning
curve
1C: near term thru
24
2018

2017


$82
$21
$102
2018


$80
$21
$101
2019


$79
$16
$95
2020


$77
$16
$94
2021


$76
$16
$93
2022


$75
$16
$91
2023


$74
$16
$90
2024


$73
$16
$89
2025


$72
$16
$88
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.7.4     Low Drag Brakes: Data and Assumptions for this Assessment

   The 2017-2025 final rule estimated the effectiveness of low drag brakes to be to 0.8 percent.
The agencies continue to use this estimate for this Draft TAR based on the 2011 Ricardo study
and the 2015 NAS report.

   The cost associated with low drag brakes is equivalent to that used in the 2012 FRM except
for updates to 2013 dollars. The costs are shown below.
                  Table 5.169 Costs for Low Drag Brakes (dollar values in 2013$)
Cost
type

DMC
1C
TC
DMC: base year
cost
1C: complexity
$62
Low2

DMC: learning
curve
1C: near term thru
1
2018

2017


$62
$15
$77
2018


$62
$15
$77
2019


$62
$12
$74
2020


$62
$12
$74
2021


$62
$12
$74
2022


$62
$12
$74
2023


$62
$12
$74
2024


$62
$12
$74
2025


$62
$12
$74
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.8 Air Conditioning: Data and Assumptions for this Assessment

   Air conditioning (A/C) system technologies include improved hoses, connectors and seals for
leakage control. They also include improved compressors, expansion valves, heat exchangers
and the control of these components for the purposes of improving tailpipe CCh emissions and
fuel economy as a result of A/C use.

   For this Draft TAR analysis, EPA is continuing to use the GHG and fuel economy
effectiveness estimates that were used in the 2012 FRM analysis, with costs adjusted to 2013
dollars (presented below). For more information on these estimates, see Section 5.1 of the 2012
TSD.
                    Table 5.170 Costs for A/C Controls  (dollar values in 2013$)
Cost type
TC
2017
$91
2018
$117
2019
$134
2020
$141
2021
$154
2022
$152
2023
$146
2024
$143
2025
$140
                    Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
5.3.4.9 Cost Tables for Individual Technologies Not Presented Above

   Costs associated with SCR-equipped diesel vehicles are equivalent to those used in the FRM
except for updates to 2013 dollars and use of new learning curve (curve 23). The costs
incremental to the baseline engine configuration for our different vehicle classes are shown
below. These costs are used to characterize technology costs in the baseline fleet; EPA does not
build OMEGA packages using this technology and instead uses the advanced diesel technology
presented below.
                                             5-413

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.171 Costs for SCR-equipped Diesel Technology for Different Vehicle Classes (dollar values in 2013$)
Tech
Small car
Standard
car
Large car
Small MPV
Large MPV
Truck
Small car
Standard
car
Large car
Small MPV
Large MPV
Truck
Small car
Standard
car
Large car
Small MPV
Large MPV
Truck
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
DMC: base
cost
1C: complexity
$2,456
$2,456
$3,019
$2,483
$2,483
$3,462
Med2
Med2
Med2
Med2
Med2
Med2






DMC: learning
curve
1C: near term
thru
23
23
23
23
23
23
2018
2018
2018
2018
2018
2018






2017
$2,223
$2,223
$2,734
$2,248
$2,248
$3,135
$941
$941
$1,156
$951
$951
$1,326
$3,164
$3,164
$3,890
$3,199
$3,199
$4,461
2018
$2,188
$2,188
$2,691
$2,213
$2,213
$3,086
$939
$939
$1,155
$949
$949
$1,324
$3,127
$3,127
$3,846
$3,162
$3,162
$4,410
2019
$2,156
$2,156
$2,651
$2,180
$2,180
$3,040
$702
$702
$863
$710
$710
$990
$2,858
$2,858
$3,515
$2,890
$2,890
$4,030
2020
$2,126
$2,126
$2,614
$2,150
$2,150
$2,998
$701
$701
$862
$709
$709
$989
$2,827
$2,827
$3,477
$2,858
$2,858
$3,986
2021
$2,098
$2,098
$2,580
$2,121
$2,121
$2,958
$700
$700
$861
$708
$708
$987
$2,799
$2,799
$3,441
$2,829
$2,829
$3,946
2022
$2,072
$2,072
$2,548
$2,095
$2,095
$2,921
$699
$699
$860
$707
$707
$986
$2,772
$2,772
$3,408
$2,802
$2,802
$3,908
2023
$2,047
$2,047
$2,517
$2,070
$2,070
$2,887
$699
$699
$859
$706
$706
$985
$2,746
$2,746
$3,377
$2,776
$2,776
$3,872
2024
$2,024
$2,024
$2,489
$2,047
$2,047
$2,854
$698
$698
$858
$706
$706
$984
$2,722
$2,722
$3,347
$2,752
$2,752
$3,838
2025
$2,002
$2,002
$2,462
$2,024
$2,024
$2,823
$697
$697
$857
$705
$705
$983
$2,700
$2,700
$3,319
$2,729
$2,729
$3,806
   Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
   Costs associated with advanced diesel vehicles (i.e., Tier 3 compliant) are equivalent to those
used in the FRM except for updates to 2013 dollars and use of new learning curve (curve 23).
The costs incremental to the baseline engine configuration for our different vehicle classes are
shown below. These costs are used when building OMEGA diesel packages.
  Table 5.172 Costs for Advanced Diesel Technology for Different Vehicle Classes (dollar values in 2013$)
Tech
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
DMC: base
cost
1C:
complexity
$2,506
$2,506
$3,069
$2,533
$2,533
$3,512
Med2
Med2
Med2
Med2
Med2
Med2






DMC:
learning
curve
1C: near
term
thru
23
23
23
23
23
23
2018
2018
2018
2018
2018
2018






2017
$2,269
$2,269
$2,779
$2,293
$2,293
$3,180
$960
$960
$1,176
$970
$970
$1,345
$3,228
$3,228
$3,955
$3,263
$3,263
$4,525
2018
$2,233
$2,233
$2,735
$2,257
$2,257
$3,130
$958
$958
$1,174
$968
$968
$1,343
$3,191
$3,191
$3,909
$3,226
$3,226
$4,473
2019
$2,200
$2,200
$2,695
$2,224
$2,224
$3,084
$716
$716
$878
$724
$724
$1,004
$2,916
$2,916
$3,573
$2,948
$2,948
$4,088
2020
$2,170
$2,170
$2,658
$2,193
$2,193
$3,041
$715
$715
$876
$723
$723
$1,003
$2,885
$2,885
$3,534
$2,916
$2,916
$4,044
2021
$2,141
$2,141
$2,623
$2,164
$2,164
$3,001
$715
$715
$875
$722
$722
$1,002
$2,856
$2,856
$3,498
$2,886
$2,886
$4,003
2022
$2,114
$2,114
$2,590
$2,137
$2,137
$2,964
$714
$714
$874
$721
$721
$1,000
$2,828
$2,828
$3,464
$2,858
$2,858
$3,964
2023
$2,089
$2,089
$2,559
$2,112
$2,112
$2,928
$713
$713
$873
$721
$721
$999
$2,802
$2,802
$3,433
$2,832
$2,832
$3,928
2024
$2,065
$2,065
$2,530
$2,088
$2,088
$2,895
$712
$712
$872
$720
$720
$998
$2,778
$2,778
$3,403
$2,808
$2,808
$3,894
2025
$2,043
$2,043
$2,503
$2,065
$2,065
$2,864
$711
$711
$872
$719
$719
$997
$2,755
$2,755
$3,374
$2,784
$2,784
$3,861
   Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.
                                                5-414

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   Costs associated with powersplit HEVs are equivalent to those used in the FRM except for
updates to 2013 dollars and use of new learning curve (curve 24). The costs incremental to the
baseline configuration for our different vehicle classes are shown below. These costs are used to
characterize technology costs in the baseline fleet; EPA does not build OMEGA packages using
this technology and instead uses the strong HEV technology presented earlier.
  Table 5.173 Costs for Powersplit HEV Technology for Different Vehicle Classes (dollar values in 2013$)
Tech
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
DMC: base
cost
1C:
complexity
$3,128
$3,482
$3,767
$4,570
$5,620
$5,620
Highl
Highl
Highl
Highl
Highl
Highl






DMC:
learning
curve
1C: near
term
thru
24
24
24
24
24
24
2018
2018
2018
2018
2018
2018






2017
$2,992
$3,330
$3,602
$4,370
$5,374
$5,374
$1,754
$1,952
$2,112
$2,563
$3,151
$3,151
$4,746
$5,282
$5,714
$6,933
$8,525
$8,525
2018
$2,934
$3,265
$3,532
$4,286
$5,270
$5,270
$1,751
$1,948
$2,108
$2,557
$3,145
$3,145
$4,684
$5,213
$5,640
$6,843
$8,414
$8,414
2019
$2,881
$3,206
$3,469
$4,209
$5,175
$5,175
$1,073
$1,194
$1,291
$1,567
$1,927
$1,927
$3,953
$4,400
$4,760
$5,776
$7,102
$7,102
2020
$2,832
$3,152
$3,410
$4,138
$5,088
$5,088
$1,071
$1,192
$1,290
$1,565
$1,924
$1,924
$3,904
$4,344
$4,700
$5,703
$7,012
$7,012
2021
$2,788
$3,103
$3,357
$4,073
$5,008
$5,008
$1,070
$1,191
$1,288
$1,563
$1,922
$1,922
$3,858
$4,293
$4,645
$5,636
$6,930
$6,930
2022
$2,747
$3,057
$3,307
$4,013
$4,935
$4,935
$1,068
$1,189
$1,286
$1,561
$1,919
$1,919
$3,815
$4,246
$4,594
$5,574
$6,854
$6,854
2023
$2,709
$3,015
$3,261
$3,957
$4,866
$4,866
$1,067
$1,188
$1,285
$1,559
$1,917
$1,917
$3,776
$4,202
$4,546
$5,516
$6,783
$6,783
2024
$2,673
$2,975
$3,219
$3,905
$4,802
$4,802
$1,066
$1,186
$1,284
$1,557
$1,915
$1,915
$3,739
$4,161
$4,502
$5,462
$6,717
$6,717
2025
$2,640
$2,938
$3,178
$3,856
$4,742
$4,742
$1,065
$1,185
$1,282
$1,556
$1,913
$1,913
$3,705
$4,123
$4,461
$5,412
$6,655
$6,655
       Note: DMC=direct manufacturing costs; IC=indirect costs; TC=total costs.

5.4    CAFE Technology Assessment

This section describes the cost and technical analysis conducted by NHTSA for this report.
Section 5.4.1 describes the development of direct and indirect costs and the application of
learning curves in the NHTSA analysis. Section 5.4.2 details GT Power and Autonomie
simulation modeling to develop technology effectiveness values for use in the CAFE model.

5.4.1   Technology Costs Used in CAFE Assessment

5.4.1.1 Direct Costs

   The majority of technology costs used by NHTSA in this analysis are the  same as those used
in the 2012 FRM. These costs, however, have been updated to 2013 dollars since all costs in this
analysis are in 2013 dollars. Based on new information, stakeholder feedback, and the 2015 NAS
report, NHTSA updated DMC for the technologies discussed below.589

5.4.1.1.1     Improved Low Friction Lubricants and Engine Friction Reduction Levels 2 & 3
(LUBEFR2 & LUBFFR3)

   For this analysis, NHTSA assumed that incremental improvements in low friction lubricants
and engine  friction reductions could be realized. Based on the Massachusetts Institute of
                                            5-415

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Technology Sloan Automotive Laboratory's "On the Road toward 2050" report, a 3 percent
combined improvement was assumed to be achievable by 2030.59° This translates into a 0.275
percent improvement compounded annually for the MY2015-2030 timeframe. The DMC basis
for this technology is the 2012 FRM EFR2_LUB2 technology cost ($12.65/cylinder). The yearly
per cylinder DMC then becomes $0.84 ($12.65/15 years). Converting this from 2010$ to 2013$
yields a DMC of $0.8875/cylinder. The yearly cost and effectiveness values are accumulated and
then applied in two discrete MYs. LUBEFR2 with an effectiveness improvement of 0.823
percent is applied in MY2018 for a DMC of $2.66/cylinder ($0.8875 x 3 years). LUBEFR3 with
an effectiveness improvement of 2.18 percent is applied in MY2023, incremental to LUBEFR2,
for a DMC of $4.44/cylinder ($0.8875 x 5 years).

5.4.1.1.2      Automatic Transmission Improvements Levels 1 & 2 (ATI1 & ATI2)

   A 1.5 percent improvement by MY2025, or 0.151 percent compounded annually for the
MY2015-2025 timeframe, was assumed based on comments received in stakeholder meetings.
The cost basis is the 2012 FRM FIEG technology cost of $202 for 2.64 percent improvement
(average improvement across all the vehicle classes) or $76.527 percent ($202/2.64 percent). This
equates to a DMC of $114.77 ($76.52/percent x 1.5 percent) for a 1.5 percent improvement. This
yields a yearly DMC of $11.48 ($114.77/10 years). Converting this from 2010$ to 2013$ yields a
yearly DMC of $12.13. The yearly cost and effectiveness values are accumulated and then
applied in two discrete MYs. ATI1 with an effectiveness improvement of 0.45 percent is applied
in MY2018 for a DMC of $36.39 ($12.13 x 3 years). ATI2 with an effectiveness improvement of
1.20 percent is applied in MY2023, incremental to ATI2,  for a DMC of $60.65 ($12.13 x 5
years).

5.4.1.1.3      High Compression Ratio Engine

   This is analogous to Mazda's SkyActiv engine technology.  The costs for the HCR technology
are from the 2015 NAS report. The NAS report's DMC include the DMC for direct injection so
these DMC are subtracted to get the DMC for HCR with direct injection.  The DMC costs for
MY2017 in 2010$ are $86 for an 14 engine, $129 for a V6 engine and $204 for a V8 engine. In
2013$ the DMC become $90.84, $136.27 and $215.50, respectively.

5.4.1.1.4      Advanced Diesel Engine (ADSL) Engine

   The DMC for the ADSL technology are also from the 2015 NAS report.  The DMC for
MY2017 in 2010$ is $3,023 for an 14 engine, $3,565 for a V6 engine and $3,795 for a V8
engine. In 2013$ the DMC become  $3,193.47, $3,766.03  and $4,009.00 respectively.

5.4.1.1.5      7-speedManual Transmission

   Due to limited availability of cost information on 7-speed manual transmissions, NHTSA is
using the DCT8 technology DMC, which is sourced from the 2012 FRM.

5.4.1.1.6      6-speedAutomatic Transmission

   The DMC for the AT6 technology is from the 2015 NAS report.  The DMC for MY2017 in
2010$ is -$13.00. In 2013$ the DMC becomes -$13.73. The AT6 technology cost is relative to
the 4-speed automatic. In contrast to this estimate,  the TSD for the earlier 2012-2016 MY CAFE
standards (EPA/NHTSA 2010) developed a cost of $101 for a six-speed automatic transmission
relative to a four-speed automatic transmission. The FEV teardown cost analysis determined that
                                            5-416

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
the six-speed transmission was $106 less costly than the five-speed transmission.591 The 2012
TSD indicated that this counterintuitive result was attributed to the six-speed transmission having
a Lepelletier-type gear set instead of a conventional planetary gear set, which requires an
additional one way clutch. Subsequent to the 2012-2016 MY TSD, the EPA/NHTSA 2017-2025
MY Technical Support Document estimated a direct manufacturing cost of-$13 (savings) for a
six-speed automatic transmission relative to a four-speed automatic transmission, which appears
to have resulted from using only the case with the Lepelletier gear set.592

5.4.1.1.7     8-speedAutomatic Transmission

   The DMC for the ATS technology is from the updated FEV teardown study.593 The DMC for
MY2012 in 2007$ is $74.81. In 2013$ the DMC becomes $82.18.  This  cost increase is relative
to the AT6 technology as indicated by the FEV teardown. The net incremental direct
manufacturing cost shown is solely based on the physical hardware evaluated. Many of the
subsystems were deemed cost neutral between the 6AT and SAT. Much  of the cost analysis work
was focused on the cost difference in the gear train and internal clutch subsystems.

5.4.1.1.8     6-speedDual Clutch Transmission

   Due to concerns regarding the challenges associated with the noise, vibration and harshness
(NVH),  integration, and drivability issues of dual clutch transmissions, NHTSA believes that the
DMC for the DCT6 is higher than the negative DMC used in the 2012 FRM. To better account
for these issues, NHTSA chose to update the DCT6 technology DMC using the upper cost for
the 6-speed Dry DCT found in the 2015 NAS report.  The DMC in this analysis for MY2017 in
2010$ is $31.00 relative to 6 speed automatic - Lepelletier type. In 2013$ the DMC becomes
$32.75.  Similarly the DMC using the upper cost for the 6-speed DCT for MY2017 is  $88.00
relative to 6 speed automatic - Lepelletier type. In 2013$ the DMC becomes $94.01. Estimated
2025 MY DMC for DCT6 dry and wet clutch costs of $26 and $75 (2010$) relative to AT6-
Lepelletier type and using the upper cost.  These costs adjusted for 2013$ are $27.78 and $80.13

   The committee found that the currently high costs of DCTs stem from the relatively low sales
volumes, compounded by the fact that DCTs used by different vehicle manufacturers have
different mechatronics for clutch and shift fork actuation. The actuation units can be
electromechanical, electrohydraulic, or a mixture of both. The clutch modules vary significantly.
Although the main difference is between wet and dry clutch configurations, other differences
include the use of torsional dampers, while others rely on a damper in the separate dual mass
flywheel. Since the hardware components from one DCT to another can  vary significantly, a
large variation in costs can be expected.589

5.4.1.1.9     8-speed Dual Clutch Transmission

   For this analysis NHTSA continued to rely on the  FEV teardown study for the DMC of the
DCTS technology. However, since the 2012 FRM, FEV has updated the teardown study for 8-
speed transmission technologies. The DMC for the DCTS technology has been updated from the
2012 FRM and is now $217.65 in MY2012 in 2007$. In 2013$ the DMC becomes $229.92.

   The committee found that the currently high costs of DCTs stem from the relatively low sales
volumes, compounded by the fact that DCTs used by different vehicle manufacturers have
different mechatronics for clutch and shift fork actuation. The actuation units can be
electromechanical, electrohydraulic, or a mixture of both. The clutch modules vary significantly.
                                            5-417

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Although the main difference is between wet and dry clutch configurations, other differences
include the use of torsional dampers, while others rely on a damper in the separate dual mass
flywheel. Since the hardware components from one DCT to another can vary significantly, a
large variation in costs can be expected. This large variation in hardware components is partly
responsible for DCTs not achieving significant cost reductions at current production volumes.

5.4.1.1.10    Continuously Variable Transmission

   The DMC for the CVT technology is sourced from the 2015 NAS report. The DMC for
MY2017 in 2010$ is $179.00.  In 2013$ the DMC becomes $189.09. NHTSA provided an
estimated 2012 MY direct manufacturing cost of $200 (2007 dollars) for the CVT relative to a
four-speed automatic transmission. Some manufacturers' estimates significantly exceeded
NHTSAs maximum range. This wide range of estimates is believed to reflect wide variations in
losses in the CVT.

5.4.1.1.11    Belt Integrated Starter Generator

   For the last FRM, NHTSA considered high-voltage BISG systems, or systems over 60V (SAE
J2232)594 In recent years, manufacturers have commercialized low-voltage BISG systems (such
as 48V)  as an  alternative to high-voltage BISG systems. With limited need for high voltage
protection, the 48V BISG systems may have lower direct manufacturing costs than their high
voltage counterparts.

   The 2015 light duty fleet has many examples of 48V BISG systems for small sized and
medium sized vehicles, but the fleet has few examples of low-voltage BISG on trucks and large
sport utility vehicles. The low voltage BISG systems operate in much the same way as the high
voltage systems but require higher current to produce a given amount of power. On trucks and
large SUVs, engineering performance of a 48V BISG system may  or may not perform as well as
high voltage BISG systems. NHTSA seeks comment on the  functionality and practicability of
low-voltage BISG systems for truck and large SUV applications. Based on an EPA teardown
study conducted by FEV of a 48V BISG system and 115V BISG system, NHTSA has lowered
the projected cost of BISG technology.595  For Small Car, Medium Car, and Small SUV the
BISG DMC is $1013.00 in MY2017.

5.4.1.1.12    Crank Integrated Starter Generator

   For this analysis, NHTSA is using the Integrated Motor Assist DMC from Table 3-47 found
in the 2012 FRM TAR. The DMC for MY2017 in 2010$ is $2008.00 for Small Car, $2541.00
for Medium Car, $2552.00 for Small SUV and Medium SUV, and  $3118.00 for Pickup. In
2013$ those costs become $2121.23, $2684.28, $2695.91, and $3293.82, respectively.

5.4.1.1.13    Electric Power Steering

   For this analysis, NHTSA is using DMC from the 2012 FRM. The DMC for MY2017 in
2010$ is $92.00 per vehicle.  In 2013$ the cost becomes $95.86 per vehicle.

5.4.1.1.14    Improved Accessories (IACC1 & IACC2)

   For this analysis, NHTSA is using DMC from the 2012 FRM. Level 1 technology (IACC1)
provides a high-efficiency alternator and level 2 (IACC2) provides a high-efficiency alternator
and incorporates mild regeneration.  For level 1, the DMC for MY2017 in 2010$ is $75.00 per
                                            5-418

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
vehicle, which becomes $77.96 after adjusting for 2013 dollars.  For level 2, the DMC for
MY2017 in 2010$ adds another $45.00 per vehicle ($120 total),  which is an additional $48.12
per vehicle ($126.08 total) in 2013 dollars.

5.4.1.1.15    Low Drag Brakes

   For this analysis, NHTSA is using DMC from the 2012 FRM. The DMC for MY2017 in
2010$ is $59.00 per vehicle. In 2013$ the cost becomes $62.03  per vehicle.

5.4.1.1.16    Secondary Axle Disconnect

   For this analysis, NHTSA is using DMC from the 2012 FRM. The DMC for MY2017 is
$82.00 per vehicle. After adjusting for 2013 dollars, the cost becomes $85.57 per vehicle.

5.4.1.1.17    Low Rolling Resistance Tires

   For this analysis, NHTSA is using DMC from the 2012 FRM. Level 1 technology (ROLL10)
provides a ten percent reduction in rolling resistance and level 2  (ROLL20) provides a twenty
percent reduction. For level 1, the DMC for MY2017 in 2010$ is $5.40 per vehicle, which
becomes $5.64 after adjusting for 2013 dollars. For level 2, the DMC for MY2017 in 2010$ is
$40.00 per vehicle, and becomes $42.77 per vehicle in 2013 dollars.

5.4.1.1.18    Aerodynamic Drag Reduction

   For this analysis, NHTSA is using DMC from the 2012 FRM. Level 1 technology (AERO10)
provides a ten percent reduction in aerodynamic drag resistance  and Level 2 (AERO20) provides
a twenty percent reduction. For level 1, the DMC for MY2017 in 2010$ is $41.00 per vehicle,
which becomes $42.86 after adjusting for 2013 dollars. For level 2, the DMC for MY2017 in
2010$ is $123.00 per vehicle, and becomes $128.57 in 2013 dollars.

5.4.1.1.19    Mass Reduction

   NHTSA awarded a contract to an engineering team consisting of Electricore, Inc. (prime
contractor), EDAG, and George Washington University to design a future midsize lightweight
vehicle (LWV). This vehicle is assumed to be manufactured using processes available in model
year 2017-2025 and be capable of high volume production (200,000 units per year). The team's
goal was to determine the maximum feasible weight reduction while maintaining the same
vehicle functionalities as the baseline vehicle, such as  performance, safety,  and crash rating.

   Furthermore, the retail price of the LWV must be within +10  percent of the original vehicle.
Based upon its production volume, market share, and five-star crash rating, the team selected the
model year 2011 Honda Accord as its baseline vehicle. Because  a lighter vehicle needs less
power, the vehicle powertrain was downsized but limited to the same naturally aspirated engine.
Any analysis of an advanced powertrain, such as a hybrid electric vehicle, was outside the scope
of this project. The major boundary conditions for this project included the following:

       •  Maintain or improve vehicle size compared to the baseline vehicle.
       •  Maintain retail price parity (±10 percent variation) with the baseline vehicle.
       •  Maintain or improve vehicle functionalities compared to the baseline vehicle,
          including maintaining comparable performance in NHTSA's New Car Assessment
                                            5-419

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
          Program (NCAP) frontal, side, side pole and IMS test programs through appropriate
          crash simulations.
       •  Powertrain may be downsized, however alternate powertrain configurations (i.e.
          hybrid electric, battery electric, and diesel) will not be considered.
       •  All advanced design, material, technologies and manufacturing processes must be
          realistically projected to be available for fleet wide production in time frame of model
          years 2017-2025 and capable of high volume production (200,000 units per year).
       •  Achieve the maximum feasible amount of mass reduction within the above listed
          constraints
   Overall, the complete LWV 1.0 achieved a total weight savings of 22 percent (332 kg) from
the baseline vehicle (1480 kg) at an incremental cost increase of $319 or $0.96 per kg. To
achieve the same vehicle performance as the baseline vehicle, the size of the engine for the LWV
was proportionally reduced from 2.4L-177 HP to  1.8L-140HP. Without the mass and cost
reduction allowance for the powertrain (engine and transmission) the mass saving for the 'glider'
was 24 percent (264kg) at a mass saving cost premium of $1.63 per kg mass saving.

   NHTSA released the first version of the report after it was peer reviewed.596 Subsequent to the
release of the report, Honda examined the report in  detail and offered their observations to
NHTSA on the components chosen to light-weight the vehicle. In addition, Honda provided
information on limitations to downsizing some of the components due to both within platform
sharing and cross-platform sharing. The other main observation from Honda was in the area of
crashworthiness, performance and drivability issues and ground clearance.597

   In 2013, NHTSA awarded a subsequent contract to Electricore with EDAG as subcontractor
to perform additional crash simulations on the light-weight Honda Accord vehicle (LWV 1.0) to
address Honda's comments. The light-weight 2011 Honda Accord (LWV 1.0) was modified to
address Honda's suggestion in areas of crashworthiness, Noise & Vibration and in drivability
performance (LWV 1.1). NHTSA used modified light-weighted 2011 Honda Accord (LWV 1.1)
to perform additional design and crash simulation to meet Insurance Institute of Highway Safety
(IIHS) evaluation  of Small Overlap Test (SOL). The light-weighted version (LWV 1.2) of 2011
Honda Accord incorporates Honda's suggestion and meets IIHS small overlap test requirements.
The following paragraph describes the progression of changes in mass reduction and cost
changes as a result of Honda's suggestion and also in meeting IIHS small overlap test
requirements relative to LWV 1.0.

    In addressing  Honda's comments, the weight of the body structure of the LWV 1.1 was
 increased by 11.5 kg and the cost was reduced by $13.08 from the original LWV 1.0 design. In
 addition, some of Honda's recommendations for NVH and durability were accepted. The total
 weight and cost of the LWV 1.1 increased by 21.75 kg and  $18.13, respectively. To address the
 IIHS SOL test (LWV 1.2) the weight of the vehicle was increased by 6.90 kg and the cost by
 $26.88. The new LWV 1.2 design was modeled  and assessed for the performance of
 crashworthiness in seven crash safety tests. The new design achieved a "good," rating in all tests
 that are comparable to the safety rating of the model year 2013 Accord. Table 5.174 shows the mass
 reduction and associated costs from light weighted vehicle version 1.0 to light-weighted version
 1.2.  The baseline is  a Honda Accord with a weight of 1,480 kg.
                                             5-420

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
    Table 5.174 Mass Reduction and Associated Costs Going From Vehicle Version 1.0 to Vehicle Version 1.2
Model
LW1.0
LWV1.1

LWV 1.2
Mass
savings (kg)
332
320.8
310.55
303.65
Percentage Mass
Reduction
22.43%
21.68%
20.98%
20.52%
Cost Increase
$ 319.00
$ 305.92
$337.13
$ 364.01
$/kg
$ 0.96
$ 0.95
$ 1.09
$ 1.20
Comments

Addressing Honda's comments, 11.5kg
was added, Cost was reduced by $13.08
NVH mass add was 10.25kg
IIHS SOL mass add was 6.6kg, cost
increase of $26.88
   The list of components that were light-weighted was rearranged in sequence based on cost
effectiveness as shown in Table 5.175. Figure 5.141 shows a graphical representation of cost per
kilogram at various levels of mass reduction plotted from Table 5.175.  As can be seen from the
cost curve in Figure 5.141, cost per kilogram increases progressively as some of the vehicle
structural components are light-weighted due to adoption of higher strength materials and in
some cases switching from steel to aluminum. The powertrain components which include
engine, transmission,  and fuel systems such as fuel filler pipe, fuel tank, fuel pump, etc., exhaust
systems and cooling systems were not considered for application of primary mass reduction but
benefits of secondary mass reduction were accounted for. These powertrain components are
assumed to be downsized only after the primary vehicle structural components (Body-In-White)
achieve certain level of mass reduction. The National Academy of Sciences (NAS) estimated
mass reduction costs assuming that powertrain downsizing be considered after the primary
vehicle mass is reduced by 10 percent of original mass. NHTSA considered the NAS approach
and applied powertrain downsizing (secondary mass savings) after the vehicle structural
components (primary mass savings) had achieved 10 percent mass reduction.  In the case of the
mass reduction study  of the 2011 Honda Accord passenger car, the baseline 2.4L engine was
replaced by 1.8L engine which was already in production. The 1.8L engine was used in Honda
Civic model which is a compact passenger car.  As a consequence of using a smaller engine, the
fuel system and exhaust system were downsized to match 1.8L engine while maintaining the
same driving range and performance. The mass reduction and cost savings from smaller
powertrain components along with primary vehicle structural components resulted in a 20
percent overall mass reduction from the baseline 2011 Honda Accord. This design configuration
is represented as the AHSS+AL solution point in Figure 5.141. Due to this approach, the cost
curve bends after 10 percent to reach the solution point as shown in Figure 5.141. As a
consequence, the cost per kilogram at the final solution point is less than the cost per kilogram at
10 percent mass reduction. Note here, at 10 percent mass reduction, no secondary mass savings
are considered.

   Additional mass reduction solution points shown in Figure 5.141 were analytically developed
such as Aluminum (AL) intensive solution and Carbon Fiber Reinforced Plastics (CFRP)
intensive solution. Note here that AL and CFRP intensive solutions are analytical solutions only
and no computational models were built to verify all the performance metrics to the baseline
2011 Honda Accord. Computational models were built for only the most cost effective light-
weight solution to verify  for all performance metrics (AHSS+AL  Solution).
                                            5-421

-------
               Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.175 Mass Reduction and Costs for Vehicle Components/System
Vehicle
Component/System
Front Bumper
Front Door Trim
Front Door Wiring Harness
Head Lamps
HVAC
Insulation
Interior Trim
Parking Brake
Rear Door Trim
Rear Door Wiring Harness
Tail Lamps
Tires
Wiring and Harness
Wheels
Rear Bumper
Instrument Panel
Body Structure
Deck lid
Hood
Front Door Frames
Fenders
Seats
Rear Door Frames
Powertrain components
(Engine, transmission, Fuel
system, Exhaust system,
coolant system), Brakes
etc.
Cumulative Mass
Saving
3.59
4.93
5.23
6.94
9.54
12.74
15.77
16.76
17.89
18
18.63
23.08
27.38
28.82
32.33
41.78
96.18
101.39
108.86
124.26
127.53
147.56
159.02
303.65
Cumulative MR%
0.24%
0.33%
0.35%
0.47%
0.64%
0.86%
1.07%
1.13%
1.21%
1.22%
1.26%
1.56%
1.85%
1.95%
2.18%
2.82%
6.50%
6.85%
7.36%
8.40%
8.62%
9.97%
10.74%
20.52%
Cumulative Cost
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-1.23
-$1.23
$0.53
$17.27
$173.13
$188.97
$211.49
$262.88
$274.98
$374.02
$428.47
364.01
Cumulative Cost
$/kg
-0.34
-0.25
-0.24
-0.18
-0.13
-0.10
-0.08
-0.07
-0.07
-0.07
-0.07
-0.05
-0.04
-0.04
0.02
0.41
1.80
1.86
1.94
2.12
2.16
2.53
2.69
1.20
                              5-422

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment

57 -

c*
*A .
I
5* S3
52 -
SO -
0
SI





NHTSA Passenger Car Cost Curve








r




M*
-*-












.X
2'













X






.*•



y

^




= i«


^


T



10



24-jf

A












- 7019.5s3 + 780 54
R- = 0.994


"*^ 	 -~—



— ^
---.


-~^,
•—^


%



15


\f +
1

	



4.02




Ife -








02522


^"^


"~~^






-•
• AHSS
Solution
%



2(|%


















• J
S

U.
olut





CTEP
Solution



on
UISS + Al Solution


25°.i
























3C
Mass Reduction % (curb Weight]
-Primary Mass Reduction • Optimized Solutions (Honda/SOL) 	 Poly. (Primary Mass Reduction)
%
                        Figure 5.141 NHTSA Passenger Car Cost Curve

   A fitted curve was developed based on the above listed mass reduction points to derive cost
per kilogram at distinct mass reduction points. These are shown in Table 5.176.
               Table 5.176  Cost Per Kilogram at Distinct Mass Reduction Points MR%

                               PC
                               MRO
                               MR1 - 5%
                               MR2 - 7.5%
                               MRS - 10%
                               MR4- 15%
                               MRS - 20%
$/kg
$0
$1.12
$1.99
$2.54
$2.33
$1.26
                                            5-423

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
c^ nnn
$2,750
$2,500
$2,250
$2,000
$1,750
$1,500
$1,250
$1,000
$750
$500
$250
$0
-$250
0.0

«N
0%
NHTSA Passenger Car Cost Curve (Direct Manufacturing
rn«f^

Ml

"

^

„-•

.*••



<+



+~

5.00%

tff










,-<










»•*


















































'












































AJ
So









_re<
151
!..<
(LWi
















/
yx
S+Al
Q













/




/
/
/


i
i
i
i
m



A

' CF
Sol








RP
uti




1
/*AL
' Solution
:








on









10.00% 15.00% 20.00% 25.00% 30.00%
Mass Reduction
          Figure 5.142 Direct Manufacturing Costs for Light-Weighting Approaches Analyzed

5.4.1.1.19.1  Light Duty Pickup Truck Light- Weighting Study

   NHTSA also awarded a contract to EDAG to conduct a vehicle weight reduction feasibility
and cost study of a 2014MY full size pick-up truck.  The light weighted version of the full size
pick-up truck (LWT) used manufacturing processes that will likely be available during the model
years 2025-2030 and with the capability of high volume production.  The goal was to determine
the maximum feasible weight reduction while maintaining the same vehicle functionalities, such
as towing, hauling, performance, noise, vibration, harshness, safety, and crash rating, as the
baseline vehicle, as well as the functionality and capability of designs to meet the needs of
sharing components across same or cross vehicle platform.  Consideration was also given to the
sharing of engines and other components with vehicles built on other platforms to achieve
manufacturing economies of scale, and in recognition of resource constraints which limit the
ability  to optimize every component for every vehicle. At the time of writing for this Draft TAR,
the report is in peer review and will be finalized by the NHTSA NPRM and EPA Proposed
Determination in 2017.

   A comprehensive teardown/benchmarking of the baseline vehicle was conducted for the
engineering analysis. The analysis included geometric optimization of load bearing vehicle
structures, advanced material utilization along with a manufacturing technology assessment that
would  be available in the 2017 to 2025 time frame. As part of the analysis, the baseline vehicle's
overall mass, center of gravity and all key dimensions were determined. Before the vehicle
teardown, laboratory torsional stiffness tests, bending stiffness tests and normal  modes of
vibration tests were performed on baseline vehicles so that these results could be compared with
the CAE model of the light weighted design. After conducting a full tear down and
                                             5-424

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
benchmarking of the baseline vehicle, a detailed CAE model of the baseline vehicle was created
and correlated with the available crash test results. The project team then used computer
modeling and optimization techniques to design the light-weighted pickup truck and optimized
the vehicle structure considering redesign of structural geometry, material grade and material
gauge to achieve the maximum amount of mass reduction while achieving comparable vehicle
performance as the baseline vehicle.  Only technologies and materials projected to be available
for large scale production and available within two to three design generations (e.g. model years
2020, 2025 and 2030) were chosen for the LWT design. Three design concepts were evaluated:
1) a multi-material approach; 2) an aluminum intensive approach; and 3) a Carbon Fiber
Reinforced Plastics approach.  The multi-material approach was identified as the most cost
effective.  The recommended materials (advanced high strength steels, aluminum, magnesium
and plastics), manufacturing processes, (stamping, hot stamping, die casting, extrusions, and roll
forming) and assembly methods (spot welding, laser welding, riveting and adhesive bonding) are
currently used, although some to a lesser degree than others. These technologies can be fully
developed within the normal product design cycle using the current design and development
methods.

   The design of the LWT was verified, through CAE modeling, that it meets all relevant crash
tests performance.  The LS-DYNA finite element software used by the EDAG team is an
industry standard for crash simulation and modeling.  The researchers modeled the
crashworthiness of the LWT design using the NCAP Frontal, Lateral Moving Deformable
Barrier, and Lateral Pole tests, along with the IIHS Roof, Lateral Moving Deformable Barrier,
and Frontal Offset (40 percent and 25 percent) tests.  All of the modeled tests were  comparable
to the actual  crash tests performed on the 2014  Silverado in the NHTSA database. Furthermore,
the FMVSS No. 301 rear impact test was modeled and it showed no damage to the fuel system.

   The baseline 2014 MY Chevrolet Silverado's platform shares components across several
platforms. Some of the chassis components and other structural components were designed to
accommodate platform derivatives, similar to the components in the baseline vehicle which are
shared across platforms such as GMT 920 (GM Tahoe, Cadillac Escalade, GMC Yukon), GMT
930 platform (Chevy Suburban, Cadillac Escalade ESV, GMC Yukon XL), and GMT 940
platform (Chevy Avalanche and Cadillac Escalade EXT) and GMT 900 platform (GMC Sierra).
As per the National Academy of Science's guidelines, the study assumes  engines would be
downsized or redesigned for mass reduction levels at or greater than 10 percent. As a
consequence of mass reduction, several of the components used designs that were developed for
other vehicles in the weight category  of light-weighted designed vehicles were used to maximize
economies of scale and resource limitations.  Examples include brake systems, fuel tanks, fuel
lines, exhaust systems, wheels, and other components.

   Cost is  a key consideration when vehicle manufacturers decide which fuel-saving technology
to apply to a vehicle.  Incremental cost analysis for all of the new technologies applied to reduce
mass of the light-duty full-size pickup truck designed were calculated.  The cost estimates
include variable costs as well as non-variable costs,  such as the manufacturer's investment cost
for tooling. The cost estimates include all the costs directly related to manufacturing the
components. For example, for a stamped sheet metal part, the cost models estimate the costs for
each of the operations involved in the manufacturing process, starting from blanking the steel
from coil through the final stamping operation to fabricate the component.  The final estimated
total manufacturing cost and assembly cost are  a sum total of all the respective cost elements
                                             5-425

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
including the costs for material, tooling, equipment, direct labor, energy, building and
maintenance.

   The information from the LWT design study was used to develop a cost curve representing
cost effective full vehicle solutions for a wide range of mass reduction levels.  The cost curve is
shown in Figure 5.143. At lower levels of mass reduction, non-structural components and
aluminum closures provide weight reduction which can be incorporated independently without
the redesign of other components and are stand-alone solutions for the LWV.  The holistic
vehicle design using a combination of AHSS and aluminum provides good levels of mass
reduction at reasonably acceptable cost. The LWV solution achieves 17.6 percent mass
reduction from the baseline curb mass. Further two more analytical mass reduction solutions (all
aluminum and all carbon fiber reinforced plastics) were developed to show additional mass
reduction that could be potentially achieved beyond the LWV mass reduction solution point. The
aluminum analytical solution predominantly uses aluminum including chassis frame and other
components. The carbon fiber reinforced plastics analytical solution predominantly uses CFRP in
many of the components.  The CFRP analytical solution shows higher level of mass reduction but
at very high costs. Note here that both all-Aluminum and all CFRP mass reduction solutions are
analytical solutions only and no computational models were developed to examine all the
performance metrics.

   An analysis was also conducted to examine the cost sensitivity of major vehicle systems to
material cost and production volume variations.
$20.00

$15.00

$1000

 $5.00

 $000
     C
•S5.00

-$10.00

•$15.00

-$2000

-$2500
                           NHTSA Light Truck Mass Reduction Cost Curve
                        172698x''-40530x4-2587.5x! + 1022x>-25.202x +0.1934
                                      Rz = 0.9965
                                                          AI. Solution  '
                                                               ' AHSS - AL
                                                                Solulioixl WYt
                                   8%   10%   12%   14%   16%   18%  20%   22%   24%  26%
                                                %MR
    Figure 5.143 NHTSA Draft Light Duty Pickup Truck Lightweighting (AHSS Frame with Aluminum
                           Intensive) Cost Curve (DMC $/kg v %MR)
   Table 5.177 lists the components included in the various levels of mass reduction for the
LWV solution. The components are incorporated in a progression based on cost effectiveness.
                                             5-426

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
              Table 5.177 Components Included for Different Levels of Mass Reduction
Vehicle Component/System
Interior Electrical Wiring
Headliner
Trim - Plastic
Trim - misc.
Floor Covering
Headlamps
HVAC System
Tail Lamps
Chassis Frame
Front Bumper
Rear Bumper
Towing Hitch
Rear Doors
Wheels
Front Doors
Fenders
Front/Rear Seat & Console
Steering Column Assy
Pickup Box
Tailgate
Instrument Panel
Instrument Panel Plastic Parts
Cab
Radiator Support
Powertrain
Cumulative Mass
Saving
1.38
1.56
2.59
4.32
4.81
6.35
8.06
8.46
54.82
59.93
62.96
65.93
77
102.25
116.66
128.32
157.56
160.78
204.74
213.14
218.66
221.57
304.97
310.87
425.82
Cumulative
MR%
0.06%
0.06%
0.11%
0.18%
0.20%
0.26%
0.33%
0.35%
2.25%
2.46%
2.59%
2.71%
3.17%
4.20%
4.80%
5.28%
6.48%
6.61%
8.42%
8.76%
8.99%
9.11%
12.54%
12.78%
17.51%
Cumulative
Cost
($28.07)
($29.00)
($34.30)
($43.19)
($45.69)
($45.69)
($45.69)
($45.69)
$2.57
$7.89
$11.04
$14.13
$28.09
$68.89
$92.53
$134.87
$272.57
$287.90
$498.35
$538.55
$565.06
$580.49
$1,047.35
$1,095.34
1246.68
Cumulative Cost
$/kg
-20.34
-18.59
-13.24
-10.00
-9.50
-7.20
-5.67
-5.40
0.05
0.13
0.18
0.21
0.36
0.67
0.79
1.05
1.73
1.79
2.43
2.53
2.58
2.62
3.43
3.52
2.93
   A fitted curve was developed based on the above listed mass reduction points to derive cost
per kilogram at distinct mass reduction points as shown in Table 5.178.

                         Table 5.178 Cost Per Kilogram of Mass Reduced
MR% $/kg
5.0%
7.5%
10.0%
15.0%
20.0%
$0.97
$2.09
$2.98
$3.27
$5.75
                                              5-427

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
As explained above, the direct manufacturing costs for the components listed above are shown in
Figure 5.144.
NHTSA Light Truck Cost Curve (Direct Manufacturing Cost)
$4 000
$3,500 -
$3,000 -
$2,500 -
$2,000 -
$1,500 -
$1,000 -
$500 -
$0 <
0<
-$500 -






». 	 , A
/b






p. +.-.+.»
*






^.^--
4






.^"





^,,'<
10%





+ 	







/
f
*('
i
i
i
i
/
	 J\HSS + AL
Solution (LW
15%
20


»



IV)
%







M

%
            Figure 5.144 NHTSA Light Truck Cost Curve ($/Vehicle vs. % Mas Reduction)
   Table 5.179 shows the direct manufacturing costs at distinct mass reduction levels.

             Table 5.179 Direct Manufacturing Costs for Different Mass Reduction Levels
LT Baseline Curb Wt. 2432 kg
MRO
MRl-5%
MR2-7.5%
MRS -10%
MR4-15%
MRS -20%
Mass
Reduction (kg)
0
122
182
243
365
486
DMC ($)
$0
$118
$381
$725
$1193
$2797
5.4.1.2 Indirect Costs

5.4.1.2.1      Methodologies for Determining Indirect Costs

   To produce a unit of output, vehicle manufacturers incur direct and indirect costs. Direct
costs include cost of materials and labor costs. Indirect costs are all the costs associated with
producing the unit of output that are not direct costs - for example, they may be related to
production (such as research and development), corporate operations (such as salaries, pensions,
and health care costs for corporate staff), or selling (such as transportation, dealer support, and
                                              5-428

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
marketing).  Indirect costs are generally recovered by allocating a share of the costs to each unit
of good sold. Although it is possible to account for direct costs allocated to each unit of good
sold, it is more challenging to account for indirect costs allocated to a unit of goods sold. To
make a cost analysis process more feasible, markup factors, which relate total indirect costs to
total direct costs, have been developed.  These factors are often referred to as retail price
equivalent multipliers.

   Cost analysts and regulatory agencies (including both NHTSA and EPA)  have frequently used
these multipliers to predict the resultant impact on costs associated with manufacturers'
responses to regulatory requirements. The best approach, if it were possible, to determining the
impact of changes in direct manufacturing costs on a manufacturer's indirect costs would be to
actually estimate the cost impact on each indirect cost element. However, doing  this within the
constraints of an agency's time or budget is not always feasible, and the technical, financial, and
accounting information to carry out such an analysis may simply be unavailable.

   The one empirically derived metric that addresses the markup of direct costs to consumer
costs is the RPE multiplier, which is measured from manufacturer 10-K accounting statements
filed with the Securities and Exchange Commission.  Over roughly a three decade period, the
measured RPE has been remarkably stable, averaging 1.5, with minor annual variation. The
National Research Council notes that, "Based on available data, a reasonable RPE multiplier
would be  1.5."  The historical trend in the RPE is illustrated in Figure 5.145.
                     RPE History,  1972-1997, and 2007
    2.00
    1.90
    1.80
    1.70
    1.60
    1.50
    1.40
    1.30
    1.20
    1.10
    1.00
                                                                 RPE
        1970
1975
1980
1985
1990
1995
2000
2005
2010
                         Figure 5.145 RPE History 1972-1997 and 2007

    RPE multipliers provide, at an aggregate level, the relationship between revenue and direct
manufacturing costs. They are measured by dividing total revenue by direct costs. However,
because this provides only a single aggregate measure, using RPE multipliers results in the
application of a common incremental markup to all technologies. It assures that the aggregate
cost impact across all technologies is consistent with empirical data, but does  not allow for
indirect cost discrimination among different technologies. Thus, a concern in using the RPE
multiplier in cost analysis for new technologies added in response to regulatory requirements is
that the indirect costs of vehicle modifications are not likely to be the same for all different
                                             5-429

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
technologies. For example, less complex technologies could require fewer R&D efforts or less
warranty coverage than more complex technologies. In addition, some simple technological
adjustments may, for example, have no effect on the number of corporate personnel and the
indirect costs attributable to those personnel. The use of RPEs, with their assumption that all
technologies have the same proportion of indirect costs, is likely to overestimate the costs of less
complex technologies and underestimate the costs of more complex technologies. However, for
regulations such as the CAFE and GHG emission standards under consideration, which drive
changes to nearly every vehicle system, overall average indirect costs should align with the RPE
value.  Applying RPE to the cost for each technology assures that alignment.

   Modified multipliers have been developed by EPA,  working with a contractor, for use in
rulemakings.598 These multipliers are referred to as indirect cost multipliers (or ICMs).  ICMs
assign unique incremental changes to each indirect cost contributor at several different
technology levels.

       ICM = (direct cost + adjusted indirect cost)/(direct cost)

   Developing the ICMs from the RPE multipliers requires developing adjustment factors based
on the complexity of the technology and the time frame under consideration: the less complex a
technology, the lower its ICM, and the longer the time  frame for applying the technology, the
lower the ICM. This methodology was used in the cost estimation for the recent light-duty MYs
2012-2016 and MYs 2017-2025 rulemaking and for the heavy-duty MYs 2014-2018 rulemaking.
The ICMs for the light-duty context were developed in a peer-reviewed report from RTI
International and were subsequently discussed in a peer-reviewed journal article.599
Importantly, since publication of that peer-reviewed journal article, the agencies have revised the
methodology to include a return on capital (i.e.,  profits) based on the assumption implicit in
ICMs (and RPEs) that capital costs are proportional to  direct costs, and businesses need to be
able to earn returns on their investments.

5.4.1.2.2      Indirect Cost Multipliers Used in this Analysis

   Since their original development in February 2009, the agencies have made some changes to
both the ICMs factors and to the method of applying those factors relative to the factors
developed by RTI and presented in their reports.  We have described and explained those
changes in several rulemakings over the years, most notably the 2017-2025 FR for light vehicles
and the more recent Heavy-duty GHG Phase 2 NPRM.600  In the 2015  NAS study, the committee
stated a conceptual agreement with the ICM method since ICM takes into account design
challenges and the activities required to implement each technology. However, although
endorsing ICMs as a concept, the NAS Committee stated that".. .the empirical basis for such
multipliers is still lacking, and,  since their application depends on expert judgment, it is not
possible to determine whether the Agencies' ICMs are  accurate or not."  NAS also states that
".. .the  specific values for the ICMs are critical since they may affect the overall estimates of
costs and benefits for the overall standards and the cost effectiveness of the individual
technologies." The committee did encourage continued research into ICMs given the lack of
empirical data for them to evaluate the ICMs used by the agencies in past analyses. EPA, for its
part, continues to study the issue surrounding ICMs but has not pursued further efforts given
resource constraints and demands in areas such as technology benchmarking and cost teardowns.
On balance, NHTSA believes that the empirically derived RPE is a more reliable basis for
estimating indirect costs. To ensure overall indirect costs in the analysis align with the RPE
                                             5-430

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
value, NHTSA has developed its primary analysis based on applying the RPE value of 1.5 to
each technology. NHTSA also has conducted a sensitivity analysis examining the impact of
applying the ICM approach using the same methodology and multiplier values described in
Section 5.3 forEPA's analysis.

   The ICMs used in NHTSA's sensitivity analysis are shown in Table 5. ISO.601 Near term
values account for differences in the levels of R&D, tooling, and other indirect costs that will be
incurred. Once the program has been fully implemented, some of the indirect costs will  no
longer be attributable to the standards and,  as such, a lower ICM factor is applied to direct costs.
                     Table 5.180 Indirect Cost Multipliers Used in this Analysis

Complexity
Low
Medium
Highl
High2
2017-2025 FRM& this TAR
Near term
1.24
1.39
1.56
1.77
Long term
1.19
1.29
1.35
1.50
   We note two important aspects to the ICM method. First, the ICM consists of two portions: a
small warranty-related term and a second, larger term to cover all other indirect costs elements.
The breakout of warranty versus non-warranty portions to the ICMs are presented in Table
5.181. The latter of these terms does not decrease with learning and, instead, remains constant
year-over-year despite learning effects which serve to decrease direct manufacturing costs.
Learning effects are described in the next section. The second important note is that all indirect
costs are forced to be positive, even for those technologies estimated to have negative direct
manufacturing costs.
                    Table 5.181 Warranty and Non-Warranty Portions of ICMs

Complexity
Low
Medium
Highl
High2
Near term
Warranty
0.012
0.045
0.065
0.074
Non-warranty
0.230
0.343
0.499
0.696
Long term
Warranty
0.005
0.031
0.032
0.049
Non-warranty
0.187
0.259
0.314
0.448
                                             5-431

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
   The ICM categories assigned to each technology and their long-term cutoffs are shown in
Table 5.182.
           Table 5.182  ICM categories and Short Term ICM Schedules for CAFE Technologies
Technology
ICM
Category
Short
Term
Through
 Low Friction Lubricants- Level 1
 Engine Friction Reduction - Level 1
 Low Friction Lubricants and Engine Friction Reduction - Level 2
 Variable Valve Timing (VVT) - Coupled Cam Phasing (CCP) on SOHC
 Discrete Variable Valve Lift (DVVL) on SOHC
 Cylinder Deactivation on SOHC
 Variable Valve Timing (VVT) - Intake Cam Phasing (ICP)
 Variable Valve Timing (VVT) - Dual Cam Phasing (DCP)
 Discrete Variable Valve Lift (DVVL) on DOHC
 Continuously Variable Valve Lift (CVVL)
 Cylinder Deactivation on DOHC
 Stoichiometric Gasoline Direct Injection (GDI)
 Cylinder Deactivation on OHV
 Variable Valve Actuation - CCP and DVVL on OHV
 Stoichiometric Gasoline Direct Injection (GDI) on OHV
 Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Small Displacement - Turbo
 Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Small Displacement - Downsize
 Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Medium Displacement -Turbo
 Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Medium Displacement -
 Downsize
 Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Large Displacement - Turbo
 Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Large Displacement - Downsize
 Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Small Displacement - Turbo
 Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Small Displacement - Downsize
 Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Medium Displacement - Turbo
 Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Medium Displacement -
 Downsize
 Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Large Displacement - Turbo
 Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Large Displacement - Downsize
 Cooled Exhaust Gas Recirculation  (EGR) - Level 1 (24 bar BMEP) - Small Displacement -
 Turbo
 Cooled Exhaust Gas Recirculation  (EGR) - Level 1 (24 bar BMEP) - Small Displacement -
 Downsize
 Cooled Exhaust Gas Recirculation  (EGR) - Level 1 (24 bar BMEP) - Medium Displacement -
 Turbo
 Cooled Exhaust Gas Recirculation  (EGR) - Level 1 (24 bar BMEP) - Medium Displacement -
 Downsize
 Cooled Exhaust Gas Recirculation  (EGR) - Level 1 (24 bar BMEP) - Large Displacement -
 Turbo
 Cooled Exhaust Gas Recirculation  (EGR) - Level 1 (24 bar BMEP) - Large Displacement -
 Downsize
 Cooled Exhaust Gas Recirculation  (EGR) - Level 2 (27 bar BMEP) - Small Displacement -
 Turbo
Low2
Low2
Low2
Low2
Medium2
Medium2
Low2
Medium2
Medium2
Medium2
Medium2
Medium2
Medium2
Medium2
Medium2
Medium2
Medium2
Medium2

Medium2
Medium2
Medium2
Medium2
Medium2
Medium2

Medium2
Medium2
Medium2

Medium2

Medium2

Medium2

Medium2

Medium2

Medium2

Medium2
2018
2018
2024
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018
2018

2018
2018
2018
2024
2018
2024

2018
2024
2018

2024

2018

2024

2018

2024

2018

2024
                                                5-432

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Small Displacement -
Downsize
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Medium Displacement -
Turbo
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Medium Displacement -
Downsize
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Large Displacement -
Turbo
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Large Displacement -
Downsize
Advanced Diesel -Small Displacement
Advanced Diesel - Medium Displacement
Advanced Diesel - Large Displacement
6-Speed Manual/Improved Internals
Improved Auto. Trans. Controls/Externals
6-Speed Trans with  Improved Internals (Auto)
6-speed DCT
8-Speed Trans (Auto or DCT)
High Efficiency Gearbox w/ dry sump (Auto or DCT)
Shift Optimizer
Electric Power Steering
Improved Accessories - Level 1
Improved Accessories - Level 2 (w/ Alternator Regen and 70% efficient alternator)
12V Micro-Hybrid (Stop-Start)
Integrated Starter Generator
Strong Hybrid (Powersplit or 2-Mode) - Level 1 - Battery
Strong Hybrid (Powersplit or 2-Mode) - Level 1 - Non-Battery
Conversion from SHEV1 to SHEV2
Strong Hybrid (P2 Parallel or 2-Mode) - Level 2 - Battery
Strong Hybrid (P2 Parallel or 2-Mode) - Level 2 - Non-Battery
Plug-in Hybrid - 20 mi range - Battery
Plug-in Hybrid - 20 mi range - Non-Battery
Plug-in Hybrid - 40 mi range - Battery
Plug-in Hybrid - 40 mi range - Non-Battery
Electric Vehicle (Early Adopter) - 75 mile range - Battery
Electric Vehicle (Early Adopter) - 75 mile range - Non-Battery
Electric Vehicle (Early Adopter) -100 mile range - Battery
Electric Vehicle (Early Adopter) -100 mile range - Non-Battery
Electric Vehicle (Early Adopter) -150 mile range - Battery
Electric Vehicle (Early Adopter) -150 mile range - Non-Battery
Electric Vehicle (Broad Market) -150 mile range - Battery
Electric Vehicle (Broad Market) -150 mile range - Non-Battery
Fuel Cell Vehicle
Charger-PHEV20
Charger-PHEV40
Charger-EV
Charger Labor
Mass Reduction - Level 1
Mass Reduction - Level 2
Mass Reduction - Level 3
Mass Reduction - Level 4
Medium2
Medium2
Medium2
Medium2
Medium2
Medium2
Medium2
Medium2
Low2
Low2
Low2
Medium2
Medium2
Low2
Low2
Low2
Low2
Low2
Medium2
Highl
Highl
Highl
Highl
Highl
Highl
High2
Highl
High2
Highl
High2
High2
High2
High2
High2
High2
High2
High2
High2
Highl
Highl
Highl
None
Low2
Low2
Low2
Low2
2018
2024
2018
2024
2018
2024
2024
2024
2018
2018
2018
2018
2018
2024
2024
2018
2018
2024
2018
2018
2024
2018
2018
2024
2018
2024
2018
2024
2018
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2024
2018
2018
2018
2018
                                                   5-433

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
 Mass Reduction - Level 5
 Low Rolling Resistance Tires - Level 1
 Low Rolling Resistance Tires - Level 2
 Low Rolling Resistance Tires - Level 3
 Low Drag Brakes
 Secondary Axle Disconnect
 Aero Drag Reduction, Level 1
 Aero Drag Reduction, Level 2
                                                         Low2
                                                         Low2
                                                         Low2
                                                         Low2
                                                         Low2
                                                         Low2
                                                         Low2
                                                         Medium2
2018
2018
2024
2024
2018
2018
2018
2024
   There is some level of uncertainty surrounding both the ICM and RPE markup factors.  The
ICM estimates used in this TAR, consistent with the 2012 final rule, group all technologies into
three broad categories and treat them as if individual technologies within each of the three
categories (low, medium, and high complexity) will have exactly the same ratio of indirect costs
to direct costs. This simplification means it is likely that the direct cost for some technologies
within a category will be higher and some lower than the estimate for the category in general.
Additionally, the ICM estimates were developed using adjustment factors developed in two
separate occasions: the first, a consensus process, was reported in the RTI report; the second, a
modified Delphi method, was conducted separately and reported in an EPA memorandum. Both
of these panels were composed of EPA staff members with previous background in the
automobile industry; the memberships of the two panels overlapped but were not the same. The
panels evaluated each element of the industry's RPE estimates and estimated the degree to which
those  elements would be expected to change in proportion to changes in direct manufacturing
costs. The method and the estimates in the RTI report were peer reviewed by three industry
experts and subsequently by reviewers for the International Journal of Production Economics.
However, the ICM estimates have not been validated through a direct accounting of actual
indirect costs for individual technologies. Finally, only a handful of technologies were examined
out of roughly 50 that will be used to meet the CAFE standards.  There is thus uncertainty
regarding both the absolute values  estimated for ICMs and their validity as representatives of the
universe of technologies.

   RPEs are also difficult to estimate because the accounting statements of manufacturers do not
neatly categorize all cost elements  as either direct or indirect costs. Hence, each researcher
developing an RPE estimate must apply a certain amount of judgment to the allocation of the
costs. We note, however, that the two independent researchers that have measured RPEs each
reached essentially identical conclusions, placing the RPE at roughly 1.5. Since empirical
estimates of ICMs are ultimately derived  from the same data used to measure RPEs, both
measures are dependent on the accuracy of RPE measurement. As noted above, the value of
RPE has not been measured for specific technologies, or for groups of specific technologies.
Thus applying a single average RPE to any given technology by definition overstates costs for
very simple technologies, or understates them for advanced technologies. This same concern
applies to ICMs within each of the general ICM complexity categories.
5.4.1.2.3
NHTSA 's Application of Learning Curves
   NHTSA applies estimates of learning curves to the various technologies that will be used to
meet CAFE standards. Learning curves reflect the impact of experience and volume on the cost
of production. As manufacturers gain experience through production, they refine production
techniques, raw material and component sources, and assembly methods to maximize efficiency
                                             5-434

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
and reduce production costs. Typically, learning curves reflect initial learning rates that are
relatively high, followed by slower learning as the easier improvements are made and production
efficiency peaks. This eventually produces an asymptotic shape to the learning curve as small
percent decreases are applied to gradually declining cost levels (see Figure 5.146)
            Cost Changes Over Time Based on Cumulative
                                     Learning
 $120
 $100
                                                                                •Cost
       1  3  5   7  9  11 13  15  17 19 21 23 25 27  29 31 33 35 37 39 41 43 45 47
           Figure 5.146 Hypothetical Illustration of Cumulative Production Based Learning
   The learning curves the agency currently uses represent our current estimates regarding the
pace of learning.  Depending on the technology, the curves assume a learning rate of 3 percent
over the previous years' cost for a number of years, followed by 2 percent over several more
years, followed by 1 percent indefinitely. In a few cases, larger decreases of 20 percent are
applied every 2 years during the initial years of production before learning decreases to the more
typical levels described above. This occurs for the changes that involve relatively new emerging
technologies that are not yet mature enough to warrant the slower learning rates.

   Table 5.183 lists the various learning schedules that NHTSA applies to technologies for the
2017-2025 FRIA. The  schedules are identified by a reference schedule number that was
originally  assigned to each schedule during the development of the agencies learning
methodology. Many other schedules were originally developed, but only those shown in Table
5.183 were considered relevant to the technology costs used in the current analysis. The table
illustrates  cost reduction rates for years 2015 through 2030. However, only a subset of these
years is relevant to each technology, depending on the year in which its direct cost estimate is
                                            5-435

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
based and the years in which the technology is applied. The learning rates that are indicated prior
to the direct manufacturing costs

    base year reflect "prior learning" that was estimated to occur before the base year direct
manufacturing cost estimate used by the agencies were developed. So, for example, if a cost
estimate for a mature technology reflects expected conditions in MY 2017, there would have
already been learning prior to that which would have impacted the MY 2017 costs. Additional
learning would then commence in MY 2018.
         Table 5.183 Learning Schedules by Model Year Applied to Specific CAFE Technologies

Model
Year
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Schedule #
6
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
11
0.913
0.885
0.868
0.850
0.833
0.817
0.800
0.784
0.769
0.753
0.738
0.731
0.723
0.716
0.709
0.702
12
1.000
0.970
0.951
0.932
0.913
0.895
0.877
0.859
0.842
0.825
0.809
0.801
0.793
0.785
0.777
0.769
16
1.000
1.000
0.970
0.941
0.913
0.885
0.859
0.833
0.808
0.784
0.760
0.745
0.730
0.716
0.701
0.687
18
1.000
1.000
1.000
1.000
0.800
0.800
0.640
0.640
0.627
0.615
0.602
0.590
0.579
0.567
0.556
0.544
19
2.441
1.953
1.953
1.563
1.563
1.250
1.250
1.250
1.250
1.250
1.000
0.970
0.941
0.913
0.885
0.859
21
1.063
1.031
1.000
0.970
0.941
0.913
0.885
0.859
0.842
0.825
0.808
0.792
0.776
0.768
0.761
0.753
24
1.250
1.000
1.000
0.970
0.941
0.913
0.885
0.859
0.833
0.808
0.784
0.768
0.753
0.738
0.723
0.708
25
1.563
1.563
1.563
1.563
1.250
1.250
1.000
0.970
0.941
0.913
0.885
0.859
0.842
0.825
0.808
0.792
26
1.146
1.114
1.095
1.065
1.029
1.000
0.973
0.944
0.920
0.898
0.876
0.859
0.842
0.827
0.812
0.798
30
0
0
0
3
3
3
3
3
3
3
3
3
3
3
3
3
31
0
0
0
0
0
0
0
0
5
5
5
5
5
5
5
5
   Table 5.184 lists the technologies that manufacturers may use to achieve higher CAFE levels,
and the learning schedule that is applied to each technology. Selection of specific learning
curves was based on the agency's best judgment as to the maturity of each technology and where
they would best fit along the learning curve, as well as the year on which their direct
manufacturing costs are based.

   For example, schedules 11, 12, and 21 are appropriate for technologies that are more mature
and have already passed through the steep portion of the learning curve, while schedules 16, 19,
24, and 25 are more appropriate for emerging technologies that will be experiencing learning
along the steep part of the curve between MYs 2014-2025.
                                             5-436

-------
                                     Technology Cost, Effectiveness, and Lead-Time Assessment
                       Table 5.184 Learning Schedules for Specific CAFE Technologies
Technology                                                                                 Learning
                                                                                           Schedule
Low Friction Lubricants- Level 1                                                                6
Engine Friction Reduction - Level 1                                                              6
Low Friction Lubricants and Engine Friction Reduction - Level 2                                      6
Variable Valve Timing (VVT) - Coupled Cam Phasing (CCP) on SOHC                                   12
Discrete Variable Valve Lift (DVVL) on SOHC                                                      12
Cylinder Deactivation on SOHC                                                                 11
Variable Valve Timing (VVT) - Intake Cam Phasing (ICP)                                             12
Variable Valve Timing (VVT) - Dual Cam Phasing (DCP)                                              12
Discrete Variable Valve Lift (DVVL) on DOHC                                                      12
Continuously Variable Valve Lift (CVVL)                                                          12
Cylinder Deactivation on DOHC                                                                 11
Stoichiometric Gasoline Direct Injection (GDI)                                                     11
Cylinder Deactivation on OHV                                                                  12
Variable Valve Actuation - CCP and DVVL on OHV                                                  12
Stoichiometric Gasoline Direct Injection (GDI) on OHV                                              11
Turbocharging and Downsizing - Level 1 (18 bar BMEP) -Small Displacement-Turbo                    11
Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Small Displacement - Downsize                 11
Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Medium Displacement-Turbo                  11
Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Medium Displacement - Downsize               11
Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Large Displacement-Turbo                    11
Turbocharging and Downsizing - Level 1 (18 bar BMEP) - Large Displacement - Downsize                 11
Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Small Displacement - Turbo                    11
Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Small Displacement - Downsize                 11
Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Medium Displacement - Turbo                 11
Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Medium Displacement - Downsize               11
Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Large Displacement - Turbo                    11
Turbocharging and Downsizing - Level 2 (24 bar BMEP) - Large Displacement - Downsize                 11
Cooled Exhaust Gas Recirculation (EGR) - Level 1 (24 bar BMEP) - Small Displacement  - Turbo            11
Cooled Exhaust Gas Recirculation (EGR) - Level 1 (24 bar BMEP) - Small Displacement - Downsize         11
Cooled Exhaust Gas Recirculation (EGR) - Level 1 (24 bar BMEP) - Medium Displacement - Turbo          11
Cooled Exhaust Gas Recirculation (EGR) - Level 1 (24 bar BMEP) - Medium Displacement- Downsize       11
Cooled Exhaust Gas Recirculation (EGR) - Level 1 (24 bar BMEP) - Large Displacement-Turbo            11
Cooled Exhaust Gas Recirculation (EGR) - Level 1 (24 bar BMEP) - Large Displacement- Downsize         11
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Small Displacement -Turbo            11
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Small Displacement - Downsize         11
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Medium Displacement-Turbo          11
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Medium Displacement- Downsize       11
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Large Displacement-Turbo            11
Cooled Exhaust Gas Recirculation (EGR) - Level 2 (27 bar BMEP) - Large Displacement- Downsize         11
Advanced Diesel - Small Displacement                                                           11
Advanced Diesel - Medium Displacement                                                        11
Advanced Diesel - Large Displacement                                                           11
6-Speed Manual/Improved Internals                                                            12
Improved Auto. Trans. Controls/Externals                                                        12
6-Speed Trans with Improved Internals (Auto)                                                     11
                                                     5-437

-------
                                      Technology Cost, Effectiveness, and Lead-Time Assessment
6-speed DCT                                                                                  11
8-Speed Trans (Auto or DCT)                                                                     11
High Efficiency Gearbox w/ dry sump (Auto or DCT)                                                 21
Shift Optimizer                                                                                21
Electric Power Steering                                                                         12
Improved Accessories - Level 1                                                                   12
Improved Accessories - Level 2 (w/ Alternator Regen and 70% efficient alternator)                       12
12V Micro-Hybrid (Stop-Start)                                                                    16
Integrated Starter Generator                                                                     16
Strong Hybrid (Powersplit or 2-Mode) - Level 1 - Battery                                            24
Strong Hybrid (Powersplit or 2-Mode) - Level 1 - Non-Battery                                         11
Conversion from SHEV1 to SHEV2                                                                N/A
Strong Hybrid (P2 Parallel or 2-Mode) - Level 2 - Battery                                            24
Strong Hybrid (P2 Parallel or 2-Mode) - Level 2 - Non-Battery                                         11
Plug-in Hybrid - 20 mi range - Battery                                                             19
Plug-in Hybrid - 20 mi range - Non-Battery                                                         11
Plug-in Hybrid - 40 mi range - Battery                                                             19
Plug-in Hybrid - 40 mi range - Non-Battery                                                         11
Electric Vehicle (Early Adopter) - 75 mile range - Battery                                            19
Electric Vehicle (Early Adopter) - 75 mile range - Non-Battery                                         21
Electric Vehicle (Early Adopter) -100 mile range - Battery                                            19
Electric Vehicle (Early Adopter) -100 mile range - Non-Battery                                        21
Electric Vehicle (Early Adopter) -150 mile range - Battery                                            19
Electric Vehicle (Early Adopter) -150 mile range - Non-Battery                                        21
Electric Vehicle (Broad Market) -150 mile range - Battery                                            19
Electric Vehicle (Broad Market) -150 mile range - Non-Battery                                        21
Charger-PHEV20                                                                               19
Charger-PHEV40                                                                               19
Charger-EV                                                                                   19
Charger Labor                                                                                 6
Mass Reduction - Level 1                                                                        21
Mass Reduction - Level 2                                                                        21
Mass Reduction - Level 3                                                                        21
Mass Reduction - Level 4                                                                        21
Mass Reduction - Level 5                                                                        21
Low Rolling Resistance Tires - Level 1                                                             6
Low Rolling Resistance Tires - Level 2                                                             25
Low Rolling Resistance Tires - Level 3                                                             N/A
Low Drag Brakes                                                                               6
Secondary Axle Disconnect                                                                      12
Aero Drag Reduction, Level 1                                                                     12
Aero Drag Reduction, Level 2                                                                     12
   5.4.1.3 Technology Cost Summary Tables
      The following tables summarize incremental costs and total costs for advanced technologies
   in 2013 dollars. Incremental costs reflect the additional costs that the Volpe model applies over
   the previous step in the technology track for a specific piece of technology.  Absolute costs
                                                      5-438

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
reflect cost to add an advanced technology and requisite enabling technologies over the low
technology baseline in the technology path.

   The following cost tables show the combined results of direct manufacturing costs, indirect
costs, learning curves, and technology progression paths for 2017MY and 2025 MY. To
calculate direct manufacturing costs for a given year from the costs listed in these tables, divide
by the RPE (1.5) and adjust for the appropriate learning schedule factors as well as incremental
costs for removing technologies that are no longer needed. The costs for all years are relevant
inputs for the CAFE model.

   Many technologies have projected costs that vary by application. For instance, the
incremental cost of many engine technologies takes into account the engine configuration, like
number of banks and number cylinders. Similarly, many advanced vehicle technologies have a
specific cost for each vehicle class.  The following tables summarize the costs for CAFE model
inputs by application.

5.4.1.3.1      Basic Gasoline Engine Costs

   This  section shows projected costs for basic gasoline engine technologies.  Table 5.185
demonstrates how technology costs may scale with application attributes.  Table 5.186 and Table
5.188 show incremental and absolute costs for advanced technologies on the basic engine path.
                                             5-439

-------
                                                                 Technology Cost, Effectiveness, and Lead-Time Assessment
                         Table 5.185 Examples of Engine Technology Costs that Scale with Engine Attributes
Gasoline Engine Technologies - Direct Manufacturer Costs - Small Displacement DOHC
Tech
LUBEFR1
LUBEFR2
LUBEFR3
VVT
VVL
SGDI
DEAC
HCR
Basis
cylinder
cylinder
cylinder
bank
cylinder
cylinder
none
none
UnitDMC
$ 13.36
$ 0.89
$ 0.89
$ 75.20
$ 51.31
$ 56.76
$ 28.19
$ 90.85 $
215.5
Learning
Factor
6
30
31
12
12
11
11
21
DMC for
4-Cylinder
1-Bank Engine
$ 53.45
$ 3.55
$ 3.55
$ 75.20
$ 205.26
$ 227.05
$ 28.19
$ 90.85
DMC for
4-Cylinder
2-Bank
EngineYYY
$ 53.45
$ 3.55
$ 3.55
$ 150.40
$ 205.26
$ 227.05
$ 28.19
$ 90.85
DMC for
6-Cylinder
1-Bank Engine
$ 80.18
$ 5.32
$ 5.32
$ 75.20
$ 307.88
$ 340.57
$ 28.19
$ 136.27
DMC for
6-Cylinder
2-Bank Engine
$ 80.18
$ 5.32
$ 5.32
$ 150.40
$ 307.88
$ 340.57
$ 28.19
$ 136.27
DMC for
8-Cylinder
2-Bank Engine
$ 106.91
$ 7.10
$ 7.10
$ 150.40
$ 410.51
$ 454.09
$ 28.19
$ 215.50
Illustrative example for cost calculation purposes, only.
                                                              5-440

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
            Table 5.186 Projected MY2017 Incremental Costs for Gasoline Engine Technology
Technology
LUBEFR1
LUBEFR2
LUBEFR3
AAAA
VVT
VVL
SGDI
DEAC
HCR
4-Cylinder
1-Bank
80.18
-
-
107.23
292.67
295.47
36.69
136.27
4-Cylinder
2-Bankzzz
80.18
-
-
214.46
292.67
295.47
36.69
136.27
6-Cylinder
1-Bank
120.27
-
-
107.23
439.01
443.21
36.69
204.41
6-Cylinder
2-Bank
120.27
-
-
214.46
439.01
443.21
36.69
204.41
8-Cylinder
2-Bank
160.36
-
-
214.46
585.35
590.95
36.69
323.26
Cost
Adjustment
SOHC
(per basis)



(30.68)
(17.24)



Cost
Adjustment
OHV
(per basis)



(30.68)
(17.24)



            Table 5.187 Projected MY2025 Incremental Costs for Gasoline Engine Technology
Technology
LUBEFR1
LUBEFR2
LUBEFR3
VVT
VVL
SGDI
DEAC
HCR
4-Cylinder
1-Bank
80.18
15.97
26.62
93.09
254.08
256.51
31.85
112.39
4-Cylinder
2-BankBBBB
80.18
15.97
26.62
186.17
254.08
256.51
31.85
112.39
6-Cylinder
1-Bank
120.27
23.96
39.93
93.09
381.12
384.76
31.85
168.58
6-Cylinder
2-Bank
120.27
23.96
39.93
186.17
381.12
384.76
31.85
168.58
8-Cylinder
2-Bank
160.36
31.95
53.24
186.17
508.15
513.02
31.85
266.60
Cost
Adjustment
SOHC
(per basis)



(30.68)
(17.24)



Cost
Adjustment
OHV
(per basis)



(30.68)
(17.24)



zzz Illustrative example for cost calculation purposes, only.
AAAA LUBEFR2 and LUBEFR3 are not available until MY2018 and MY2023, respectively.
BBBB illustrative example for cost calculation purposes, only.
                                                 5-441

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
             Table 5.188 Projected MY2017 Absolute Costs for Gasoline Engine Technology
Technology
LUBEFR1
LUBEFR2
LUBEFR3DDDD
VVT
VVL
SGDI
DEAC
HCR
4-Cylinder
1-Bank
80.18
80.18
80.18
187.41
480.08
775.56
812.25
948.52
4-Cylinder
2-Bankcccc
80.18
80.18
80.18
294.64
587.31
882.78
919.48
1,055.75
6-Cylinder
1-Bank
120.27
120.27
120.27
227.50
666.51
1,109.72
1,146.41
1,350.82
6-Cylinder
2-Bank
120.27
120.27
120.27
334.73
773.74
1,216.95
1,253.64
1,458.05
8-Cylinder
2-Bank
160.36
160.36
160.36
374.82
960.16
1,551.11
1,587.80
1,911.06
             Table 5.189 Projected MY2025 Absolute Costs for Gasoline Engine Technology
Technology
LUBEFR1
LUBEFR2
LUBEFR3
VVT
VVL
SGDI
DEAC
HCR
4-Cylinder
1-Bank
80.18
96.15
122.77
215.86
469.94
726.45
758.30
870.69
4-Cylinder
2-BankEEEE
80.18
96.15
122.77
308.95
563.03
819.53
851.39
963.78
6-Cylinder
1-Bank
120.27
144.23
184.16
277.25
658.36
1,043.13
1,074.98
1,243.56
6-Cylinder
2-Bank
120.27
144.23
184.16
370.34
751.45
1,136.21
1,168.07
1,336.65
8-Cylinder
2-Bank
160.36
192.31
245.55
431.72
939.88
1,452.90
1,484.75
1,751.35
cccc illustrative example for cost calculation purposes, only.
DDDD LUBEFR2 and LUBEFR3 are not available until MY2018 and MY2023, respectively.
EEEE illustrative example for cost calculation purposes, only.
                                                 5-442

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.1.3.2      Gasoline Turbo Engine Costs
      Table 5.190 Projected MY2017 Incremental Costs for Turbo and Turbo-Downsize Technology
Technology
Turbo-lSOHC
Downsize-lSOHC
Turbo-lSOHC
Downsize-lSOHC
Turbo-lSOHC
Downsize-lSOHC
Turbo-1 DOHC
Downsize-1
DOHC
Turbo-1 DOHC
Downsize-1
DOHC
Turbo-1 DOHC
Downsize-1
DOHC
Turbo-1 OHV
Downsize-1 OHV
Turbo-1 OHV
Downsize-1 OHV
Turbo-1 OHV
Downsize-1 OHV
Engin
e Type
SOHC
SOHC
SOHC
SOHC
SOHC
SOHC
DOHC
DOHC
DOHC
DOHC
DOHC
DOHC
OHV
OHV
OHV
OHV
OHV
OHV
Displacemen
t
Small
Small
Medium
Medium
Large
Large
Small
Small
Medium
Medium
Large
Large
Small
Small
Medium
Medium
Large
Large
Learnin
g Factor
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
Incrementa
ICost
577.57
-40.46
577.57
-182.22
973.57
-39.92
577.57
-40.46
577.57
-260.79
973.57
-130.53
0.00
0.00
577.57
386.00
973.57
450.33
Downsizing
Costs
Adjustmen
t

-36.69

-36.69

-36.69

-36.69

-36.69

-36.69

-36.69

-36.69

-36.69
Technolog
y Costs
After
Downsizing
577.57
-77.15
577.57
-218.91
973.57
-76.61
577.57
-77.15
577.57
-297.48
973.57
-167.22
0.00
-36.69
577.57
349.31
973.57
413.64
Incrementa
1 Combined
Tech Cost
500.42
358.66
896.96
500.42
280.08
806.35
-36.69
926.87
1,387.21
                                             5-443

-------
                         Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.191 Projected MY2025 Incremental Costs for Turbo and Turbo-Downsize Technology
Technology
Turbo-lSOHC
Downsize-lSOHC
Turbo-lSOHC
Downsize-lSOHC
Turbo-lSOHC
Downsize-lSOHC
Turbo-1 DOHC
Downsize-1
DOHC
Turbo-1 DOHC
Downsize-1
DOHC
Turbo-1 DOHC
Downsize-1
DOHC
Turbo-1 OHV
Downsize-1 OHV
Turbo-1 OHV
Downsize-1 OHV
Turbo-1 OHV
Downsize-1 OHV
Engin
eType
SOHC
SOHC
SOHC
SOHC
SOHC
SOHC
DOHC
DOHC
DOHC
DOHC
DOHC
DOHC
OHV
OHV
OHV
OHV
OHV
OHV
Displacemen
t
Small
Small
Medium
Medium
Large
Large
Small
Small
Medium
Medium
Large
Large
Small
Small
Medium
Medium
Large
Large
Learnin
g Factor
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
Incrementa
ICost
491.37
-34.42
491.37
-155.02
828.28
-33.96
491.37
-34.42
491.37
-221.87
828.28
-111.05
0.00
0.00
491.37
328.39
828.28
383.13
Downsizing
Costs
Adjustmen
t

-31.85

-31.85

-31.85

-31.85

-31.85

-31.85

-31.85

-31.85

-31.85
Technolog
y Costs
After
Downsizing
491.37
-66.27
491.37
-186.87
828.28
-65.82
491.37
-66.27
491.37
-253.73
828.28
-142.90
0.00
-31.85
491.37
296.54
828.28
351.27
Incrementa
1 Combined
Tech Cost
425.10
304.50
762.47
425.10
237.65
685.38
-31.85
787.91
1,179.55
                                       5-444

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.192 Projected MY2017 and MY2025 Absolute Costs for Turbo and Turbo-Downsizing Technology
Technology
Turbo-lSOHC
Downsize-lSOHC
Turbo-lSOHC
Downsize-lSOHC
Turbo-lSOHC
Downsize-lSOHC
Turbo-1 DOHC
Downsize-1 DOHC
Turbo-1 DOHC
Downsize-1 DOHC
Turbo-1 DOHC
Downsize-1 DOHC
Turbo-1 OHV
Downsize-1 OHV
Turbo-1 OHV
Downsize-1 OHV
Turbo-1 OHV
Downsize-1 OHV
Engine Type
SOHC
SOHC
SOHC
SOHC
SOHC
SOHC
DOHC
DOHC
DOHC
DOHC
DOHC
DOHC
OHV
OHV
OHV
OHV
OHV
OHV
Displacement
Small
Small
Medium
Medium
Large
Large
Small
Small
Medium
Medium
Large
Large
Small
Small
Medium
Medium
Large
Large
Learning
Factor
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
MY2017
Absolute
Combined
Tech Cost
1,419.89
1,505.07
2,150.60
1,419.89
1,426.50
2,059.99
882.78
2,073.29
2,640.85
MY2025
Absolute
Combined
Tech Cost
1,276.49
1,379.48
1,930.53
1,276.49
1,312.63
1,853.45
819.53
1,862.89
2,347.62
                                            5-445

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.1.3.3       Other Advanced Gasoline Engine Technologies

    Table 5.193 Direct Manufacturing Costs and Learning Schedules for Advanced Engine Technologies
Technology
SEGR
DWSP
TURBO2
CEGR1
CEGR1P
CEGR2
HCR2
DMC
(Small Displacement)
307.20
54.93
9.89
255.59
0.00
443.85
28.19
DMC
(Medium Displacement)
307.20
54.93
221.91
255.59
0.00
443.85
28.19
DMC
(Large Displacement)
307.20
54.93
374.05
255.59
0.00
443.85
28.19
Learning
Schedule
11
11
11
11
11
11
11
      Table 5.194 Projected MY2017 Incremental Costs for Advanced Gasoline Engine Technologies
Technology
SEGR
DWSP
TURBO2
CEGR1
CEGR1P
CEGR2
HCR2
4-Cylinder
1-Bank
399.78
71.48
196.81
332.62
0.00
577.62
36.69
4-Cylinder
2-BankFFFF
399.78
71.48
196.81
332.62
0.00
577.62
36.69
6-Cylinder
1-Bank
399.78
71.48
288.78
332.62
0.00
577.62
36.69
6-Cylinder
2-Bank
399.78
71.48
288.78
332.62
0.00
577.62
36.69
8-Cylinder
2-Bank
399.78
71.48
486.79
332.62
0.00
577.62
36.69
      Table 5.195 Projected MY2025 Incremental Costs for Advanced Gasoline Engine Technologies
Technology
SEGR
DWSP
TURBO2
CEGR1
CEGR1P
CEGR2
HCR2
4-Cylinder
1-Bank
347.06
62.05
170.86
288.76
0.00
501.45
31.85
4-Cylinder
2-BankGGGG
347.06
62.05
170.86
288.76
0.00
501.45
31.85
6-Cylinder
1-Bank
347.06
62.05
250.70
288.76
0.00
501.45
31.85
6-Cylinder
2-Bank
347.06
62.05
250.70
288.76
0.00
501.45
31.85
8-Cylinder
2-Bank
347.06
62.05
422.59
288.76
0.00
501.45
31.85
FFFF Illustrative example for cost calculation purposes, only.
GGGG niustrative example for cost calculation purposes, only.
                                                5-446

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
        Table 5.196 Projected MY2017 Absolute Costs for Advanced Gasoline Engine Technologies
Technology
SEGR
DWSP
TURBO2
CEGR1
CEGR1P
CEGR2
HCR2
4-Cylinder
1-Bank
1,819.67
1,891.16
2,087.97
2,420.59
2,420.59
2,998.21
3,034.91
4-Cylinder
2-BankHHHH
1,819.67
1,891.16
2,087.97
2,420.59
2,420.59
2,998.21
3,034.91
6-Cylinder
1-Bank
1,826.28
1,897.76
2,186.54
2,519.17
2,519.17
3,096.79
3,133.48
6-Cylinder
2-Bank
1,826.28
1,897.76
2,186.54
2,519.17
2,519.17
3,096.79
3,133.48
8-Cylinder
2-Bank
2,459.77
2,531.25
3,018.04
3,350.67
3,350.67
3,928.29
3,964.98
        Table 5.197 Projected MY2025 Absolute Costs for Advanced Gasoline Engine Technologies
Technology
SEGR
DWSP
TURBO2
CEGR1
CEGR1P
CEGR2
HCR2
4-Cylinder
1-Bank
1,623.55
1,685.60
1,856.46
2,145.22
2,145.22
2,646.67
2,678.52
4-Cylinder
2-Bank1111
1,623.55
1,685.60
1,856.46
2,145.22
2,145.22
2,646.67
2,678.52
6-Cylinder
1-Bank
1,659.69
1,721.74
1,972.44
2,261.20
2,261.20
2,762.65
2,794.51
6-Cylinder
2-Bank
1,659.69
1,721.74
1,972.44
2,261.20
2,261.20
2,762.65
2,794.51
8-Cylinder
2-Bank
2,200.51
2,262.56
2,685.15
2,973.91
2,973.91
3,475.36
3,507.21
HHHH niustrative example for cost calculation purposes, only.
1111 Illustrative example for cost calculation purposes, only.
                                                 5-447

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.1.3.4      Diesel Engine Costs
          Table 5.198 Projected MY2017 Incremental Costs for Diesel Engines by Engine Type
Technology
ADSL
TURBODSL
DWSPDSL
EFRDSL
CLCDSL
LPEGRDSL
DSIZEDSL
4 Cylinder
3977.96
26.94
44.37
101.41
107.75
210.75
0.00
6 Cylinder
4395.41
26.94
44.37
152.12
161.63
263.04
0.00
8 Cylinder
4425.70
26.94
44.37
152.12
161.63
263.04
0.00
          Table 5.199 Projected MY2025 Incremental Costs for Diesel Engines by Engine Type
Technology
ADSL
TURBODSL
DWSPDSL
EFRDSL
CLCDSL
LPEGRDSL
DSIZEDSL
4 Cylinder
3192.32
22.22
36.59
83.64
88.87
173.81
0.00
6 Cylinder
3490.87
22.22
36.59
125.46
133.30
216.94
0.00
8 Cylinder
3474.76
22.22
36.59
125.46
133.30
216.94
0.00
                                              5-448

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
           Table 5.200 Projected MY2017 Absolute Costs for Diesel Engines by Engine Type
Technology
ADSL
TURBODSL
DWSPDSL
EFRDSL
CLCDSL
LPEGRDSL
DSIZEDSL
4 Cylinder
3977.96
4004.90
4049.27
4150.68
4258.43
4469.18
4469.18
6 Cylinder
4395.41
4422.35
4466.72
4618.84
4780.47
5043.51
5043.51
8 Cylinder
4425.70
4452.64
4497.01
4649.13
4810.76
5073.80
5073.80
           Table 5.201 Projected MY2025 Absolute Costs for Diesel Engines by Engine Type
Technology
ADSL
TURBODSL
DWSPDSL
EFRDSL
CLCDSL
LPEGRDSL
DSIZEDSL
4 Cylinder
3192.32
3214.53
3251.13
3334.77
3423.63
3597.44
3597.44
6 Cylinder
3490.87
3513.08
3549.67
3675.13
3808.43
4025.37
4025.37
8 Cylinder
3474.76
3496.98
3533.57
3659.03
3792.33
4009.26
4009.26
5.4.1.3.5
Transmission Costs
   The transmission technology paths for manual and automatic transmissions are separate.
Emerging advanced transmissions have learning schedules with greater opportunity for future
cost reduction than the learning schedules for transmissions that have been widely used for many
years.
                                              5-449

-------
                           Technology Cost, Effectiveness, and Lead-Time Assessment
      Table 5.202 Direct Manufacturing Costs and Learning Schedules for Transmissions
Transmission
MT5
MT6
MT7
ATS
AT6
AT6P
ATS
AT8P
DCT6
DCT8
CVT
Direct manufacturing Cost
0.00
247.42
239.10
0.00
-13.73
0.00
82.18
194.00
32.75
239.10
189.09
Learning Factor
12
12
11
12
21
21
11
21
21
11
21
Table 5.203  Projected MY2017 Incremental Costs for Transmission Technologies by Vehicle Class
Transmission
MT5
MT6
MT7
ATS
AT6
AT6P
ATS
AT8P
DCT6
DCT8
CVT
SmallCar
0.00
352.80
311.16
0.00
-6.87
0.00
106.95
291.00
49.12
311.16
283.64
MedCar
0.00
352.80
311.16
0.00
-6.87
0.00
106.95
291.00
49.12
311.16
283.64
SmallSUV
0.00
352.80
311.16
0.00
-6.87
0.00
106.95
291.00
49.12
311.16
283.64
MedSUV
0.00
352.80
311.16
0.00
-6.87
0.00
106.95
291.00
49.12
311.16
283.64
Pickup
0.00
352.80
311.16
0.00
-6.87
0.00
106.95
291.00
49.12
311.16
283.64
                                          5-450

-------
                            Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.204 Projected MY2025 Incremental Costs for Transmission Technologies by Vehicle Class
Transmission
MT5
MT6
MT7
ATS
AT6
AT6P
ATS
AT8P
DCT6
DCT8
CVT
SmallCar
0.00
300.15
264.72
0.00
-5.55
0.00
90.99
235.20
39.70
264.72
229.25
MedCar
0.00
300.15
264.72
0.00
-5.55
0.00
90.99
235.20
39.70
264.72
229.25
SmallSUV
0.00
300.15
264.72
0.00
-5.55
0.00
90.99
235.20
39.70
264.72
229.25
MedSUV
0.00
300.15
264.72
0.00
-5.55
0.00
90.99
235.20
39.70
264.72
229.25
Pickup
0.00
300.15
264.72
0.00
-5.55
0.00
90.99
235.20
39.70
264.72
229.25
 Table 5.205 Projected MY2017 Absolute Costs for Transmission Technologies by Vehicle Class
Transmission
MT5
MT6
MT7
ATS
AT6
AT6P
ATS
AT8P
DCT6
DCT8
CVT
SmallCar
0.00
352.80
663.95
0.00
-6.87
-6.87
100.08
391.08
49.12
360.28
283.64
MedCar
0.00
352.80
663.95
0.00
-6.87
-6.87
100.08
391.08
49.12
360.28
283.64
SmallSUV
0.00
352.80
663.95
0.00
-6.87
-6.87
100.08
391.08
49.12
360.28
283.64
MedSUV
0.00
352.80
663.95
0.00
-6.87
-6.87
100.08
391.08
49.12
360.28
283.64
Pickup
0.00
352.80
663.95
0.00
-6.87
-6.87
100.08
391.08
49.12
360.28
283.64
                                           5-451

-------
                         Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.206 Projected MY2025 Absolute Costs for Transmission Technologies by Vehicle Class
Transmission
MT5
MT6
MT7
ATS
AT6
AT6P
ATS
AT8P
DCT6
DCT8
CVT
SmallCar
0.00
300.15
564.87
0.00
-5.55
-5.55
85.44
320.64
39.70
304.42
229.25
MedCar
0.00
300.15
564.87
0.00
-5.55
-5.55
85.44
320.64
39.70
304.42
229.25
SmallSUV
0.00
300.15
564.87
0.00
-5.55
-5.55
85.44
320.64
39.70
304.42
229.25
MedSUV
0.00
300.15
564.87
0.00
-5.55
-5.55
85.44
320.64
39.70
304.42
229.25
Pickup
0.00
300.15
564.87
0.00
-5.55
-5.55
85.44
320.64
39.70
304.42
229.25
                                        5-452

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.1.3.6
Electric Vehicle and Accessory Costs
Table 5.207 Direct Manufacturing Costs and Learning Schedules for Electric Vehicle and Accessory Systems
                                  by Vehicle Technology Class
Technology
EPS
IACC1
IACC2
SS12V
BISG
CISC
SHEVP2
SHEVP2_battery
SHEVP2_non-battery
SHEVPS
SHEVPS_battery
SHEVPS_non-battery
PHEV30
PHEV30_battery
PHEV30_non-battery
PHEV30 C
CHRG_L
PHEV50
PHEV50_battery
PHEV50_non-battery
PHEV50_C
CHRG L
EV200
EV200_battery
EV200_non-battery
EV C
CHRG L
FCV
Learning
Factor
12
12
12
16
24
18

24
11

24
11

19
11
19
6

19
11
19
6

19
21
19
6
26
SmallCar
95.86
77.96
48.12
273.49
1,013.00
2,121.23

783.27
1,799.26

783.27
1,799.26

3,365.03
3,156.00
177.50
1,000.00

4,594.74
3,156.00
195.68
1,000.00

8,733.63
406.34
213.86
1,000.00
15,566.29
MedCar
95.86
77.96
48.12
300.29
1,013.00
2,684.29

1,015.74
2,378.46

1,015.74
2,378.46

5,330.99
5,716.96
177.50
1,000.00

7,838.02
5,716.96
195.68
1,000.00

12,048.97
2,214.28
213.86
1,000.00
15,566.29
SmallSUV
95.86
77.96
48.12
322.52
1,013.00
2,695.91

843.17
1,936.34

843.17
1,936.34

3,894.70
3,656.28
177.50
1,000.00

5,408.55
3,656.28
195.68
1,000.00

10,741.72
132.17
213.86
1,000.00
15,566.29
MedSUV
95.86
77.96
48.12
330.44
1,162.72
3,293.83

938.71
2,217.03

938.71
2,217.03

5,330.99
5,716.96
177.50
1,000.00

7,838.02
5,716.96
195.68
1,000.00

12,048.97
2,214.28
213.86
1,000.00
15,566.29
Pickup
95.86
77.96
48.12
373.60
1,277.30
3,293.83

1,089.51
2,339.96

1,089.51
2,339.96

5,330.99
5,716.96
177.50
1,000.00

7,838.02
5,716.96
195.68
1,000.00

12,048.97
2,214.28
213.86
1,000.00
15,566.29
                                               5-453

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.208 Projected MY2017 Incremental Costs for Electric Vehicle and Accessory Systems by Vehicle
                                           Class
Technology
EPS
IACC1
IACC2
SS12V
BISG
CISC
SHEVP2
SHEVPS
PHEV30
PHEV50
EV200
FCV
SmallCar
136.69
111.17
68.61
397.93
805.10
2,467.45
1,996.92
334.57
12,469.24
5,675.96
10,754.55
7,999.51
MedCar
136.69
111.17
68.61
436.92
766.12
3,273.05
3,099.39
592.45
20,459.21
9,418.11
10,344.76
-4,835.07
SmallSUV
136.69
111.17
68.61
469.27
733.77
3,258.13
2,265.16
-259.20
14,403.82
6,508.38
13,191.12
4,964.27
MedSUV
136.69
111.17
68.61
480.79
946.82
4,143.48
2,549.19
-647.47
20,784.83
9,418.11
10,344.76
-4,835.07
Pickup
136.69
111.17
68.61
543.59
1,055.89
4,080.68
2,763.50
-261.29
20,398.65
9,418.11
10,344.76
-4,835.07
Table 5.209 Projected MY2025 Incremental Costs for Electric Vehicle and Accessory Systems by Vehicle
                                           Class
Technology
EPS
IACC1
IACC2
SS12V
BISG
CISC
SHEVP2
SHEVPS
PHEV30
PHEV50
EV200
FCV
SmallCar
116.29
94.58
58.37
311.88
609.78
1,335.51
1,722.00
996.27
7,395.11
3,638.09
5,027.54
10,056.47
MedCar
116.29
94.58
58.37
342.43
579.23
1,813.71
2,636.57
1,402.09
12,264.90
5,554.07
4,492.06
2,356.14
SmallSUV
116.29
94.58
58.37
367.79
553.88
1,798.85
1,944.19
699.21
8,521.32
4,064.29
5,932.68
8,281.87
MedSUV
116.29
94.58
58.37
376.82
720.86
2,330.06
2,191.27
582.07
12,534.18
5,554.07
4,492.06
2,356.14
Pickup
116.29
94.58
58.37
426.04
806.34
2,280.85
2,369.96
895.45
12,220.79
5,554.07
4,492.06
2,356.14
                                              5-454

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
Table 5.210 Projected MY2017 Absolute Costs for Electric Vehicle and Accessory Systems by Vehicle Class
Technology
EPS
IACC1
IACC2
SS12V
BISG
CISC
SHEVP2
SHEVPS
PHEV30
PHEV50
EV200
FCV
SmallCar
136.69
247.85
316.46
714.40
1,519.50
3,181.85
3,516.42
3,516.42
15,985.66
21,661.62
32,416.17
40,415.68
MedCar
136.69
247.85
316.46
753.38
1,519.50
4,026.44
4,618.89
4,618.89
25,078.10
34,496.20
44,840.97
40,005.89
SmallSUV
136.69
247.85
316.46
785.73
1,519.50
4,043.87
3,784.66
3,784.66
18,188.48
24,696.86
37,887.98
42,852.25
MedSUV
136.69
247.85
316.46
797.26
1,744.08
4,940.74
4,293.27
4,293.27
25,078.10
34,496.20
44,840.97
40,005.89
Pickup
136.69
247.85
316.46
860.06
1,915.95
4,940.74
4,679.45
4,679.45
25,078.10
34,496.20
44,840.97
40,005.89
Table 5.211 Projected MY2025 Absolute Costs for Electric Vehicle and Accessory Systems by Vehicle Class
Technology
EPS
IACC1
IACC2
SS12V
BISG
CISC
SHEVP2
SHEVPS
PHEV30
PHEV50
EV200
FCV
SmallCar
116.29
210.86
269.24
581.11
1,190.90
1,916.63
2,912.90
2,912.90
10,308.01
13,946.10
18,973.64
29,030.11
MedCar
116.29
210.86
269.24
611.67
1,190.90
2,425.38
3,827.47
3,827.47
16,092.36
21,646.43
26,138.49
28,494.63
SmallSUV
116.29
210.86
269.24
637.02
1,190.90
2,435.88
3,135.09
3,135.09
11,656.41
15,720.70
21,653.38
29,935.25
MedSUV
116.29
210.86
269.24
646.06
1,366.91
2,976.12
3,558.19
3,558.19
16,092.36
21,646.43
26,138.49
28,494.63
Pickup
116.29
210.86
269.24
695.27
1,501.61
2,976.12
3,871.57
3,871.57
16,092.36
21,646.43
26,138.49
28,494.63
                                               5-455

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.1.3.7      Vehicle Technology Costs
        Table 5.212 Direct Manufacturing Costs and Learning Schedules for Vehicle Technologies
Technology
ROLL10
ROLL20
LDB
SAX
AERO10
AERO20
MR1
MR2
MRS
MR4
MRS
Direct Manufacturing Costs
5.64
42.77
62.03
85.57
42.86
128.57
Refer to and Table 5.176 in the previous
Mass Reduction section of the CAFE
technology assessment.
Also, refer to Figure 5.144 and Table
5.178.
Learning Factor
6
25
6
12
12
12
21
21
21
21
21
              Table 5.213 Projected MY2017 Incremental Costs for Vehicle Technologies
Technology
ROLL1
ROLL2
LDB
SAX
AERO1
AERO2
SmallCar
8.46
100.25
93.04
122.01
61.11
183.32
MedCar
8.46
100.25
93.04
122.01
61.11
183.32
SmallSUV
8.46
100.25
93.04
122.01
61.11
183.32
MedSUV
8.46
100.25
93.04
122.01
61.11
183.32
Pickup
8.46
100.25
93.04
122.01
61.11
183.32
              Table 5.214 Projected MY2025 Incremental Costs for Vehicle Technologies
Technology
ROLL10
ROLL20
LDB
SAX
AERO10
AERO20
SmallCar
8.46
56.80
93.04
103.80
51.99
155.96
MedCar
8.46
56.80
93.04
103.80
51.99
155.96
SmallSUV
8.46
56.80
93.04
103.80
51.99
155.96
MedSUV
8.46
56.80
93.04
103.80
51.99
155.96
Pickup
8.46
56.80
93.04
103.80
51.99
155.96
                                                5-456

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
               Table 5.215 Projected MY2017 Absolute Costs for Vehicle Technologies
Technology
ROLL10
ROLL20
LDB
SAX
AERO10
AERO20
SmallCar
8.46
108.70
93.04
122.01
61.11
244.43
MedCar
8.46
108.70
93.04
122.01
61.11
244.43
SmallSUV
8.46
108.70
93.04
122.01
61.11
244.43
MedSUV
8.46
108.70
93.04
122.01
61.11
244.43
Pickup
8.46
108.70
93.04
122.01
61.11
244.43
               Table 5.216 Projected MY2025 Absolute Costs for Vehicle Technologies
Technology
ROLL10
ROLL20
LDB
SAX
AERO10
AERO20
SmallCar
8.46
65.26
93.04
103.80
51.99
207.95
MedCar
8.46
65.26
93.04
103.80
51.99
207.95
SmallSUV
8.46
65.26
93.04
103.80
51.99
207.95
MedSUV
8.46
65.26
93.04
103.80
51.99
207.95
Pickup
8.46
65.26
93.04
103.80
51.99
207.95
5.4.2   Technology Effectiveness Modeling Method and Data Used in CAFE Assessment

   This section provides an overview of Argonne National Laboratory simulation modeling
conducted to estimate energy consumption reductions from combinations of light-duty
powertrain and vehicle technologies. The modeling work was conducted under contract to
NHTSA and provides inputs to DOT Volpe's CAFE Compliance and Effects Modeling System
(commonly referred to as the Volpe Model) for light- and medium-duty vehicles.602 >603' The
section provides a description of baseline vehicles, model validation, technology assumptions,
and methodology.

   For this TAR, NHTSA is employing a world recognized full vehicle simulation model
Autonomie developed by Argonne National Laboratory over the past 15 years under funding
from the US DOE Vehicle Technologies Office. Autonomie has been developed and validated
over a very wide range of powertrain  configurations and component technologies leveraging
vehicle test data from Argonne Advanced Powertrain Research Facility (APRF) and component
performance data from the US National Laboratories, including Oak Ridge National Laboratory
(ORNL), Idaho National Laboratory (INL) and the National Renewable National Laboratory
(NREL). Using Autonomie will not only improve  the transparency of the process, but also
increase the robustness of the process by simulating every single combination of individual
technologies. Input data for Autonomie has been created through a combination of benchmarking
activities and high fidelity component modeling. Benchmarking is a commonly used technique
that is intended to create a detailed characterization of a vehicle's operation and performance.
                                            5-457

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.1 Volpe Model Background

   The Volpe model combines technologies in sequence dictated by what are referred to as
"decision trees." In the model there are seven vehicle classes and eight decision trees for each
class. The decision trees include the following  sub-systems: engine; transmission; powertrain
electrification; hybridization; light-weighting; aerodynamics; rolling resistance; and dynamic
load reduction. Each of the sub-systems is evaluated independently of each other, starting with
the top-most technology and progressing down the decision tree. Figure 5.147 shows the
decision trees for basic engine and transmission technologies.
                Basic Engine Paths
Basic Engine Path
SOHC
\l






DEAC


DOHC
I
WT
4
WL
I
SGDI
l/\



OHV
I/






HCR

Transmission Paths
                                                     Manual Transmission Path
                                                     Automatic Transmission Path
                 Figure 5.147 Volpe Model Engine and Transmission Decision Trees

   In past rulemakings, the model relied on estimates of synergies between technologies,
recognizing that multiple technologies can address the same inefficiency. An example of this is a
combination of variable valve timing, cylinder deactivation, and 6-speed automatic transmission
technologies. For a specific vehicle platform, each technology individually offers a reduction in
energy consumption. However, when modeled in combination, the package provides a reduction
that is somewhat less than the sum of the individual technology benefits. The reason for this is
that each of the three technologies reduces a portion of the throttling loss encountered at part
loads. When a portion of the loss has been addressed by a technology, the loss has been
eliminated and cannot be reduced by another technology. In some cases, combining technologies
may produce fuel savings that are greater than the sum of the savings from the two technologies
- or positive synergies. The synergy factor used previously in the Volpe model estimated the
extent to which combinations of technologies result in less than additive (negative synergies) or
more than additive (positive synergies) energy  consumption savings. Synergy factors used in the
Volpe model for prior rulemakings were based on engineering judgement of the impact on
energy consumption from the combination of technologies.

   To more accurately estimate the impact on light-duty energy consumption of combined
powertrain and vehicle technologies in the Volpe model, NHTSA contracted with Argonne
National Laboratory to simulate powertrain and vehicle technology combinations as shown in
Figure 2. Modeling conducted for the light-duty MTE Draft TAR is the first time the results of
the Argonne simulation results have been used directly in the Volpe model.
                                             5-458

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                 Current Process
        DECISION TREES
                                    LU
                                    Q
                                    o

                                    ui
                                    O
                                                            New Process
                                               Large Scale Simulation Process

                                               Individual Vehicle Simulation
                                                Q
                                                O
                                                u
                                                "v.
                                                LLJ
                                                CL
                                                _l
                                                O
     Figure 5.148 Model Input - Replacing Decision Trees and Synergies with Individual Simulations

   This new process allow NHTSA to directly use Autonomie vehicle system simulation results
as input to the Volpe model. The process workflow can be summarized as shown below:
             ANL APRF's Vehicle
             Test Data for Validation
       Component
     Performance Data
     from Test and High
      Fidelity Models
                    Full Vehicle
                 Simulation Results
                for Every Technology
                   Combination

AUTDNDMIE
                                    Volpe Model
             Vehicle Technology
             Assumptions
                      Figure 5.149 Autonomie Directly Feeds the Volpe Model
                                              5-459

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.2 Autonomie Vehicle Simulation Tool

5.4.2.2.1     Overview

   Many of today's automotive control-system simulation tools are suitable for modeling, but
they provide rather limited support for model building and management. Setting up a simulation
model requires more than writing down state equations and running them on a computer. With
the introduction of EDVs, the number of components that can populate a vehicle has increased
considerably, and more components translate into more possible drivetrain configurations and
powertrain control options. In addition, building hardware is expensive. Traditional design
paradigms in the automotive industry often delay control-system design until late in the
process—in some cases requiring several costly hardware iterations. To reduce costs and
improve time to market, it is imperative that greater emphasis be placed on modeling and
simulation. This only becomes truer as time goes on because of the increasing complexity of
vehicles and the greater number of vehicle configurations.

   With the large number of possible advanced vehicle architectures and time and cost
constraints, it is impossible to manually build every powertrain configuration model. As a result,
processes have to be automated.

   Autonomie is a MATLABO-based software environment and framework for automotive
control-system design, simulation, and analysis.604 The tool is designed for rapid and easy
integration of models with varying levels of detail (low to high fidelity) and abstraction (from
subsystems to systems and entire architectures), as well as processes (e.g., calibration,
validation). Developed by Argonne National Laboratory (ANL) in collaboration with General
Motors, Autonomie was designed to serve as a single tool that can be used to meet the
requirements of automotive engineering throughout the development process from modeling to
control.  Autonomie was built to accomplish the following:

      •  Support proper methods, from model-in-the-loop, software-in-the-loop, and
          hardware-in-the-loop to rapid-control prototyping;
      •  Integrate math-based engineering  activities through all stages of development, from
          feasibility studies to production release;
      •  Promote re-use and exchange of models industry-wide through its modeling
          architecture and framework;
      •  Support users' customization of the entire software package, including system
          architecture, processes, and post-processing;
      •  Mix and match models of different levels of abstraction for execution efficiency with
          higher-fidelity models where analysis and high-detail understanding are critical;
      •  Link with commercial  off-the-shelf software applications, including GT-Power©,
          AMESim©, and CarSim©, for detailed, physically based models;
      •  Provide configuration and database management.

   By building models automatically, Autonomie allows the quick simulation of a very  large
number  of component technologies and powertrain configurations. Autonomie can do the
following:

      •  Simulate subsystems, systems, or entire vehicles;
                                             5-460

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
          Predict and analyze fuel efficiency and performance;
          Perform analyses and tests for virtual calibration, verification, and validation of
          hardware models and algorithms;
          Support system hardware and software requirements;
          Link to optimization algorithms; and
          Supply libraries of models for propulsion architectures of conventional powertrains as
          well as ED Vs.
   Autonomie was used to assess the energy consumption of advanced powertrain technologies.
Autonomie has been validated for several powertrain configurations and vehicle classes using
Argonne's Advanced Powertrain Research Facility vehicle test data.605

   With more than 400 different pre-defined powertrain configurations, Autonomie is an ideal
tool for analyzing the advantages and drawbacks of the different options within each family,
including conventional, parallel, series, and power-split HEVs. Various approaches have been
used in previous studies to compare options ranging from global optimization to rule based
control.606

   Autonomie also allows users to evaluate the impact of component sizing on fuel consumption
for different powertrain technologies as well as to define the component requirements (e.g.,
power, energy) to maximize fuel displacement for a specific application.607 To properly evaluate
any powertrain-configuration or component-sizing impact, the vehicle-level control is critical,
especially for ED Vs. Argonne has extensive expertise in developing vehicle-level controls based
on different approaches, from global optimization to instantaneous optimization, rule-based
optimization,  and heuristic optimization.608

   The ability to simulate a large number of powertrain configurations, component technologies,
and vehicle-level controls over numerous drive cycles has been used to support many DOE and
manufacturer studies. These studies focused on fuel efficiency, cost-benefit analysis, or
greenhouse gases.609 All the development performed in simulation can then be implemented in
hardware to take into account non-modeled parameters,  such as emissions and temperature.610

   Autonomie is the primary vehicle simulation tool selected by DOE to support its U.S. DRIVE
Program and Vehicle Technologies Office (VTO). Autonomie has been used for numerous
studies to provide the U.S. government with guidance for future research.611

   The vehicle models in Autonomie are developed under in Matlab/Simulink/Stateflow and are
open for users to view and modify any equation or algorithm.  Several hundreds of powertrain
configurations and more than 100 full vehicle models including controls are included in the tool.
Figure 5.150 shows the high level vehicle organization.
                                             5-461

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                                      tf
                                         i,L
                                             a .t
                      Figure 5.150 Autonomie Vehicle Model Organization

   The following section will describe the plant models and controllers.

5.4.2.2.2     Plant Model Overview

5.4.2.2.2.1    Internal Combustion Engine Model

   All Autonomie engine models use performance maps to predict fuel rate, operating
temperature and, in some cases when maps are available, emissions. The output torque of the
engine is computed from the engine controller command which takes a percentage of the spread
between the maximum engine torque map and the minimum engine torque map.  These maps are
based on two sources: from test data which are measured from engines running at steady state
points on an engine dyno or from high fidelity engine models such as GTPower.  These GT
Power engine maps can incorporate technologies such as GDI, VVL, VVT, camless and other
advanced engine technologies. In addition, to these performance maps, the engine models also
include a single time constant to represent the transient response of the engine output torque to
the engine command.

   However, some engine models use specific logic to represent some specific technology or
fuels. For example, Autonomie uses a specific model for spark ignition engine with a turbo
charger. The maps for turbo technologies were developed using GT-POWER©.  With turbo
engines, there is a 'lag' in torque delivery due to the operation of the turbo charger.  This impacts
vehicle performance, as well as the vehicle shifting on aggressive cycles.  Turbo lag has been
modelled for the turbo systems based on principles of a  first order delay, where the turbo lag
kicks in after the naturally aspirated torque limit of the turbo engines has been reached.  The
figure below shows the response of the turbo engine model for a step command.
                                            5-462

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                            Time(s)
                   Figure 5.151 Turbo-charged Engine Response for a 1L Engine

   It should be noted that the turbo response changes with engine speed (i.e., at higher speeds,
the turbo response is faster due to higher exhaust flow rates).

   Autonomie also uses a specific engine model for cylinder deactivation, as this model has a
more advanced fuel calculation subsystem which includes different maps. Due to NVH
considerations in production vehicles, cylinder deactivation operation is not performed during
several vehicle operation modes, like vehicle warm-up, lower gear operation, idle, and low
engine speed. In order to provide a realistic  evaluation of the benefits of cylinder deactivation
technology, cylinder deactivation is not been used under the following vehicle and engine
conditions:

       •   Cylinder deactivation is disabled if the  engine is at idle or any speed below 1000
           RPM or above 3000 RPM.
       •   Cylinder deactivation is disabled if the  vehicle is in the  1st of the 2nd gear.
       •   Cylinder deactivation is disabled if the  engine load is above half the max BMEP of
           the engine (and a certain hysteresis is maintained to prevent constant activation and
           deactivation).
   Typically, cylinder deactivation is not performed  during the vehicle warm up phase, i.e. for a
cold start. Since  all the simulations considered in this study assume a 'hot start', where in the
engine coolant temperature is steady around 95 degrees C, the cold start condition was not a
factor for the simulations. In addition, changes in the transmission shifting calibration (like
lugging speed limits) and additional torque converter slippage during cylinder deactivation have
also been disregarded.

   Autonomie also has a separate engine model for the spark ignition engine with fuel cut off.
This engine has a specific torque calculation to calculate the engine torque loss when the engine
fuel is cut off during deceleration events.  In general, engine models in Autonomie are of two
types, throttled engines and un-throttled engines. As shown in the figure below, both types of
models provide motoring torque when fuel is cut to the engine (e.g. fuel cut off during
deceleration). With throttled engines, the motoring torque is a function of throttle position.
                                              5-463

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                                 Torque
                                                      Max torque
                                                     curve (WOT)
                                (slightly
                              below idle,0
                    Figure 5.152 Engine Operating Regions for Throttled Engines
                  Figure 5.153 Engine Operating Regions for Un-throttled Engines
5.4.2.2.2.2    Transmission Mode Is

   Automatic Gearbox Model

   The gearbox model allows for torque multiplication and speed division based on the gear
number command from the powertrain controller. As for all the other models, the losses are
taken into account using torque losses to easily deal with regenerative conditions.
                   Gear # command

                   Torque in, Inertia

                   Rotational speed out
Information

Torque out, Inertia

Rotational speed in
                      Figure 5.154 Automatic Gearbox Model Input / Output

   The drivetrain is considered rigidly attached to the wheels. Since the wheel speed and
acceleration are calculated in the wheel model and propagated backward  throughout the rest of
the drivetrain model, the gearbox unit is modeled as a sequence of mechanical torque gains. The
torque and speed are multiplied and divided, respectively, by the current  ratio for the selected
gear. Furthermore, torque losses corresponding to the torque/speed operating point are subtracted
                                               5-464

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
from the torque input. Torque losses are defined on the basis of a three-dimensional efficiency
lookup table that has as inputs input shaft rotational speed, input shaft torque, and gear number.

   When a gear is selected, the input inertia is fed forward to the next component after being
reflected to the output shaft using the square of the gear ratio. When the neutral gear is engaged,
the input gearbox rotational speed is calculated on the basis of the input shaft inertia.

   Since this is an automatic gearbox model, it can be shifted in sequence from one gear to
another without having to pass through neutral and without a complete torque interruption at its
output. The torque passing through  the transmission during shifting is reduced, but does not go to
zero as it does for a manual gearbox. Also, the torque converter model is separate from the
automatic gearbox model.

   Dual Clutch Transmission

   Dynamic models of the dual-clutch transmission  are obtained including the clutch and gear-
train, but no synchronizer dynamics. Figure 5.155 illustrates an example of DCT system that can
be considered as a combination of two manual transmissions, with one providing odd gears
connected to clutchl, and the other  providing even gears connected to clutch2. With alternating
control of the two clutches, the oncoming clutch engages and the off-going clutch releases to
complete the shift process without torque interruption. It is necessary to preselect the gears to
realize the benefits of the DCT system. The different plant models and controls have been
validated using vehicle  test data.
                                              Clutch
                                                            Gear-train
                                                                          Final drive
                     Figure 5.155 Dual Clutch Gearbox Model Input / Output

   The pre-selection of gears can be implemented by considering the operating conditions of the
DCT system. For example, if the first synchronizer is at the first-gear position, and the third
through fifth synchronizers are at the neutral position (as they must be), then the gear ratio
between shaftl and the output shaft is first gear. At same time, the gear ratio between shaft2 and
the output shaft can be selected in the same manner for the pre-selection mode. To achieve a
desired input-output gear ratio, the corresponding synchronizer and clutch have to be applied.

   Continuously Variable Transmission

   A metal V-belt CVT model is considering hydraulic and mechanical loss. The hydraulic loss
constitutes the majority of the total loss at low vehicle speed, whereas the mechanical loss is the
main source of inefficiency at high speed. The operating conditions of the metal V-belt CVT
system can be described by the following five parameters:
                                             5-465

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
       1) Primary clamping force (FP) or primary pressure (PP);
       2) Secondary clamping force (FS) or secondary pressure (PS);
       3) Primary revolution speed (&>p);
       4) Input torque (TIN); and
       5) Pulley ratio (i).
   On both the primary and the secondary pulleys, the belt is clamped by the forces produced by
the hydraulic pressures in the cylinders. These two clamping forces, FP and Fs, counteract each
other. Therefore, when the pulley ratio is constant, there is a balance between FP and FS. A ratio
change occurs when their balance is lost.

   In addition, CVT ratio control and clamping force control strategies, including the CVT shift
dynamics, were developed. The following are considered in the low-level controller:

       •  The demanded CVT ratio is determined from the engine OOL;
       •  The secondary pressure is determined for the given input torque and CVT ratio; and
       •  The primary pressure is controlled to meet the demanded CVT ratio.
   Figure 5. 156 shows a block diagram of the model -based ratio control and plant.
    CVT control Block
                                  CVT plant Block
                           Figure 5.156 CVT Model Block Diagram
   Torque Converter
   The torque converter is modeled as two separate rigid bodies when the coupling is unlocked
and as one rigid body when the coupling is locked. The downstream portion of the torque
converter unit is treated as being rigidly connected to the drivetrain. Therefore, there is only one
degree of dynamic freedom, and the model has only one integrator.

   The effective inertias are propagated downstream until the point where actual integration
takes place. When the coupling is unlocked, the engine inertia is propagated up to the coupling
input, where it is used for calculating the rate of change of the input speed of the coupling. When
the coupling is locked, the engine inertia is propagated all the way to the wheels.
                                             5-466

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The torque converter model is based on a lookup table, which determines the output torque
depending on the lockup command. The upstream acceleration during slip and the downstream
acceleration are taken into account in calculating the output speed.

5.4.2.2.2.3    Electric Machine Models

   Electric machine plant models in Autonomie can take in Torque or Power as the command
and produce a Torque output. Operating speed of the motor is determined by the components
connected to the motor. In a vehicle, the vehicle speed and gear ratios determine the operating
speed of the motor.

   The lookup table used in a motor model estimates the operational losses over the entire
operating region of the motor. This map is typically derived from the efficiency map provided in
the initialization file.

   Typically, every motor has a continuous operating region (region under the continuous torque
curve as shown  in figure), and a transient region where the motor can operate for a short period
of time (peak torque capability of a motor is defined for a specific duration, e.g. 30s). The
maximum torque output gets de-rated to the continuous torque levels, when the electric machine
temperature increases. The electric machine model in Autonomie has this general logic built into
it.

   Autonomie provides a logic to scale an existing motor to a different power rating. The shape
of the efficiency map is kept the same, but the torque  axis is scaled to meet the desired power
rating.

5.4.2.2.2.4    Energy Storage Models

   Autonomie includes  several energy storage models depending on the application (i.e. high
power, high energy).  The default battery model is  a charge reservoir and an equivalent circuit
whose parameters are a function of the remaining charge in the reservoir, also known as the state
of charge (SOC). The equivalent circuit accounts for the  circuit parameters of the battery pack as
if it were a perfect open circuit voltage source in series with an internal resistance. Another
battery model in Autonomie is the  one used for high energy batteries. The equations and
schematic of this type of battery is  shown in Figure 5.157. This model uses two time constants to
represent the polarization behavior of the battery pack. This lumped parameter model can
represent many  different battery chemistries for the internal resistances, capacitances and open
circuit voltage are all maps based on SOC and, in some cases, temperature.
                                             5-467

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                                               AAAAA
                                                  Ro
                          v,
                                                                  ocv
o
)\
Rpi
_/\AAAA_

)\
Rp2
_/V AAAA—

                                                           Ip2
    Equation (1)        1000*(OCV-VL)/IL = R =
                      Where,          OCV = open circuit voltage, V
                                        VL = cell voltage, V
                                        R = total cell impedance, milliohms
                                        Ro = cell internal ohmic resistance, milliohms
                                       Rpi = first internal polarization resistance, milliohms
                                       Rp2 = second internal polarization resistance, milliohms
                                        IL = cell load current, A
                                        Ip-i = current through first polarization resistance, A
                                        Ip2 = current through second polarization resistance, A

                        Figure 5.157 High Energy Battery Model Schematic
   Another important aspect to consider for sizing is the pulse power limits of the battery pack.
There are several different options to represent the maximum power of the battery in Autonomie.
The most basic represents maximum power as a function of SOC. Other models introduce a time
constraint for the maximum power. These battery packs have different power limits for 10
second, 2 second and continuous power. The Autonomie model accounts for the duration of the
pulse and limits the power accordingly. This aspect is not necessary a feature of the plant, but is
handled by the low level control and is dependent on the battery chemistry and plant's
performance characteristics.

5.4.2.2.2.5     Chassis Models

   The chassis plant model in Autonomie translates  the force from wheel to vehicle acceleration
and linear speed.  The losses related to moving the vehicle is  estimated in this model. Two types
of initialization data can be used for estimating this behavior.

       •  Coefficients  derived from a coast down test. The  losses estimated from these
          coefficients will cover both rolling resistance & aero dynamic losses. Dyno set values
          for nearly every vehicle is available from EPA.
       •  Values for coefficient of drag, frontal area, rolling resistance of tires etc.
                                              5-468

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   The model based on coast down is used for validation purposes while the model based on the
aerodynamic equations is used to predict the impact of non-existing vehicles

5.4.2.2.2.6    Tire Models

   Just as the two chassis models, there are two wheel models corresponding to each of the
chassis models. The initialization data for the wheel rolling resistance can be provided by the
user in many ways. Wheel radius can be provided by the user, or this could be computed by
Autonomie from a sidewall label of the tire e.g. P225/50/R17. The tire losses model uses a
constant and a speed term to represent the losses.

5.4.2.2.2.7    A uxiliaries Model

   Most powertrains in Autonomie have two accessory models. The mechanical accessories
driven by the engine through a belt and the electrical accessories connected to the lower voltage
bus.

   The main electrical accessory model in Autonomie is a constant power draw. If the vehicle
has a high voltage bus, a step down power conditioner is connected between the high voltage bus
and low voltage bus to supply the electrical accessories. When a vehicle contains thermal
models, a current draw is added to represent the  electrical power draw of the cooling fans.

5.4.2.2.2.8    Driver Models

   Autonomie uses a look-ahead driver to better approximate the behavior of a real driver.
Forward looking models are especially sensitive to how well the driver follows  the trace  and how
aggressive the driver is in doing so. Both of these factors can noticeably affect fuel economy
results when simulating advanced vehicles. For example, a driver which is too aggressive can
add additional engine on events for a hybrid or delay transmission shifts for a conventional, both
of these events lower fuel economy. For this reason, Autonomie employs a look ahead driver,
which at its core, is a PI controller with a feedforward part that, in addition, uses time advanced
copies of the trace to replicate the ability of a human driver to look a few seconds ahead on the
driver's aid to  anticipate accelerations and decelerations. The result is a smoothing of the pedal
demand from the driver, which leads to an overall more representative fuel  economy.  The added
complexity yields several additional dimensions of tuning to the model, for the relative
weightings of the time advanced copies have to be optimized.

   The driver model also uses an additional layer of logic to manage the accelerator pedal
demand, specifically, during shift events when the engine is disconnected from the wheels. On a
manual transmission, during the shift through neutral, the driver must be capable of expecting a
decrease in vehicle speed and not aggressively stomp on the accelerator pedal in an attempt to
compensate for the decrease in vehicle speed.

5.4.2.2.2.9    Environment Models

   The environment model in Autonomie outputs all of the relative information about the
operating environment of a vehicle during a simulation such as ambient temperature, ambient
pressure, relative humidity, air density and grade. There are two versions of the environment
model in Autonomie, one for which the grade  is  a function of time, such as would be
encountered on a chassis dynamometer test which follows a preset grade schedule, and the other
for which the grade is a function of distance as when following a mapped route.
                                             5-469

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.2.3
Control Overview
   All the vehicle-level control algorithms used in the study were developed on the basis of
vehicle test data collected at Argonne's Advanced Powertrain Research Facility. It is important
to note that while the logic for the vehicle-level control algorithms were developed on the basis
of test data, only the logic has been used for the present study, since the calibration parameters
have been adapted for each vehicle to ensure energy consumption minimization with acceptable
drive quality (i.e., number of engine on/off conditions, and shifting events).

5.4.2.2.3.1    Transmission Shifting A Igorithm

   The transmission shifting logic has a significant impact on vehicle fuel economy and should
be carefully designed to maximize the powertrain efficiency while maintaining acceptable drive
quality. The logic used in the simulated conventional light-duty vehicle models relies on two
components: (1) the shifting controller, which provides the logic to select the appropriate gear
during the simulation; and (2) the shifting initializer, the algorithm that defines the shifting maps
(i.e., values of the parameters of the shifting controller) specific to a selected set of component
assumptions.

   Shifting Controller

   The shifting controller determines the  appropriate gear command at each simulation step. A
simplified schematic of the controller is shown in Figure  5.158. The letters and numbers in the
discussion that follows correspond to those shown in the figure.
                Downshifting
                           K(t)
                                                          a(t) > Out OR «/ter(r)
           a(t): Accelerator pedal
           V(t): Vehicle speed
           y(t): Gear
                                                 State-machine
7,,p (t) : Upshifting Gear
/„„(*): Downshifting Gear
Ycmd (t): Gear Command
                                           ant : Accelerator pedal position
                                           above which Hmer is ^-Passed
                                           r: Minimu™ Hme shift criteria must be
                                           met before shift occurs
                      Figure 5.158 Shifting Controller Schematic in Autonomie
                                               5-470

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The controller is based on two main shifting maps — one for upshifting (a), moving from a
lower gear to a higher gear, and another one for downshifting (b), moving from a higher gear to a
lower gear — as well as a state-machine (c) that defines the status of the system (e.g., no
shifting, upshifting).  Each shifting map outputs a next-gear command 7dn(0 and 7up(t) based
on the current accelerator pedal position a (t) and vehicle speed V(t).  The state machine is
composed of different states, of which only one is active at any time step;  a change in state
occurs whenever a transition condition from the active state becomes true  (i.e., an upshift will
occur only if a set of conditions is true). The state that is active most of the time is the hold-gear
state (d), which makes sense because, most of the time, the vehicle should be in gear and not
shifting for drivability reasons. An upshift occurs when the upshifting gear yup(f) is strictly
higher than the current gear y(t) (1) (e.g., yup(t)  = 5 and y(t) = 4). For all vehicles, the shift
does not necessarily happen instantly when the command to shift is given, depending on the
current pedal position.  In aggressive driving, i.e., at high accelerator-pedal positions (5), the
shift happens as  soon as the gear transition (1)  becomes true, ensuring optimal performance.  In
contrast, in "normal" driving, i.e., at low pedal positions (2), there is an intermediate state (e) that
allows the shift only when the gear condition (1) is true for a minimum time T. This constraint  is
imposed to avoid an excessive number of shifting events, which would lead to unacceptable
drive quality and increased energy consumption. The upshifting itself is executed in state (f), in
which the shift command ycmd(0 is incremented (i.e., the next upper gear is selected); once the
shifting is completed (6), the state machine comes back to the hold-gear state (d). Downshifting
occurs in a similar way.

   Currently, in Autonomie, a shifting event can only result in moving one gear up or one gear
down: there is no gear-skipping. Gear skipping is usually used under very  specific conditions
that are not encountered during the standard FTP and HFET drive cycles considered in the study.
As an additional level of robustness in the Autonomie control algorithm, an upshift or downshift
cannot occur if the resulting engine speed would be too low or too high, respectively. This
approach ensures that the engine is not operated below idle or above its maximum rotational
speed as shown in Figure 5.159.
                                             5-471

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
    Wheel torque demand to Pedal %
                                                   Upshift map
                                                                        Downshift map
    Engine speed calculation for next gear
                         Figure 5.159 Shifting Calculations in Autonomie
   Shifting Initializer

   The shifting controller uses shifting maps to compute the gear command. In the controller, the
shift map is a two-dimensional (2-D) look-up table indexed by vehicle speed and accelerator-
pedal position. Defining such a map is equivalent to defining the "boundaries" of each gear area;
those boundaries are the shifting speeds. Figure 5.160 illustrates that equivalence.
              0      33     40      60
                       Speed (irfs)
0       20      40
         Speed (rrfs)
              Figure 5.160 Upshifting Gear Map (left), Upshifting Vehicle Speeds (right)

   For each shifting curve, there are two key points: the "economical" shifting speed (at very low
pedal position) and the "performance" shifting speed (at high pedal position). The objective of
the control engineer is to combine both goals of the shifting control to fulfill the driver
                                              5-472

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
expectations: minimization of energy consumption on the one hand and maximization of vehicle
performance on the other.
   The economical shifting speed for an upshift or a downshift is the speed at which the
upshift/downshift occurs when the accelerator pedal position is very lightly pressed. Vek^k+l is
the economical vehicle speed for upshifting from gear k to gear k+1. l^co1^ i§ the downshifting
speed for this same set of gears. The vehicle speed shift points are computed from the engine
shift points &>ecofc+1 and ^eco1^- Figure 5.161 shows the engine speed shift points for an engine
associated with a 5-speed transmission.
                     Gear
5
4
r
3


«£?

,,2-sl
w«o
•
.",
....




CJ4^3

/



N
7
b*tCQ
^/


	 T
\
-

s
,;,-.'
W«Tffl


s
3—^4
	 1st
— 3rd




4

,.4-iS
"«ro
7"

gear
gear
:==•
:==•
jear




                           503    1QW    1500    2000    2500    3000    3503
                                      Engine Speed (RPM)

        Figure 5.161 Example Engine Speed Range in Economical Driving, and Economical Shift

   The initializing algorithm for the shifting controller computes the up-  and downshifting
speeds at zero pedal position based on the four "extreme" shift points: upshifting from lowest
gear (&>eco2X upshifting into highest gear (w^o1^"). downshifting into lowest gear (wf^1), and
downshifting from highest gear (<^ecoN~1)- N is the number of gears. The speeds can be set by
the user or left at their default values. Below is a description of their default values in
Autonomie:
            + <*>
               margin
                               - engine idle speed; a)margin- speed margin, -50-100 rpm]
        = Midie 71 (1 + eud) [kl,k2: gear ratios for gears 1,2; eud : margin to avoid overlap,
0.05-0.1]
    >eco1^N'- Engine speed at which best efficiency can be achieved
(JL)C
                                  1,000 rpm]
   Once those four speeds are computed, the remaining ones are computed by linear
interpolation to allow consistent shifting patterns that are acceptable to the drivers. For example,
any upshifting speed is given by Equation 1:
                 ,1->H
                 Jeco

                           eco
                                  _
                                   l->2
                                   eco
                               N -2
                                        (i-l)+  a)£02,l
-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   In a shifting map, the vehicle upshifting speed from gear i to i+1 shall be strictly higher than
the downshifting speed from gear i+1 to i. Otherwise, the downshifting speed will always request
gear i while gear i+1 is engaged and vice-versa, resulting in oscillations between gears that
would be unacceptable to the driver. For this study, the algorithm in the initialization file
prevents that by making sure the following relation is true:
                                 .
                                 eco
   The values of the engine economical shifting speeds at lowest and highest gears are
automatically defined on the basis of the engine and transmission characteristics.

   Finally, the vehicle economical up- and downshifting speeds can be computed using the
engine up- and downshifting speeds, the gear ratio, the final drive ratio and the wheel radius:
V*
                   R
                    wh
   Where:     kFD is the final drive ratio and Rwh is the wheel radius.
   During performance, the gears are automatically selected to maximize the torque at the wheel.
Figure 5.162 illustrates that gear selection, which consists of finding the point where the engine
peak torque (reported at the wheels) curve at gear k falls under the one at gear k+1.
           Figure 5.162 Maximum Engine Torque at Wheels and Performance Upshift Speeds

   The performance downshifting speed is given by the performance upshifting speed and the
difference between the economical shifting speeds:
                 = a
Pf.ec
'perf
'perf
                                                    = a
                                                          •
                                                     pf,ec  \vperf
                                                                         perf
   The definition of the final shifting curves is critical to properly evaluating the benefits of
transmission technologies while maintaining acceptable performance. Figure 5.163 shows how a
set of upshifting and downshifting curves for two adjacent gears is built, based on selected
vehicle speeds and accelerator pedal positions. At low pedal positions (i.e., below a^0 ), the
                                             5-474

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
upshifting speed is the economical upshifting speed. Similarly, below ag™0, the downshifting
speed is the economical downshifting speed.  This approach ensures optimal engine operating
conditions under gentle driving conditions. At high pedal positions (i.e., above aper/), the
shifting speed is the performance shifting speed, ensuring maximum torque at the wheels under
aggressive driving conditions.
                                                         — Upshifting
                                                         — Downshifting
       Figure 5.163 Design of Upshifting and Downshifting Speed Curves for Two Adjacent Gears

   Torque Control during Shifting Events

   Figure 5.164 shows the transmission clutch pressure, output torque, and engine speed curves
during a change from  1st to 2nd gear. The output torque experienced both a trough period (lower
than the torque in the original gear) and a crest period (higher than the torque in the original
gear). The trough period is called a torque hole, while the crest period is called a torque
overshoot. The torque hole is defined by depth and width, where the depth is the difference
between minimum torque and the torque in previous gear, and the width is the half value of the
maximum width of the torque hole.
                                     Torque
                                     Phase
                                                               Time
                                    Poorly controlled pressures
                                    Well controlled pressures
                   Figure 5.164 Generic Shift Process for Automatic Transmission
                                              5-475

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The bigger the torque hole, the larger the decrease of torque in torque phase, which results in
a more significant reduction in acceleration. Because the decrease in acceleration causes
discomfort for both the driver and passengers, the torque hole should be as shallow and  narrow
as possible. Torque reduction behavior is a well-known phenomenon, observed during vehicle
testing and referenced in several papers and presentations.

   Autonomie integrates a low-level control algorithm that reproduces the torque hole
phenomenon. Figure 5.165 illustrates, in detail, the behavior of the vehicle model for a short
period of time [205 sec to 205.8 sec]. The area highlighted by the grey circle indicated the torque
hole during a shifting event.
                 10000

                  9000

                  8000

                  7000

                  6000

                  5000

                  4000

                  3000

                  2000
-vehicle speed*100 mph
 gear*1000
-engine torque*100 Mm
-engine speed rpm
                        204.2  204.4 204.6  204.8  205  205.2 205.4  205.6  205.8
                                          seconds

                   Figure 5.165 Torque Hole in Autonomie during Shifting Event

   Engine Lugging Limits

   Engine lugging limits are a critical NVH parameter. The assumptions shown in Table 5.217
describe the logic implemented in Autonomie to prevent lugging for multiple transmissions. The
logic and values were developed based on APRF vehicle test data analysis.

   Shift parameters are selected such that low speed, high torque operation is avoided. The
selected shifting limits are based on test data observations relative to the number of gears
available.

                    Table 5.217 Vehicle and Powertrain Technologies Evaluated

Lugging speed
(rad/s)
5 speed Trans.
140
6 speed Trans.
130
7 speed Trans.
120
8 speed Trans.
110
   Figure 5.166  Example of Engine Operating Conditions to Prevent Lugging shows an example
of how engine operating conditions are restricted to prevent lugging for multiple transmissions (5
and 8 speed automatic) on the UDDS driving schedule.
                                              5-476

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   200
   150
            5 Speed Automatic
                          8 Speed Automatic
                            Max torque curve
                            Eng Operating Points for 5spd Trans.
                       Areas where
                       engine Operation
                       is Not Allowed
         100
               200
                    300    400
                   Engine Speed rad/s
                               500
                                      	Max torque curve
                                       • Er>g Operating Points for 8spd Trans
                                          700
                               300    400
                              Engine Speed rad/s
              Figure 5.166 Example of Engine Operating Conditions to Prevent Lugging
   Shifting Maps

   All shifting maps used for the simulations are presented below. The shifting maps have been
developed to ensure minimum energy consumption across all transmissions while maintaining an
acceptable  drivability. While plant models with higher degree of fidelity would be necessary to
accurately model the impact of each technology on the drivability, using such models was not
appropriate for the current study. As a result, the work related to the drive quality was focused on
number of shifting events, time in between shifting events, engine time response and engine
torque reserve.
                                    Upshift and downshift Table
                150
                         UpSft Gear Map
                             Gear Map
                             20
     40         60
Normalized Accel Demand (%)
80
100
         Figure 5.167 5-Speed Automatic Up (plain lines) and Down (dotted lines) Shifting Map
                                              5-477

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                                    Upshift and downshift Table
                150
                   ™"M9W""500™	"JO*"'
                        40C
                             20         40         60         80
                                   Normalized Accel Demand (%)
100
         Figure 5.168 6-Speed Automatic Up (plain lines) and Down (dotted lines) Shifting Map

                                    Upshift and downshift Table
                150
                                                                    -500100-
                                                                ••••••••  400
                       600         ^^**S?
                    • BBB^jQ^BBBBBBBBBB IBBBI-700"aBBBBBBI
                   uuupff       000
                             20         40         60         80
                                   Normalized Accel Demand (%)
         Figure 5.169 8-Speed Automatic Up (plain lines) and Down (dotted lines) Shifting Map
5.4.2.2.3.2    Torque Converter Lock-up Assumptions

   A torque converter is a hydrodynamic fluid coupling used to transfer rotating power from a
prime mover, such as an internal combustion engine, to a rotating driven load. It is composed of
an impeller (drive element); a turbine (driven component); and a stator, which assist the torque
converter function. The torque converter is filled with oil and transmits the engine torque by
means of the flowing force of the oil. The device compensates for speed differences between the
engine and the other drivetrain components and is therefore ideally suited for start-up function.
                                              5-478

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   The torque converter is modeled as two separate rigid bodies when the coupling is unlocked
and as one rigid body when the coupling is locked. The downstream portion of the torque
converter unit is treated as being rigidly connected to the drivetrain. Therefore, there is only one
degree of dynamic freedom, and the model has only one integrator. This integrator is reset when
the coupling is locked, which corresponds to the loss of the degree of dynamic freedom. Figure
5.170 shows the efficiency of the torque converter used for the study.

   The effective inertias are propagated downstream until the point where actual integration
takes place. When the coupling is unlocked, the engine inertia is propagated up to the coupling
input, where it is used for calculating the rate of change of the input speed of the coupling. When
the coupling is locked, the engine inertia is propagated all the way to the wheels.
                        25
                      =§ 1.5
                        0.5
                          7
                                                                 0.8
                                                                 0.6s?
                                                                 0.4?=;
                                                                 0.2
                          0
                                 0.2
                                         0.4       0.6      0.8
                                      Output/Input Speed Ratio
                       Figure 5.170  Torque Converter Efficiency Example

   Figure 5.171 describes the conditions under which the torque converter will be locked. The
same algorithm is used to represent current torque converter lockup logic, as well as future
aggressive lockup logic. The torque converter is used as a start-up device in the first gear, with
very low slip (torque ratio of 0.95) at higher speeds, in the first gear. Recent trends in torque
converter technology suggest operation in locked or controlled slip mode, in the 2nd and higher
gears. In general, the torque converter is in controlled slip or mechanically locked based on
vehicle speed and pedal position, for each gear apart from the 1st. In order to suggest advances in
torque converter technology, it was assumed that the torque converter would be in a
mechanically locked state for the 2nd and higher gears. This approach has been applied to all
transmissions with 6 gears or more.
                                              5-479

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
  Vehicle Soee-
  Pedal Position^

   Gear Number
If the shifting is not in progress, the torque converter is
  locked (1) at a specific gear number and (2) pedal
         position for a given vehicle speed.
Lockuo Sianal
                     Figure 5.171 Torque Converter Lockup Control Algorithm
5.4.2.2.3.3    Fuel Cut-off Algorithm

   Engine fuel cut-off control algorithms used in the study have been developed on the basis of
vehicle test data collected at Argonne's Advanced Powertrain Research Facility. The fuel cut-off
controller is implemented for gasoline and diesel engines through analysis as shown in Figure
5.172  Engine Fuel Cut-off Analysis Based on Test Data (data source APRF). In Autonomie,
engine control and plant blocks are organized for idle fuel rate and fuel off conditions. Engine
fuel is cut off under the following conditions:

   Vehicle is actively braking, for a certain minimum time.

   Engine speed is above a minimum threshold (e.g. 1000 RPM).
                          Fuel flow
                         —Engine speed /100
                          fKtelPosilion'lOlvr
                         —ihrotUe position (in«le|
       2500 2600  2700  2800 2900 3000 3100 3200  3300 3400 3500-i
       A 	i	L tn
                                       1000  1500  2000  2500  3000  3500
                                         Engine speed [RPM]
           Figure 5.172  Engine Fuel Cut-off Analysis Based on Test Data (data source APRF)
5.4.2.2.3.4    Vehicle Level Control for Electrified Power trains

   The task of achieving fuel savings with a hybrid architecture depends on the vehicle
performance requirements and the type of powertrain selected as well as the component sizes and
technology, the vehicle control strategy, and the driving cycle. The overall vehicle-level control
strategy is critical to minimize energy consumption while maintaining acceptable drive quality.
Figure 5.173 illustrates a simple acceleration, cruising and braking cycle for a full FIEV,
demonstrating the best usage of different power sources based on the vehicle's power demand.
During small accelerations, only the energy storage power is used (EV mode) and during
                                               5-480

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
braking, some of the energy is absorbed and stored. The engine does not start to operate during
low power demands, owing to its poor efficiency compared to the electrical system. The engine
is only used during medium and high power demands, where its efficiency is higher.
                Auto-start
               with torque
                smoothing
               Figure 5.173 Hybrid Electric Vehicle Principles [source: www.gm.com]

   While different vehicle-level control strategy approaches have been studied for electric drive
vehicles (e.g., rule based, dynamic programming, instantaneous optimization), the vast majority
of current and future electric drive vehicles are using and expected to use rule-based control
strategies. The vehicle level control strategies logics used in the study will be described below.

   It is important to note that while the control algorithms have been developed based on
extensive vehicle test data, the calibration parameters used for the study were adapted to the
component technologies and performance characteristics (i.e., power, energy, and efficiency) of
each individual vehicle.

   Micro and Mild HEV

   The vehicle-level control strategies of the micro- and mild (i.e., BISG and CISG) micro-
HEVs are similar in many aspects due to the low peak power and energy available from the
energy storage system.

   For the micro HEV case, the engine is turned off as soon as the vehicle is fully stopped and
restarted as soon as the brake pedal is released. No regenerative braking is considered for that
powertrain.

   For the mild HEV cases, the engine is turned off as soon as the vehicle is fully stopped.
However, since some regenerative braking energy is recovered, the vehicle is propelled by the
electric machine during vehicle launch, allowing the engine to be restarted later.

   Single-mode power split HEV

   The vehicle-level control strategy algorithm of a single-mode power split HEV was based on
the Toyota Prius APRF test data analysis. The control logic implemented can be divided into
                                             5-481

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
three areas: engine-on condition, battery SOC control, and engine operating condition. Each
algorithm is described below.

   The operation of the engine determines the mode, such as pure electric vehicle mode or HEV
mode. The engine is simply turned on when the driver's power demand exceeds a predefined
threshold. As shown in Figure 5.174, the engine is turned on early if the SOC is low, which
means that the system is changed from PEV mode to HEV mode to manage the battery SOC.
Wheel power demand (kW)
(j,O tOJ^^OOOtO-t^^OO
Engine-on Condition






. •


F • F F
• Eng
° Eng


.
,'+&&
«• •



ine-on points
ine-on points at high torq
,
•
t •
,^$fj
•
•



•
&&.'£
^fef
0
o
o


ue demand

•
't

§S
u
o (9
n
'• c

0 55 60 65 70 7
Battery SOC (%)
   Figure 5.174 Engine-On Condition - 2010 Prius Example Based on 25 Test Cycles (data source APRF)
   The engine is turned off when the vehicle decelerates and is below a certain vehicle speed.

   The desired output power of the battery is highly related to the energy management strategy.
When the vehicle is in HEV mode, the battery power is determined by the current SOC, as
shown in Figure 5.175. The overall trend shows that the energy management strategy tries to
bring the SOC back to a regular value around 60 percent. Both the engine on/off control and the
battery power control are robust approaches to manage the SOC in the appropriate range for an
input-split hybrid. If the SOC is low, the engine is turned on early, and the power split ratio is
determined to restore the SOC to its target value so that the SOC can be safely managed without
charge depletion. In summary, the battery SOC is controlled by raising (low SOC) or lowering
(high SOC) the engine power demand required to meet the vehicle speed trace.
                                            5-482

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                       Battery Output Power
                                             60
                                          SOC (%)

 Figure 5.175 SOC Regulation Algorithm - 2010 Prius Example Based on 25 Test Cycles (data source APRF)

   For engine operation control, the two previously described control concepts determine the
power split ratio. The concepts do not, however, generate the target speed or torque of the engine
because the power split system could have infinite control targets that produce the same power.
Therefore, an additional algorithm is needed to determine the engine speed operating points
according to the engine power, as shown in Figure 5.176. An engine operating line is defined on
the basis of the best efficiency curve to select the optimum engine speed for a specific engine
power demand.
                           4000
                                      Engine Operating Targets
                           1000
                                        Engine Power (kW)

  Figure 5.176  Example of Engine Operating Target - 2010 Prius Example Based on 25 Test Cycles (data
                                       source APRF)

   In summary, the engine is turned on based on the power demand at the wheel along with the
battery SOC. If the engine is turned on, the desired output power of the battery is determined on
the basis of the current SOC and the engine should provide appropriate power to drive the
vehicle. The engine operating targets are determined by a predefined line, so the controller can
produce required torque values for the electric machine and the generator on the basis of the
engine speed and torque target.
                                              5-483

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Pre-transmission HEV

   The vehicle-level control strategy logic of a pre-transmission HEV is based on the VW Jetta
HEV APRF test data analysis. In the pre-transmission HEV, the engine is a main power source
and the electric machine assists the engine according to the vehicle operating conditions and the
driver request. Three driving modes are used: EV mode, engine mode, and HEV mode. When the
vehicle is driving at low speed or the demanded power is low, the vehicle is operated only by the
electric machine in EV mode. During high-speed operation, start-up, or aggressive acceleration,
the vehicle is operated by the engine in Engine mode or HEV mode.

   The driving mode control strategy is determined by the engine on/off state. When the vehicle
drives at low speed, the system is operated only by the electric machine, without engine
operation. Figure 5.177 (left panel) shows the vehicle speed and wheel demand torque when the
engine is turned on. The right figure shows the operating area of pure electric driving in the same
index.
1200

1000

era

«M.
                 Engin* Turn On Common
                                                          Pun- Electric V«hlcl« Mod*
m


          s

                               1'
                                                                        HEV mode
                                                                        PEV mode
                16    20
                Veh Spd. nv'm
                                                            16   20
                                                             Veh Spd. ml*
  Figure 5.177 Cycles Wheel Torque vs. Vehicle Speed, 2014 Jetta HEV Based on Test Cycles (data source
                                         APRF)

   In HEV and engine mode, the engine is operated to manage the demanded power at high
speed or acceleration. In these modes, the engine is controlled to operate at higher engine
thermal efficiency. However, since the range of the multi-gear transmission gear ratio is limited,
the electric machine is used to provide additional control of the engine operating points.
Therefore, one other important control concept at the vehicle level is how to manage the battery
demand power within the appropriate SOC range. Figure 5.178 (left panel) shows the battery
SOC target when the engine is turned on. Under the engine on/off condition, the proportional
demand power for the battery sustains the SOC level at an appropriate range near specific range.
On the right, engine power vs. wheel power is shown for a 2014 Jetta HEV example based on
test cycles.
                                             5-484

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                     SOC Target
                                                            Engin* Pow«f
                                                            .
       Figure 5.178 SOC vs. Time (left) Engine Power vs. Wheel Power (right) (data source APRF)
   Plug-in Hybrid Electric Vehicle - Blended PHEV

   The vehicle-level control strategy logic of a single-mode power split blended PHEV was
based on the Toyota Prius PHEV APRF test data analysis. The PHEV is able to run with the
electric machine only if SOC is high enough and the demand power does not exceed the power
limit of the electric machine and the battery. Figure 5.179 shows all points when the engine is
turned on.
                      I
                      I
                                           39    «    *    fl
Figure 5.179 2013 Prius PHEV Wheel Speed and Demand Torque, Based on Test Cycles (data source APRF)

   Another control strategy logic is necessary to distribute the power between the engine and the
battery, which determines the behaviors of SOC on the hybrid driving mode. Figure 5.180 shows
the overall control strategy to manage the SOC according to the CD or CS mode.

   In Figure 5.180, the points are obtained only during the hybrid driving mode because the
battery provides all demand power if the electric machine is the only power source. First, the
battery provides no power or constant power under the CD mode if the SOC is greater than 28
percent. The engine is turned on under the CD mode when the battery does not provide  all the
demand power, and the engine provides all demand power. However, if the vehicle speed
exceeds 100 km/h, the battery provides a constant power (here about 10 kW).
                                            5-485

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment

                       I
                        \
                        7    :"--;T       •  .-•       •   .. -
                        £-    '• ^mm ;   * ••'             • •
                        5 ,   J
                        =    ;
                        -
                       *
                             • i    Vi     i     V,    ...     TIJ •"—
 Figure 5.180 2013 Prius PHEV Output Power of the Battery for SOC Balancing Based on Test Cycles (data
                                      source APRF)

   This control is designed to constantly consume electric energy under the CD mode, so that
drivers have consistent experiences during the CD mode. In contrast, the control strategy to
manage the SOC in the CS mode is similar to the Prius HEV, where the desired power of the
battery decreases as the SOC decreases. Further, rapid recuperation is also observed in the very
low SOC range, like below 20 percent, and there is no specific control for the SOC balancing
according to the battery temperature just as for the Prius HEV. In Figure 5.181, for a 2013
Toyota Prius PHEV, the power constraints are observed in the regenerative operation because the
electric machine must provide the demanded propulsion torque over the constraints until the
engine is turned on, whereas the mechanical brake is able to quickly respond to compensate for
the required braking torque.
                          P~
                               '
                                          MX. {*.»


 Figure 5.181 2013 Prius PHEV Battery Output Power According to SOC based on Test Cycles (data source
                                         APRF)
                                             5-486

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   If the engine is turned on and the desired battery power is calculated according to the
strategies in the previous two sections, the desired engine power can be calculated by the demand
power and the desired battery power. However, the engine operating target is not fixed because
the engine could operate at a number of operating points to produce the same power. Therefore,
the operating target of the engine should be controlled as well as a function of temperature.
Figure 5.182 shows the two different engine operating targets according to the coolant
temperature, which are  almost the same as the operating targets of the Prius HEV. The line that
can be inductively assumed from the red points in Figure 5.182 shows that the desired torque and
speed can be determined if the desired power is given.
                                   150   200   250   3UO
                                        Engine speed (rad's)
 Figure 5.182 2013 Toyota Prius PHEV Engine Operating Target Based on Test Cycles (data source APRF)
   Plug-in Hybrid Electric Vehicle - Range Extender PHEV

   The vehicle-level control strategy logic of a range extender PHEV was based on the GM Volt
Gen 1 PFIEV APRF test data analysis. The control implemented can be divided into four areas:
engine-on condition, transmission mode, battery SOC control during charge sustain mode, and
engine operating condition. If the battery is fully charged, a charge-depleting mode is selected,
wherein the battery is the main power source. Since it is considered that all driving should be
covered by "EV Drive," the vehicle is propelled by utilizing stored electric energy. If the battery
SOC drops to a predetermined level, a charge-sustaining mode is automatically selected. The
vehicle is then propelled by using a combination of the engine and battery while the SOC is
maintained.

   The engine  is  turned ON when the driver's demand power is over a threshold line, as shown
in Figure 5.183, where the demand power is determined by the wheel axle torque and current
vehicle speed.
                                             5-487

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                               0.205
                                      0.21   0.215   0.22
                                         Battery SOC
                                                       0.225
                                                              0.23
 Figure 5.183 Engine On Points - 2011 GM Volt PHEV Example Based on Test Cycles (data source APRF)

   The combined electric machine efficiency map and gear spin loss determines the EV drive
mode, such as EV1 and EV2. When the EV2 drive is in operation, the most efficient combination
of electric machine input speeds can be selected to meet the output speed and torque. With this
two-electric machine arrangement, electric machine speeds can be adjusted continuously, for
greatest tractive effort or greatest overall efficiency. The EV2 mode is used when the vehicle
speed exceeds a predefined threshold and the driver demands light load, as shown in Figure
5.184 in the gearbox (GB) axle torque - vehicle speed domain.
8
                         1000

                          800

                          600

                          400

                        -  200

                           °

                         -200

                         -400

                         -600
                                       10    15    20
                                        Vehicle Speed, mis
                                                         25
                                                               30
    Figure 5.184 EV Operating Mode - 2011 GM Volt PHEV Based on Test Cycles (data source APRF)

   In Figure 5.185, the mode selection rule is defined on the basis of the speed ratio, which is
defined as the ratio of the ICE input speed to vehicle speed. The power-split mode is used if the
speed ratio is low, which means that the system is changed from series mode to power-split
mode to avoid low system efficiency. In a high-speed ratio range, the system efficiency of the
power-split mode is low because electrical machines have relatively low efficiency. Low system
efficiency at a high-speed ratio range  can be avoided by propelling the vehicle by using the series
mode instead of the split mode. The EV2 drive and split operation offered by the Volt powertrain
system provides advantages over the more conventional EV drives and series operation.
                                             5-488

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                         450

                         400

                         350

                         300
•D
0>
£.250

c 200
'tn
c
m 150

  100

   50
                   .•'
                                             * Series Mode (Engine On)
                                             •  Split Mode (Engine On)
                                    10        20       30
                                        Vehicle Speed, mis
                                                               40
   Figure 5.185 HEV Operating Mode - 2011 GM Volt PHEV Based on Test Cycles (data source APRF)

   The desired battery power is linked to the energy management strategy. We found that the
battery power can be determined by the wheel power demand and the current SOC, as  shown in
Figure 5.186, when the vehicle is in HEV mode. The results are obtained by analyzing data
during HEV mode. Although some points are away from the line, the overall trend shows that the
energy management strategy tries to  avoid low power operation of the engine and bring the SOC
back to a regular range between  21 percent and 22 percent. Both the engine on/off control and
the battery power control are robust approaches to manage SOC in the appropriate range. If the
SOC is low, the engine is turned on,  and the power-split ratio is determined to restore the SOC to
a narrow range, so that the SOC  can be managed safely without depletion.
       100
        50
        -60
                                                  60
                                                  40
                                                  20
                                                1°
                                                01
                                                  -20
  09
                      246
                     Wti**l Power, W
             3    10
                X10'
                                                                           (EnaineOn)
0,216   022   0.226   0.23   0.235
        Battery SOC
   Figure 5.186 Battery Output Power - 2011 GM Volt PHEV Based on Test Cycles (data source APRF)

   The control concepts previously stated are used to determine the transmission mode and the
power-split ratio. The concepts do not, however, generate the engine target speed or torque
because the series and power-split system can de-couple the engine and wheels speed as long as
the output power demand is met, which provides greater flexibility to choose the engine working
point to optimize energy consumption. Therefore, an additional control concept to determine the
operating target is needed to complete the control strategy, for which engine speed operating
points are obtained according to the engine power, as shown Figure 5.187.
                                             5-489

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                          450

                          400

                        -2 350
                          300
                          250
                          150

                          100
Series Mode (Engine On)
 Split Mode (Engine On)
                                           3      4
                                        Engine Power, W
                                                            x10
  Figure 5.187 Engine Operating Targets - 2011 GM Volt PHEV Based on Test Cycles (data source APRF)

   In summary, the engine status is determined on the basis of the power demand or the need for
performance. If the engine is turned on, the desired power of the battery is determined on the
basis of the current SOC, and then the engine should provide appropriate power to drive the
vehicle. Finally, the engine operating targets are determined by a predefined line, and so the
controller can produce the required torque values for the electric machine and the generator on
the basis of the engine speed and torque target.

   Fuel Cell Hybrid Electric Vehicle

   Unlike the other vehicle-level controls previously discussed, the algorithm for the fuel cell
HEVs was not derived from test data, due to the lack of test vehicles. Instead, dynamic
programming was used to define the optimum vehicle-level control algorithms for a fuel cell
vehicle. A rule-based control is then implemented to represent the rules issued from the dynamic
programming. Overall, owing to the high efficiency of the fuel cell system, energy storage only
recuperates energy during deceleration and propels the vehicle under low-load operations — the
fuel cell system does not recharge the battery.  Unlike electric drive powertrains with an engine,
the battery does not smooth the transient demands. An example of fuel cell hybrid operations is
shown in Figure 5.188.
                                              5-490

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                Vehicle Speed (m/s)
               Component Powers (W) ,|
          Electric Machine
                                            Battery
            u 59

            0.56

            0.54

            0.52

            D5
 Figure 5.188 Component Operating Conditions of a FCV on the Urban EDC using Dynamic Programming
5.4.2.3 Vehicle Model Validation
5.4.2.3.1
Vehicle Benchmarking
   Benchmarking is commonly used by vehicle manufacturers, automotive suppliers, national
laboratories, and universities in order to gain a better understanding of how vehicles are
engineered and to create large datasets that can be applied in modeling and other analyses.
NHTSA has been leveraging the extensive existing vehicle test data collected by Argonne
National Laboratory under funding from the US DOE Vehicle Technologies Office.612 Specific
instrumentation lists and test procedures have been developed over the past 20 years to collect
sufficient information to be able to develop and validate full vehicle models. Over the coming
years, NHTSA intends to benchmark additional vehicles at the APRF to inform the Proposed and
Final Determination.

   Since its inception in the nineties 1, the APRF has been focused technology assessment of
advanced technology vehicles for the U.S. Department of Energy and its partners through the
generation  and analysis of laboratory data. The staff also supports the development of
automotive standards through its expertise and public data. The team has tested a large number
of vehicles of different types, such as advanced technology conventional vehicles, hybrid electric
vehicles, plug-in hybrid electric vehicles, battery electric vehicles, and alternative fuel vehicles.

   The researchers at the APRF have developed a broad and fundamental expertise in the testing
of the next generation of energy-efficient vehicles. Over the last twenty years, many methods of
vehicle instrumentation and evaluation have continuously been refined. The instrumentation
                                             5-491

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
intends to capture component level information while the powertrain is in the vehicle. This "in-
situ" instrumentation and testing approach enables the APRF is capture vehicle level and
component level data over dynamic drive cycles as well as specific powertrain mapping tests.

   Instrumentation approach

   Two levels of instrumentation and testing exist today. The first level (Level-1) involves
comprehensive, but non-invasive, instrumentation of a vehicle, leaving the vehicle unmarked
after the testing. The second level (Level-2) involves comprehensive invasive instrumentation of
a vehicle and its powertrain components, which leaves the vehicle with irreversible alterations,
but provides an in-depth assessment of the technology. The goal of the instrumentation is to
provide usage information and efficiencies (if possible) of the different powertrain components,
operating envelops, and powertrain behavior.

   Typically, Argonne receives Level  1 test vehicles on loan; therefore, the vehicles need to
leave the test facility in the "as-received" and road worthy condition. This requirement limits
instrumentation to sensors that can be  easily installed  and removed without leaving any damage.
The Level 2 benchmark, which included in-depth, testing, and analysis of new and emerging
vehicle technologies, is specific to each vehicle. If the vehicle has an internal combustion engine,
instrumentation is applied to measure the engine speed, fuel flow and engine oil temperature. For
electrified vehicles, a power analyzer is used to record the voltage and current from the high
voltage energy storage system. If the vehicle requires  charging, the electric power from the grid
to the charger is measured. The recording of messages from the vehicle's information buses
(diagnostic and broadcast network messages) is another expertise of the APRF staff. The
instrumentation is focused to a particular technology,  or technologies that enable the increased
energy efficiency of a powertrain.

   Facility capabilities

   The APRF has a 4WD wheel drive chassis dynamometer and 2WD chassis dynamometer. The
4WD chassis dynamometer is in in a thermal chamber to evaluate the powertrain across a range
of environmental conditions. The thermal chamber and an air-handling unit with a large
refrigeration system that enables vehicle testing from  -20°C to  40°C. All  temperatures can be
evaluated with or without solar emulation lamps providing up to 850 W/m2  of radiant sun
energy. Some highlights of the APRF  capabilities include: rated to test hydrogen powered
vehicles; 5  cycle capable; several emissions measurement systems; and research focused test
cell.613 Figure 5.189 illustrates the two chassis dynamometer test cells available at the APRF.
                                             5-492

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
 Figure 5.189 Illustration of testing at 95°F with sun emulation (left) and at 20°F cold ambient temperature
                                         (right).

   The APRF benchmark program goes well beyond the standard tests performed for EPA
certification of fuel economy and emissions. To fully characterize the powertrain and the
individual components the instrumented powertrains are tested on a wide range of ambient
temperatures, drive cycles, performance tests and vehicle/component mapping tests.

   Independent and Public Data

   A major goal of the benchmarking activity is to enable petroleum displacement through data
dissemination and technology assessment. The data generated from the vehicle testing as well as
the analyses are shared through several mechanisms, such as raw data, processed data,
presentations and reports.

   The independent and public data is a foundation enabling the development of rigorous and
technology neutral codes and standards. The data also serves to develop and validate several
modeling and simulation tools within the DOE system (i.e., Autonomie) as well as outside (i.e.,
EPA Alpha model, University modeling,  and economic  models). These activities in turn impact
the modification of test plans and instrumentation for current and future test vehicles. Partners in
the testing include U.S. manufacturers and suppliers, through the U.S. Council for Automotive
Research. Many of the research activities of the  DOE rely on the benchmark laboratory and fleet
testing results to make progress towards their own goals. Figure 5.190 details some of these DOE
research activities and partners.
                                             5-493

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
              USDrive Tech
              team support


               OEM /
               supplier
              collaboration
                                     Dynamometer
                                     Downloadable —>•  Online
                                       Database
EVs
PHEVs
                                                                        DOE Activities
                                                                         Independent
                                                                      technology validation
                                                  Baseline for
                                               component technical
                                                 target and goal
                                                   setting
             DATA AND ANALYSIS
                                                 EV testing and
                                               charging evaluation
                                                                         ORNL missions
                                                                         ^^^^^^^H
                                                                       NREL AC evaluation
                              Codes and
                              standards
                            (Data for procedures
                        Modeling support
                        (Data and validation)
                         Figure 5.190 Data Dissemination and Project Partners.

   Downloadable Dynamometer Database (D3)

   An additional avenue for data distribution is Argonne's Downloadable Dynamometer
Database (D3).614 The D3 website provides access to a subset of data and reports.

   D3 is a public web portal of highly detailed accurate public and independent vehicle test data,
of critical utility in the research community. This web-based portal to Argonne vehicle test data
is designed to provide access to dynamometer data that are typically too expensive for most
research institutions to generate. Shared data is intended to enhance the understanding of system-
level interactions of advanced vehicle technologies for researchers, students, and professionals
engaged in energy-efficient vehicle research, development, or education.  Figure 5.191 shows the
structure and content of the database.
  Database Landing Page
    Select your vehicle type
Vehicle Type Page
  Select your vehicle
                                    V, r,,J
                    Vehicle Page
                 Download data and analysis
                                                           Test
                                                                            lOHz Data signals   Analysis
                                                                                            Reports
                                                             Vehicle information
                                                             Fuel Economy per cyde
                                                             Energy consumption per
                                                             cycle (DC Wh)
                                                             Recharge energy (AC Wh)
                                                      Drive cyde speed, Vehicle
                                                      speed, Tractive effort
                                                      Temperatures: ambient
                                                      cabin, engine oil
                                                      Fuel flow, engine speed
                                                      Battery voltage, current
 II
 Battery
performance
                      Figure 5.191 Map of Downloadable Dynamometer Database.
                                                   5-494

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.3.2      Vehicle Validation Examples

   Argonne has been validating the Autonomie vehicle models with vehicle test data for more
than 15 years. Test data were collected at the Argonne National Laboratory APRF from more
than 60 vehicles, spanning model years 2000-2015. A large number of signals were collected on
each vehicle with specific focus on model development and validation. While sensors were
different across vehicles, they included: torque sensors (axles); components speeds; coolant flow
sensors; coolant component temperatures; exhaust temperatures; emissions; fast CAN data; scan
tool data; power analyzer on many nodes; dynamometer loads and speeds; and direct fuel
measurement. These readings were all integrated into one data acquisition system. Some
additional parameters were then estimated based on measured data and other advanced
technology vehicles. After each individual model was independently validated, vehicle system
models were developed and the validation quality was quantified using normalized cross
correlation power (NCCP). JJJJ Vehicles were tested over a large number of cycles and runs. For
example, the MY2010 Toyota Prius HEV was run on 11 separate cycles for a total of 26
tests.KKKK

   Autonomie vehicle models have been validated within test to test repeatability for a wide
range of technologies and powertrain configurations.  The following section highlights some of
the validation performed using Argonne APRF vehicle test data. While much work has been
performed at Argonne under DOE VTO funding, NHTSA is currently evaluating the ability to
perform additional  vehicle benchmarking activities on specific vehicles, focusing on
conventional powertrains.

   NHTSA is also very much aware that subtle differences between modeled and physical shift
schedules can impact vehicle energy consumption. Some of these differences can be due to drive
quality limitations amongst other constraints. While numerous constraints have been already
taken into account (i.e., shift frequency), NHTSA welcomes any feedback that would contribute
to improving the accuracy  of the shifting  algorithm, especially for future technologies that are
not currently in the market.

5.4.2.3.2.1     Transmission Shifting A Igorithm

   As discussed in  Section 5.4.2.2.3.1,  a generic shifting algorithm has been developed,
continuously improved and validated over the past 15 years. When new transmission
technologies are introduced in the market, that algorithm is regularly validated with the latest test
data. This section highlights how the algorithm logic was modified when 8 speed automatic
transmissions were introduced. Preliminary analysis led to the development of a new calibration
and algorithm for 8 speed transmissions as the initial algorithm developed and validated for 6
speed transmissions did not provide sufficient accuracy.

   Figure 5.192 shows the simulation results of the vehicle speed, the engine speed, and the
engine torque in UDDS compared with testing results for both shifting algorithms.

   Initial shifting initializer (simulation l),New algorithm and calibration (simulation 2)
JJJJ See SAE 2011-01-0881, "Test Correlation Framework for HEV System Model," Ford Motor Company
KKKK The prius was evaluated on the following cycles: UDDS, LA92, NEDC, JC08, NYCC, SC03, Accels, cycle
  505, Highway, US06, and SS.
                                             5-495

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
                          60

                          50

                        | 40

                        TJ- 30
                        •

                        <|ZO

                          10

                           0

                          400


                          300
                       J»
                       -o
                       TO

                       •d  20°
                       •
                       0}

                       "100
                         250
                                           UDDS-Vehicle Speed
        	Vehicle Speed (Simulationl)
        	Vehicle Speed (Simulation2)
        	Vehicle Speed (ANL Test)
                         200

                         150

                         100

                          50
                      UDDS - Engnie Speed
                            	Engine Speed (Simulation-!)
                            	Engine Speed (Simulation^)
                            	Engine Speed (ANL Test)
                                          UDDS -Engnie Torque
                             — Engine Torque (Simulation 1)
                          I	Engine Torque (Simulation2)
                           	Engine Torque (ANL Test)
                                            200       300
                                               Time, sec
  Figure 5.192 2013 Sonata 6ATX Simulation and Testing Results on UDDS (0-505 s) (data source APRF)
   In Figure 5.193, the gear numbers over the HDDS (0-505 s) are compared with the test data
for two transmission types (6 speed and 8 speed). The first is a 2013 Sonata conventional 6ATX
(left) and the second is a 2013 Chrysler 300 8ATX (right). Both simulations show closed shifting
performance with the test results, but the results of simulation with the new algorithm show
higher accuracy than those of the current algorithm, especially for the eight speed transmission.
     •
   3,
     e
   .. I
   I 4
   1;
                   UODS-fttirnumbtr
                         	C««rnumb«f (SlmuUttonl)
                          — Our number (ANL Tut)
    -	G«§r number (Slmul«lon2)
     •-- e*irnumlMr(ANL T«»t|
            ITU
Vi^-ftrt/v--"	*f-4 Thr
, I i I i I |I I     i' '  El
11   1:01
          200
                                1000   1200
                                                                  UOOS-C*«r number
I-
                                                   	G«»r number ISlmulMon!)
                                                      Ge.r number |ANL T«.t|
At
i •«
J  u
                         G««r number |Slmulelon2)
                        -C««rnufflb*r(ANLT**t)
                                                          200
                                                                                     1200   1400
                       Tltn*. ««c
   Figure 5.193 Simulation and Testing Results for 6ATX (left) and 8ATX (right) (test data source APRF)
                                                 5-496

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
   In Figure 5.194, additional simulation results over the NEDC are compared with test data. In
this case, a conventional 2012 Fusion with a 6ATX transmission (right) is compared with a 2013
Chrysler 8 ATX (left).
                  NEDC -Vehlek Sp««d
  i
  I"
  $
         - Vehicle Speed (SimuUtionI)
          Vehicle Speed (Simulation!)
         • Vehicle Speed (ANL T*«Q
    20
   w «
   t,
      I-
                   NEOC -Gear number
                      	Gear number (Simula on1)
                      	Ceir number (ANL Tett)
                   NEOC-Gearnumber
                       	Gear number (Slmul I.on21
                       	Gearnumber(ANL Te»0
                                                                NEOC -Gearnumber
                                                §4

                                                I
                                                  2
                                                5
- Gear number (Slmulalonl I
• Gear number (ANL Te»t)
 Gear number (Simula«onj)
 Gearnumber(ANLT**Q
           200
                  400     600
                      Time, tec
                               too
                                     1000
                                                        200
        400     600
             Tim * >CC
                                                                            800
                                                                                  1000
     Figure 5.194  Comparison of Simulation and Test Results over the NEDC (test data source APRF)
   The CVT model and shifting control strategy developed in Autonomie were validated by
comparing the simulation results with the experimental results from Argonne ANL's APRF for
multiple vehicles. Figure 5.195 shows the validation results for the target HEV system on the
UDDS (city driving on left) and HWFET (highway driving on the right) cycles for the 2012
Honda Civic CVT.  The CVT shift dynamic model was validated by comparing the CVT gear
ratios: the simulation result for the CVT gear ratio agreed well with the experimental result. The
battery was charged or discharged according to the driving mode control strategy. The simulated
vehicle speed, gear ratio, engine torque and battery SOC are comparable with the experimental
results,  demonstrating the validity of the simulation model and control strategy.
                                               5-497

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                                                     10   20   300   400   500    600   700
                                                     1 0   200   300   400   500   600   700
   Figure 5.195 Simulation and Test Results Compared for a Honda Civic HEV (test data source APRF)

5.4.2.3.2.2   Power split HEV

   The power-split HEV model was validated under different thermal conditions. An example of
a comparison between the simulation results and the test data for engine operating points is
shown in Figure 5.196 for the 2010 Toyota Prius HEV. In Figure 5.196 (left), the engine
operating points obtained from simulation results are close to the test data, especially for engine
ON/OFF conditions. In addition, the energy consumption and the SOC behavior are also close to
the test data.
  Figure 5.196 Simulation and Testing Results over the UDDS for 2010 Toyota Prius HEV (test data source
                                          APRF)
                                              5-498

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.3.3
Pre-transmission HEV
   The pre-transmission HEV control logic was validated using Argonne's APRF test data from
the 2013 Jetta DCT Hybrid. Comparing the simulation results for the vehicle speed, gear
number, and battery SOC on the HDDS cycle with test results, as shown in Figure 5.197, showed
good correlation.
                                         UDDS-Vehicle Speed
   Figure 5.197 Simulation and Testing Results over the UDDS for 2013 Jetta DCT HEV (test data source
                                         APRF)
                                             5-499

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.3.3.1    Range Extender PHEV

   The range extender PHEV model was validated under different thermal conditions using
Argonne's APRF test data from the Gen 1GM Volt. The vehicle speed, component speed, and
component torque under normal ambient temperature were successfully compared with the
testing results shown in Figure 5.198.
                 UODJCJ|I2:! OSflC|-VeHCIe Speed
                                                                         HC1 Speed|Slmuaton|
                                                                         MCI speed nert
                              Engine Torque 1S mutalon}
                              Engine Torque
nt8tttdFuei|SimuHton|
InfegrafedFuellTeitl
            200    440
                                                               CM   w«
                                                                Time. «ec
   Figure 5.198 Simulation and Testing Results over the UDDS for 2011 GM Volt PHEV (test data source
                                          APRF)

   In Figure 5.199, the simulated SOC for a 2011 GM Volt PHEV over the HDDS matches well
with the testing results during the first 200-seconds, since the controller tends to maintain the
engine turned on to warm up the engine, and so the results of simulation show an increase in the
SOC at the start of the engine. In addition, the simulation results show that the pattern of the
coolant temperature is similar to that from test under normal ambient temperature.
                                              5-500

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                             Battery Power |ilmuaioti |
                             Battery Power fTertl
                           	Batbiv SOC |ilmuatkjn|
                           	Batbrj JOC flesU
Tempenatjre, oc
_S S 8 B 8 %















	 a

liteiy CootantTemp |4linii&ton}
ittsry CooHntTemp (Tett|
i i _^



*—* ^~. _^


	 i





Ml 4** SCO 800 1000 1200 U(
                                                                TT* H;
     Figure 5.199 Simulation and Testing Results for a 2011 GM Volt PHEV (test data source APRF)

5.4.2.4 Simulation Modeling Study Overview

   It is widely acknowledged that full-scale physics-based vehicle simulation modeling is the
most thorough approach for estimating future benefits of a package of new technologies. This is
especially important for quantifying the efficiency of individual technologies and their synergies
with other, especially for those that do not currently exist  in the fleet or as prototypes.
Developing and executing tens or hundreds of thousands of constantly changing vehicle
packages models in real-time is extremely challenging. While this approach was until recently
considered generally not practical to implement, the process developed by Argonne in
collaboration with NHTSA and the Volpe Center does just exactly that. This approach offers
multiple advantages, including the ability to apply varying levels  of technologies across the
vehicle fleet to account for the full range of vehicle attributes and performance requirements.

   As part of rulemakings, the objective of the modeling described in this section is to simulate
all of the possible technology combinations in the Volpe model and eliminate the use of synergy
factors. The result of this work is a comprehensive understanding of the impact of combined
vehicle technologies on energy consumption.  To achieve this objective, individual vehicles were
simulated to represent every combination of vehicle, powertrain, and component technologies
considered for the assessment. The sequential addition of these technologies to the five vehicle
classes currently considered results in 140,000 unique vehicle combinations. In addition,
powertrain sizing algorithms needed to be run in Autonomie to ensure similar vehicle
performances, resulting in over one million simulations.

   GT POWER simulation modeling of engine technologies was conducted by IAV Automotive
Engineering,  Inc. (IAV). GT-Power is a commercially available engine simulation tool  with
detailed cylinder model and combustion analysis.  GT-POWER is used to predict engine
performance quantities such as power, torque, airflow, volumetric efficiency, fuel consumption,
turbocharger performance and matching and pumping losses, and other parameters. Engine maps
resulting from this  analysis were then used by ANL in Autonomie.

   The current vehicle system simulations included:

       •   5 vehicles Classes (Compact, Midsize, Small SUV, Midsize SUV, Pickup)
       •   14 engine technologies
                                             5-501

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
       •   11 electrification levels (conventional is equivalent to no electrification level)
       •   9 transmissions technologies (applied to Low Electrification Level Vehicles only)
       •   6 Light Weighting levels
       •   3 Rolling Resistance levels
       •   3 Aerodynamics levels
   NHTSA is planning to simulate all the vehicle classes considered in the Volpe model,
including high performance vehicles in the near future. In addition, NHTSA is considering
adding new component technologies based on feedback from the Draft TAR and on-going and
future benchmarking activities.

   The process developed includes the following steps as shown in Figure 5.149:

       1)  Collect/develop all the technology assumptions
       2)  Create fuel maps for engine technologies.
       3)  Develop a process to automatically create the vehicle models.
       4)  Size the individual vehicles to all meet the similar vehicle technical specifications
           (note that some vehicles inherit component and energy  from previous decision tree
           steps).
       5)  Simulate each vehicle model on the standard driving cycles.
       6)  Create a database with all the required input for the Volpe model.
       7)  Create post-processing tool to validate the database content.
   Since this process has to be performed in an acceptable amount of time, distributed computing
was extensively used for vehicle sizing and simulation

   The remaining subsections of this  chapter describe each step of the analysis methodology.

5.4.2.5 Selection of Technologies for Modeling1111

   Table 5.218 lists the engine, transmission, and vehicle technologies simulated in this  study.
                    Table 5.218 Vehicle and Powertrain Technologies Evaluated
Engine Technologies
Variable Valve Timing
Variable Valve Lift
Stoichiometric gasoline direct injection
Cylinder Deactivation
High Compression Ratio
Engine Friction Reduction
Turbocharging and downsizing
Stoichiometric Exhaust Gas Recirculation
Downspeeding
Cooled EGR
Miller Cycle
Advanced Diesel
Improved turbocharger efficiency
Drivetrain Technologies
6-Speed Manual Transmission
7-Speed Manual Transmission
6-Speed Automatic Transmission
8-Speed Automatic Transmission
Continuously Variable Transmission
6-Speed Dual Clutch Transmission
8-Speed Dual Clutch Transmission
Secondary Axle Disconnect
Stop Start 12 Volt
Mild Alternator Regenerative Braking
48 Volt Belt ISG
100 Volt Crank ISG
Strong Hybrid Power Split
       ajj Of ^ technologies in the Volpe model decision tree were evaluated by Argonne. Compressed natural
  gas, liquid natural gas, liquid propane gas, and LGDI were not modeled by Argonne and are not included in Table
  5.218.
                                              5-502

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
Injection pressure increase
Downspeeding with increased boost pressure
Closed loop combustion control
Low pressure EGR

Vehicle Technologies
Aerodynamic Drag Reduction
Mass Reduction
Improved Tire Rolling Resistance
Low Drag Brakes
Strong Hybrid P2
Plug-in Hybrid (30 mile all-electric range)
Plug-in Hybrid (50 mile all electric range)
Electric Vehicle (200 mile range)
Fuel Cell Vehicle
Improved Accessories
Electric Power Steering
Electric Water Pump
Electric Cooling Fan
High Efficiency Alternator
 5.4.2.6 Modeling Assumptions

    Section 5.2 presented the agencies'joint assessment of the current state of technologies and
 the advancements that have occurred since the publication of the FRM. As stated earlier, the
 agencies have reexamined every technology considered in the FRM, as well as assessing some
 technologies that are currently commercially available but did not play a significant role in the
 FRM analysis, as well as emerging technology for which enough information is known that it
 may be included in this Draft TAR. The categories of technologies discussed in Section 5.2
 include: engines, transmissions, electrification, aerodynamics, tires, mass reduction, and other
 vehicle technologies such as improved accessories and low drag brakes. For a descriptions of
 these technologies, please refer to that section. This section adds information specific to the
 NHTSA CAFE analysis of engines, transmissions, electrification, aerodynamics, tires, mass
 reduction, and other vehicle technologies.
 5.4.2.6.1
Vehicle Level
    Table 5.219 provides the reference specifications used for the five vehicle classes modeled by
 ANL. The vehicles were sized to meet each vehicle technical specification (for example
 performance and range for electric vehicles).
                Table 5.219 Reference Vehicle Assumptions for all Classes in Autonomie
Wheel mass (kg)
Wheel radius (m)
Glider mass (kg)
Frontal Area (m2)
Drag Coefficient
Rolling resistance
Electrical Base Ace Load (W)
EXTRA: Electrical Ace Load for
cooling for EV & PHEV 30&40 (W)
Fuel Tank Size for Conventional (gal
Fuel Tank Size for HEV/PHEVs (gal)
Fuel Tank size for Fuel Cell
Compact Car
85
0.31725
820
2.3
0.32
0.0075
240
220
Midsize Car
85
0.31725
1000
2.4
0.31
0.008
240
220
Small SUV
90
0.35925
1150
2.8
0.36
0.0084
240
220
Midsize SUV
95
0.3677
1260
2.9
0.37
0.0084
240
220
Pickup
95
0.38165
1500
3.3
0.45
0.009
240
220
               12
               10
               320 miles
17
13
320 miles
17
13
320 miles
22
17
320 miles
26
20
320 miles
    Autonomie has multiple driver and chassis models that can either use vehicle dynamometer
 coefficient or the aerodynamic equations. The first option is usually only selected when
                                               5-503

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
performing vehicle validation. The aerodynamic equations, leveraging Cd, FA, and Crr, were
used to perform all simulations.

   While only five vehicle classes were simulated by ANL, the Volpe model includes additional
vehicle classes. As a result, effectiveness results from the two non-modeled classes have been
defined based on results from the five modeled classes. In the next round of simulations, all the
vehicle classes required by the Volpe model will be simulated with Autonomie.

5.4.2.6.2     Gasoline and Diesel Engines

   IAV provided wide-open-throttle engine performance values and brake-specific fuel
consumption (BSFC) maps for the engine technologies listed in Table 5.218 IAV validated the
GTPower model with existing dynamometer measurements for several engines. The models were
trained over the entire engine operating range and have predictive combustion capability. This is
essential, since the BSFC prediction needs to be accurate while the engine setup is subject to
change.

   Relevant engine geometries/parameters were measured and modeled with friction/flow losses,
heat transfer, and other parameters and calibrated to match measurements. Displacement
normalized mechanical friction was modeled as a function of engine speed and specific load. A
combustion model was trained to predict fuel heat release rates in response to physical effects
such as cylinder geometries, pressure, temperature, turbulence, residual gas concentration and
other parameters. A knock correlation based on in-cylinder conditions and fuel octane rating
predicts if knock will occur and at what intensity. A combustion stability threshold prediction
was trained using covariance of IMEP data and is used for understanding EGR tolerance,
especially at low loads. Load controllers were developed for fuel/air path actuators and targeting
controllers drive optimal and knock limited combustion phasing just as in a physical engine.
Careful modeling practice was used to provide confidence that calibrations will scale and predict
reasonable/reliable as parameters are changed throughout the various technology concept studies.

   IAV provided 14 engine maps in total: eight of these are naturally aspirated gasoline  engines,
five are turbocharged gasoline engines, as well as one diesel engine. One naturally aspirated
gasoline engine map was developed based on benchmarking of a 2014 SKYACTIV-G 2.0L
engine from a Mazda 6 by EPA. Finally,  one Atkinson engine map generated using Argonne test
data was used for electrified vehicles with power split architecture. Thus, the total number of
engine maps used in the study is 16.

   For all engines, engine speed, BMEP,  brake torque, fuel flow rate, PMEP and FMEP data
were provided in a standardized format to Argonne. These channels were provided from 1,000
RPM to the max engine speed and from 0 bar BMEP to full load to provide a full operation map.
Fuel flow rates at zero output torque were provided separately  from 650  RPM (defined idle) to
6000 RPM. Negative torque data was also provided which included the minimum fueled torque
curve from the baseline engine concept; 1) unfueled motoring curves from the baseline concept;
and 2) unfueled motoring curve from cylinder deactivation concept at wide open throttle. IAV
used gasoline with LHV = 41.3 MJ/kg for the mapping but the naturally aspirated engines were
calibrated with 87 (R+M)/2 rating fuel and the turbocharged engines used 93 octane fuel. IAV
did not use certification fuel and so ANL adjusted the vehicle fuel economy results to represent
certification fuel by using the ratio of the lower heating values of the test and certification fuels.
Values for brake specific fuel consumption at different engine loads are shown in Figure 5.200.
                                             5-504

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   IAV Engine 1 is a naturally aspirated PFI 2.0-L gasoline engine with VVT from a MY2013
vehicle. A brake specific fuel consumption (bsfc) engine map was generated from dynamometer
testing of the existing engine, which then served as the baseline map for all simulated naturally
aspirated engines (Engines l-8a). Figure 4 shows the 2L, 4-cylinder naturally aspirated PFI with
DOHC and dual cam VVT. The engine calibrations were fully optimized for best BSFC and
maximum torque.

   Each subsequent engine (bsfc map) represents an incremental increase in technology advance
over the previous technology. Engines 2-4 add variable valve lift (VVL), direction injection (DI),
and cylinder deactivation (deac) sequentially to the base engine map. Engine 5a converts Engine
1 from DOHC to SOHC. Engines 5b, 6a, 7a, and 8a add some friction reduction to Engines 5a, 2,
3 and4MMMM
              1,
              I.
                    bsfc
                  1000
                                     3000        4000
                                        speed |rpm|
                           Figure 5.200 IAV Gasoline Enginel Map

   For Engine 2, a VVL system was added to the intake valves to Engine 1. Both valve lift and
timing were optimized. This engine allows for reduced pumping work at low loads and more
torque at low speeds due to reduced intake duration.
MMMM In stage 1, FMEP is reduced by 0.1 bar and in level 2 FMEP is reduced by 25 percent over the entire operating
  range.
                                             5-505

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                    3000     4000
                      speed [rpm[
                                                            vs Eng01 - Efficiency difference
                                                     1000   2000    3000   4000   5000
        Figure 5.201 IAV Gasoline Engine! Map (left), Incremental Improvement vs Engl (right)
   PFI Engine 2 was converted to direct injection to model engine 3. The compression ratio was
raised from 10.2 to 11.0 and injection timing optimized. Direct injection provides greater knock
tolerance, allowing higher compression ratio and increased efficiency over entire map.
                                                       Eng03 vs Eng02 - Efficiency difference
        Figure 5.202 IAV Gasoline Engines Map (left), Incremental Improvement vs Eng2 (right)

   Cylinder deactivation was added to engine 3 to model engine 4. This technology deactivates
the intake and exhaust valves and prevents fuel injection into some cylinders during light-load
operation. The engine runs temporarily as though it were a smaller engine which substantially
reduces pumping losses. For 4 cylinder applications, the engine fires only 2 cylinders at low
loads and speeds below 3000 RPM and less than 5  bar BMEP by deactivating valves on 2
cylinders. The main benefit is that the effective load is doubled on 2 cylinders providing less
pumping work and higher efficiency.
                                              5-506

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                                                         Eng04 vs Eng03 - Efficiency difference
                                                     1000    2000   3000    4000   5000   6000
        Figure 5.203 IAV Gasoline Engine4 Map (left), Incremental Improvement vs Eng3 (right)

   Engine 5b was developed to assess the benefit of reduced friction. Reduction in engine
friction can be achieved through low-tension piston rings, roller cam followers, improved
material coatings, more optimal thermal management, piston surface treatments, cylinder wall
treatments  and other improvements in the design of engine components and subsystems that
improve engine operation. A SOHC engine with VVT was used and its FMEP reduced by 0.1 bar
over its entire operating range. Valve timing was optimized for fixed overlap camshaft.
                                                       EngSb vs Eng01 - Efficiency difference
            *--«'—y-
        Figure 5.204 IAV Gasoline EngineSb Map (left), Incremental Improvement vs Engl (right)
                                              5-507

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
Engine 6a was developed to assess the friction reduction impact on Engine 2. Reduced friction will
              improve efficiency at all load points as well as raise the full load line.
                                                          Eng6a vs EngSb - Efficiency difference
                                                     1000    2000    3000    4000    5000    6000
     Figure 5.205 IAV Gasoline Engine6a Map (left), Incremental Improvement vs EngSb (right)



Engine 7 a was developed to assess the friction reduction impact on Engine 3.

                                                          Eng7a vs Eng6a - Efficiency difference
                                                                                         • -;
                                                                                        |
                  3000      4000
                    sp*«d |(pm]
                                                      1000    2000    3000    4000    5000    6000
     Figure 5.206 IAV Gasoline EngineTa Map (left), Incremental Improvement vs Eng6a (right)

           Engine 8a was developed to assess the friction reduction impact on Engine 4.
                                                           EngSa vs Eng7a - Efficiency difference
                                                      1000   2000    3000    4000    5000    6000
     Figure 5.207 IAV Gasoline EngineSa Map (left), Incremental Improvement vs Eng7a (right)
                                                5-508

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   IAV Engine 12 is the base engine for all the simulated turbocharged engines (Engines 13-16)
and was also validated using engine dynamometer test data. Turbocharging and downsizing
increases the available airflow and specific power level, allowing a reduced engine size while
maintaining performance. This reduces pumping losses at lighter loads in comparison to a larger
engine. Engine 12 represents a 1.6L, 4 cylinder turbocharged, direct injection DOHC engine with
dual cam VVT and intake VVL. A compression ratio of 10.5:1 was used along with side
mounted direct fuel injectors and a twin scroll turbocharger. The calibrations were fully
optimized for best BSFC. Figure 5.208 shows fuel consumption at given engine speeds and
loads.
                    XflO     4000
                      speed |rpm]
        Figure 5.208  IAV Gasoline Enginel2 Map (left), Incremental Improvement vs Engl (right)

   Engine 12 has been further downsized to a 1.2L to represent engine 13. The turbocharger
maps scaled to improve torque at low engine speeds. All the turbocharged direct injection
engines described below have been developed using 93 octane. NHTSA understands that using
such fuel might lead to overestimating the effectiveness of the technology, especially for high
BMEP engines. While the engine maps will be updated to represent regular grade octane
gasoline, NHTSA does not expect significant effectiveness change on the standard driving cycles
as the engines operate at lower loads.
                                                       Eng13 vs Eng12 - Efficiency difference
       Figure 5.209 IAV Gasoline EnginelS Map (left), Incremental Improvement vs Engl2 (right)
   High pressure cooled EGR was added to engine 13 to develop engine 14. Exhaust gas
recirculation boost increases the exhaust-gas recirculation used in the combustion process to
                                             5-509

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
increase thermal efficiency and reduce pumping losses. Levels of exhaust gas recirculation
approach 25 percent by volume in these highly boosted engines (this, in turn raises the boost
requirement by approximately 25 percent). Cooled EGR target set points were optimized.
     i;
                                                         Eng14 vs Eng13 - Efficiency difference
        WOO     2000
                                                 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500
       Figure 5.210  IAV Gasoline Enginel4 Map (left), Incremental Improvement vs Engl3 (right)

   Engine 14 was further downsized to l.OL to develop Engine 15. Cooled EGR target set points
were re-optimized and turbocharger maps were re-scaled. Downsizing with cooled EGR reduces
in-cylinder temperatures and knock, and lower the need for enrichment to protect emission
control devices.
                                                       Eng15 vs Eng14 - Efficiency difference
                                                                                   -80

                                                                                   -100
                    3000     4000
                     sp««l [rpm]
                                                        2000    3000    4000   5000
       Figure 5.211  IAV Gasoline EnginelS Map (left), Incremental Improvement vs Engl4 (right)
   Engine 15 was converter to a 3 cylinder l.OL concept to develop engine 16. To do so, intake
and exhaust piping were scaled to account for larger mass flows through each cylinder and
cooled EGR target set points were re-optimized.
                                               5-510

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                                                         Eng16 vs Eng15 - Efficiency difference
                     3000     4000
                      speed [tpm\
                                                 500  1000 1500 2000 2500 3000 3500 4000 4500 5000 5500
       Figure 5.212  IAV Gasoline Enginel6 Map (left), Incremental Improvement vs EnglS (right)

   Figure 5.213 shows the engine map for the diesel engine. Diesel engines have several
characteristics that give superior fuel efficiency, including reduced pumping losses due to lack of
(or greatly reduced) throttling, and a combustion cycle that operates at a higher compression
ratio, with a very lean air/fuel mixture, than an equivalent-performance gasoline engine. This
technology requires additional enablers, such as NOx trap catalyst after-treatment or selective
catalytic reduction NOx after-treatment. For the diesel engine, measured data, including engine
speed, BMEP, brake torque, brake power, BSFC channels were provided.
                                                          Eng17 vs Eng16 - Efficiency difference
                      2SOO
                     speed [rpm]
                                                 1000    1500   2000   2500    3000   3500    4000
        Figure 5.213 Diesel IAV Enginel? Map (left), Incremental Improvement vs Engl6 (right)
   The last engine modeled for conventional powertrains was a high compression ratio engine.
Higher compression ratio improves piston power stroke while helping to prevent knock.
Atkinson cycle engines combine an increase in compression ratio and variable intake camshaft
timing. Although producing lower overall power for a given displacement, this engine has
specific high efficiency operating points and is capable of significant CCh reductions when
properly matched to a strong hybrid system. The engine map was developed based on the 2014
SkyActiv 2.0L engine from a Mazda 6 using test data collected by the U.S. EPA and is shown in
Figure 5.214.
                                              5-511

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                       0008


                       kOOOC


                       J0004
                                         *
                                                           S
                                             "**•••    2000
                                                1000
       Figure 5.214. High Compression Ratio Engine Map Developed From Dynamometer Test Data

   Atkinson engine technology was also used for power split hybrid powertrains. The engine
map was developed based on APRF test data and published literature. It is important to note that
pre-transmission hybrids as well as multi-mode hybrids have also been simulated. In those cases,
all the engine technologies previously described have been considered.

   NHTSA is planning to continue to work with IAV to update the existing engine maps for the
technologies considered so far as based on feedback and comments received as part of the Draft
TAR as well as develop new high fidelity models for additional technologies to represent
potential future technologies. NHTSA will ensure that all future engine model development is
performed with regular grade octane gasoline. NHTSA will also continue to gather information
on the latest engine technologies, both from public and proprietary sources, to compare the
effectiveness each of those specific OEM engines with the GTPower models.

5.4.2.7 Description of Engine Technologies Evaluated

   This next sections provides NHTSA-specific details on the engine technologies modeled in
the gasoline and diesel engines. Please refer to section 5.2 for a general description of variable
valve timing and lift, friction reduction, EGR, and developments in the technologies since the
publication of the FRM.
5.4.2.7.1
Friction reduction
   Friction reduction has been shown to offer significant improvements in vehicle fuel
consumption. Engines were subjected to two levels of reduction in friction mean effective
pressure.

       1) A reduction in FMEP by 0.1 bar across the entire engine speed range.
       2) An extreme friction reduction (25 percent FMEP) across the entire speed range.
   In the Volpe modeling, only the first level of friction reduction has been considered.
Predictive friction equation was calibrated from test data used in Engines l-8b to allow for a
smooth and systemic friction study but may under predict FMEP at high loads with late
combustion phasing. Map based FMEP lookup compiled from test data was used for Engines 12-
16. Due to different methods, we cannot draw direct conclusions on naturally aspirated vs.
downsized engine friction.
                                             5-512

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.7.2
Cylinder Deactivation
   Cylinder deactivation operates the remaining, firing cylinders at higher BMEP under light
load conditions. This moves operation of the remaining cylinders to an area of engine operation
with less throttling and thus lower pumping losses and reduced BSFC.

   Two separate engine maps are used to model the cylinder deactivation benefits. A logic
described below is then used to decide when to use or not the functionality. Due to NVH
considerations, cylinder deactivation operation is not performed in several vehicle operation
modes, such as vehicle warm-up, low gears, idle, and low engine speed in production
vehicles.NNNN Cylinder deactivation was disabled under the following vehicle and engine
conditions:

       1) If the engine is at idle or any  speed below 1000 RPM and above 3000 RPM.
       2) If the vehicle is in the 1st or the 2nd gear.
       3) If the engine load is above half the max BMEP of the engine (and a certain hysteresis
          is maintained to prevent constant activation and deactivation).
   Changes in the transmission shifting calibration (like lugging speed limits) and additional
torque converter slippage during cylinder deactivation have not currently been considered.
5.4.2.7.3
Turbocharged Engines
   In addition to the naturally aspirated engines, maps for turbo technologies were also
developed using GT-Power. With turbo engines, there is a 'lag' in torque delivery due to the
operation of the turbo charger. This impacts vehicle performance, and vehicle shifting on
aggressive cycles. Turbo lag has been modelled in Autonomie for the turbo systems based on
principles of a first order delay, where the turbo lag kicks in after the naturally aspirated torque
limit of the turbo engines has been reached. Figure 5.215 shows the response of the turbo engine
model for a step command.
                                            Time(s)
                Figure 5.215 Turbo Charged Engine Response for a One Liter Engine
    Cold start conditions were not a factor for the simulations since the study assumed "hot start," for all
  simulations, with the engine coolant temperature steady around 95 degrees C.
                                              5-513

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   The turbo response varies with engine speed, i.e. at higher speeds, the turbo response is faster
due to higher exhaust flow rates. It should be noted that the baseline engine maps (Engine 1 and
Engine 12) for the naturally aspirated and the turbo engines were validated with test data.
Maximum torque line on boosted engines is adjustable based on boost pressure (Engine 12
especially could have higher torque potential).

5.4.2.8 Transmissions

   To represent the current market distribution and trends, NHTSA considered AT, MT, DCT,
and CVT transmission technologies in the current assessment.

   As was discussed in Section 5.2, above certain values, additional gearing and ratio spread
provide minimal  additional fuel economy benefits. For this reason, the maximum gear number
considered in the present analysis was limited to 8.

   Based on the current market distribution, trends and benefit limitations of very high gear
numbers, NHTSA, Argonne and Volpe selected the following configurations for use in the Volpe
model:

       •   5-speed automatic (5AU - baseline vehicle)
       •   6-speed automatic (6AU)
       •   8-speed automatic (8AU)
       •   6-speed dual-clutch (6DCT)
       •   8-speed dual-clutch (8DCT)
       •   Continuously variable (CVT)
       •   5-speed manual (5DM)
       •   6-speed manual (6DM)
       •   7-speed manual (TDM)
   Progressive transmission gear ratios have been designed for each transmission type
considering trends in gear span and ratios, as well as expected differences in vehicle performance
and energy consumption based on the transmission technology. On the basis of a literature
review and evaluation of Argonne's APRF chassis dynamometer test data for multiple
conventional vehicles, the following criteria were selected for the design of transmission gear
ratios, final drive ratios, and shift parameters.

       •   The vehicle should shift to top gear above a certain vehicle speed (i.e. 45 mph).
       •   In top gear, the engine should operate at or above a minimum engine speed (i.e.
          1,250 rpm) to prevent engine lugging.
       •   The number of gear shifts for specific transmission on each cycle was defined using
          APRF vehicle test data. For example, for  a 6-speed transmission, on the Urban
          Dynamometer Driving Schedule cycle, the number of shifts should be around 110 to
          120 based on a review of chassis dynamometer test data. Note that this constraint is
          only evaluate after the simulations  and  is  only used to highlight vehicles with
          potential drive quality issues.
       •   Gear span and final drive ratios should  be based on industry trends.
       •   Engine operation will be restricted in the  low-speed/high torque region to prevent
          noise, vibration, and harshness issues and ensure drive quality.
       •   The span of the 8-speed transmissions is higher than that of the 6-speed transmission.
                                             5-514

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
       •  The span of the 8-speed DCT is slightly higher than the span of the 8-speed automatic
          to compensate for the lack of torque multiplication of the torque converter for the
          automatic transmission.
       •  The vehicle should be able to meet or exceed Vehicle Technical Specifications
          (VTSs) related to grade (in first and top gear) and passing performance.
   Dual clutch transmissions with torque converters are being introduced in the market. But,
based on the 2014 EPA Report on light -duty vehicles, a significant majority of the DCT
transmissions in the market today  are without the use of a torque converter device.615 Therefore,
in this study, it is assumed that a torque converter is not used with the DCT.

   Transmission design parameters that substantially affect engine operation - gearing ratios,
ratio spread, and shift control strategy - are all used to optimize the engine operation point, and
thus the effectiveness of these transmission parameters depend in large part on the engine it is
coupled with. Advanced engines incorporate new technologies, such as variable valve timing and
lift, direct injection, and turbocharging and downsizing, which improve overall  fuel consumption
and broaden the area  of high-efficiency operation. With these more advanced engines, the
benefits of increasing the number of transmission gears (or using a continually variable
transmission) diminish as the efficiency remains relatively constant over a wider area of engine
operation. Due to the impact of transmission design, Argonne conducted a review of current
transmissions in the market to select the design parameters for the study.

   Based on publicly available data, the gear spans, transmission gear ratios, and final drive
ratios for several vehicles were reviewed. Table 5.220 lists the minimum and maximum values
for gear ratio span, final drive ratio, and engine speed in top gear at 45 mph (indicator of top gear
ratio). The table also  lists the selected values for the 6-speed transmission. A  similar selection
was made for the 8-speed case, as well.
   Table 5.220 Gear Ratio, Final Drive Information for Sample 6-Speed Automatic Transmission Vehicles

Span
Final Drive
Engine Speed (45 mph)
Minimum Value
5.6
3.2
1,234 RPM
Maximum Value
6.15
4.58
1,604 RPM
Selected Value for Study
6.00
3.74
1,420 RPM
   A gear span of 6 was selected for the 6-speed case, because current trends in transmission
technology reflect increasing gear spans, thus driving selection of a span closer to the maximum
observed value.

   Similarly, span and final drive ratios for the 8-speed AU transmission were chosen,
considering available transmissions in the market today as well as the criteria listed above. It
should be noted that there are very few compact cars currently in the market with 8-speed
transmissions, and most of the available data suggest the use of 8-speed transmissions in the
large sedan (and higher) segments, luxury cars, and sports cars. Therefore, the decision on gear
span and final drive ratio was made so as to meet the criteria listed above.

   Table 5.221 lists the span, final drive ratio, and engine speed at 45 mph for the 6-speed AU,
8-speed AU, and 8-speed DCT transmissions. With a start-stop (BISG) powertrain configuration,
the electric machine provides additional torque during vehicle launch, thus aiding in vehicle
                                             5-515

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
acceleration and performance. Therefore, it is possible to have a lower final drive ratio than for a
conventional powertrain with the same transmission. A very small final drive ratio would result
in increased transmission gear ratios to attain the same performance and grade ability
requirements, and therefore, an inherent trade-off exists between higher transmission gear ratio
and final drive ratio. Finding an optimum trade-off between transmission gear ratio and final
drive ratio for the BISG is beyond the scope of this study. Table 5.221 shows gear span, final
drive and engine speed in top gear at 45 mph for a 6-speed AU, 8-speed AU, and an 8-speed
OCT.
      Table 5.221 Comparison of Gear Span, Final Drive and Engine Speed for Three Transmissions

Span
Final Drive
Engine Speed (45 mph)
6-speed AU
6
3.7
1,420 RPM
8-speed AU
7.5
3.5
1,290 RPM
8-speed DCT
7.7
3.5
1,290 RPM
   With the gear span, final drive ratio, and expected engine speed at 45 mph in top gear all
preselected, the progressive gear ratios were calculated for each transmission type using the
following formula from:
                                       Span
   Where:

   z = total number of gears,

   n = gear number in consideration for design (varies from 1 to z),

   ^2 = progression factor (independent variable — normally between 1 and 1.2),
   I
    z = top gear ratio, and
   I
      = nth gear ratio.
   The independent variable ^ can normally take a value between 1 and 1.2 based on industry
trends. The selection of ^ causes a trade-off between energy consumption and performance. For
this study, the independent variable, for each transmission, was chosen so as to minimize the
energy consumption over a combined HDDS (Urban) and FEWFET (Highway) drive cycle.
Figure 5.216 shows the fuel economy and performance (IVM-60 mph) for different values of the
independent variable for a UDDS cycle.
                                             5-516

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                       40.6
                                            40.5
                               1      1.06    1.07     1.08     1.2

                             I Fuel Eco UDDS (MPG)  • Performance IVM-60 (s)
  Figure 5.216 Fuel Economy and Performance Variations with Choice of Progression Factor for a 6-Speed
                                        Transmission

   As shown, a value of 1.07 provides the maximum fuel economy and was therefore chosen to
decide the gear ratios of the multi-speed transmissions for the study. Figure 5.217 shows the gear
ratios obtained with three different values of (p2 for a 6-speed transmission.
4
3.5
3

% 2.5
ro
o
ro 2
1.5
1
0.5
Gear Ratio Steps across Phi 2
V
\
\
- \v\
\
\

\ \
\ x^
_. \ \
\ \
\ \ ^\
\ \. N.
r r r r r r r


Phi 2: 1.0




_

-
-
- — — ^^~~^==^
r r







1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
Gear Ratio Number

  Figure 5.217 Gear Ratios Obtained with Three Values of Progression Factor for a 6-Speed Transmission
   A similar process was used for the 8-speed transmissions.

   To validate the approach described above for selection of the intermediate gear ratios, the
intermediate gear ratios calculated by the algorithm were compared to actual vehicles for two
vehicles in the compact class. Gear span, final drive ratio, and top gear ratio were inputs to the
equation above. As Figure 5.218 and Figure 5.219 show, with proper selection of the
independent variable ^, the calculated gear ratios are very close to the actual gear ratios.
                                              5-517

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
                                 Mazda 3 Gear Ratio Vs. Least Square (interpolated Ratios
                                         Mazda 3 Ratios
                                         Ratios derived based on LSE method on Phi2
                                             345
                                           Gear Ratio Number
              Figure 5.218 Comparison of Actual Gear Ratios and Gear Ratios Calculated
                                  Golf Gear Ratio Vs. Least Square (interpolated Ratios
                                         -Golf Ratios
                                         "Ratios derived based on LSE method on Phi2
                                             345
                                           Gear Ratio Number
              Figure 5.219 Comparison of Actual Gear Ratios and Gear Ratios Calculated

   A similar validation was performed with the Ford Focus and the Chevy Cruze. Table 5.222
shows the value of ^, which was calculated to minimize the LSE (Least Square Error) between
calculated and actual gear ratios for the vehicles, in comparison to the value of ^ chosen for the
study.

                 Table 5.222 Progression Ratio for Numerous Vehicles with 6-speed AU

lf.t
Ford Focus
1.09
Chevy Cruze
1.04
Mazda 3
1.08
Volkswagen Golf
1.08
Study
1.07
   Table 5.223 summarizes gear and final drive ratios for the different transmissions evaluated in
the study.
                                                 5-518

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                             Table 5.223 Transmission Attributes


5AU

5 DM

6AU

6DM

6DCT

7DM

8AU

80 M

8DCT



Gear
ratio
Total
gear
Gear
ratio
Total
gear
Gear
ratio
Total
gear
Gear
ratio
Total
gear
Gear
ratio
Total
gear
Gear
ratio
Total
gear
Gear
ratio
Total
gear
Gear
ratio
Total
gear
Gear
ratio
Total
gear
Gear
1
3.85
12.74
4.235
14.02
3.9
14.59
4.225
15.80
4.225
15.80
4.48
16.13
4.725
16.54
4.914
17.20
4,914
17.20
2
2.3262
7.70
2.4935
8.27
2.3805
8.90
2.5379
9.49
2.5379
9.49
2.7351
9.85
2.8923
10.12
2.9912
10.47
2.9912
10.47
3
1.5039
4.98
1.5773
5.22
1.5547
5.81
1.6312
6.10
1.6312
6.10
1.7867
6.43
1.8944
6.63
1.9482
6.82
1.9482
6.82
4
1.0403
3.44
1.0654
3.53
1.0865
4.06
1.1218
4.20
1.1218
4.20
1.2488
4.50
1.3276
4.65
1.3577
4.75
1.3577
4.75
5
0.77
2.55
0.77
2.55
0.8124
3.04
0.8255
3.09
0.8255
3.09
0.934
3.36
0.9956
3.48
1.0124
3.54
1.0124
3.54
6
n/a
n/a
n/a
n/a
0.65
2.43
0.65
2.43
0.65
2.43
0.7474
2.69
0.7988
2.80
0.8078
2.83
0,8078
2.83
7
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.64
2.30
0.6853
2.40
0.6897
2.41
0.6897
2.41
8
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.63
2.21
0.63
2.21
0.63
2.21
Final
Drive

3.31
3.31
3.74
3.74
3.74
3.5
3.5
3.5
3.5
   Conventional vehicles were simulated with an automatic transmission, manual transmission,
dual clutch transmission, and continuously variable transmission. Power-split HEV and PHEV
20 AER transmissions have a planetary gear set with 78 ring teeth and 30 sun teeth, similar to the
Toyota Prius. The PHEV 30 and PHEV50 AER have a planetary gear set with 83 ring teeth and
37 sun teeth, similar to the GM Voltec Genl. Fuel cell vehicles use a two-speed manual
transmission to increase the powertrain efficiency as well as allow them to achieve a maximum
vehicle speed of at least 100 mph. BEVs are fixed gear. Table 5.224 gives the characteristics of
all transmission used in the study.
                                            5-519

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                           Table 5.224 Transmission Peak Efficiency
Peak Efficiency (%): Automatic Trans.
Peak Efficiency (%): CVT
Peak Efficiency (%): DCT
Peak Efficiency (%): Manual Trans.
Peak Efficiency (%): Planetary gearset/VoltecGenl
Peak Efficiency (%): Final Drive
97.5
97.5
98
98.5
98
98
   In the current analysis, similar performance data was used for transmissions (i.e., the 1:1 ratio
of the 6 and 8 speed transmissions use the same performance maps). This approach was used to
be able to estimate the effectiveness impact of transmissions with higher gear numbers (i.e.
increased gear spread) and advanced controls (i.e., earlier torque converter lockup).
Benchmarking data collected by EPA and its contractors for a current 6 speed automatic
transmission and 8 speed transmission, show that the transmissions currently in the market do
not have the same efficiencies since they were designed at different timeframes. As a result,
NHTSA has applied a fixed additional effectiveness to represent the benefits of improved
efficiency between existing 6 and 8 speed transmissions. Future simulations runs will include
multiple efficiency options for each transmission to account for changes in transmission design
over time. Additional benchmarking performed by NHTSA and other agencies will also be
leverages when they become available to update the transmission technology, assumptions and
decision tree steps.

5.4.2.9 Torque Converter

   Multiple torque converter performance maps were used for the vehicle simulations depending
on the  engine maximum input torque. An example of data set is provided in Figure 5.220.
                       2.6

                       2.4

                       2.2

                       2

                       1.8

                       1.6

                       1.4

                       1.2

                       1

                       0.8
                                       Torque Ratio and K Factor
	K Factor (rad/s)/sqrHNm]
^—Torque Ratio
                            0.1   0.2   0.3   0.4   0.5   0.6   0.7   0.8
                                          Speed Ratio
                                                                  0.9
                      Figure 5.220  Torque Converter Specification Example
                                              5-520

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.10      Electric Machines

   Electric machine performance data were provided by Oak Ridge National Laboratory. The
performance maps, developed under DOE Vehicle Technologies Office funding, are shown
below for:

       •   micro-HEV, BISG and CISD (Figure 5.221),
       •   HEV and blended PHEV (Figure 5.222),
       •   E-REV PHEV (Figure 5.223) and
       •   BEV and FCHEV (Figure 5.224)
       •   The performance maps were developed assuming normal temperature operating
          conditions. Electric machine inverter losses are included in the maps.
       •   The figures below represent the electric machine peak torque curves. A constant ratio
          was assumed between the continuous and peak torque curves, as follows:
       •   2 for the micro-HEV, BISG, and CISG
       •   2 for the electric machine 1 and 1.5 for the electric machine 2 of the power-split HEV
          and blended PHEV
       •   1 for EREV, BEVs, and fuel cell HEV
   The electric machine specific weight is 1,080 W/kg and its controller 12,000  W/kg. The peak
efficiency is set to 90 percent. This specific weight value was provided by electric machine
experts (DOE, OEMs) and was intended to represent the expected state of the technology by
2020. The value may not, however, represent the most optimistic case, and Argonne is planning
to update the value based on information from DOE and OEM experts that has recently been
received.

   The main focus of BISG hybrid vehicles is to capture regenerative braking energy as well as
provide minimal assist to the engine during high-transient operating modes. Because the electric
machine is linked to  the engine through a belt, its power is usually limited. A value of 7 kW was
assigned to the BISG for the midsize car.

   CISG hybrid vehicles focus on the same areas of improvement as BISG vehicles. However,
owing to its position, the electric machine can be larger; consequently, additional benefits can be
obtained from regenerative braking and assist in a CISG vehicle than in a BISG vehicle. An
electric machine size of 15 kW was selected for the midsize car.
                                            5-521

-------
                             Technology Cost, Effectiveness, and Lead-Time Assessment
                                      Motor Efficiency Map (Torque)
                                                CS5 Mo«« EISci«ney Map
                                                     Ptop*lmg Max Torque Cuiw(H m)
                                                         Max Torque Cun*(N m)
                           •1000
                                     1: 1
                                                        500
                                                                 1000
                                           Speed (fad's)
   Figure 5.221 Electric Machine Map for Micro- and Mild HEV (data source ORNL)
           Motor Efeitncy Map (Tnque,                                         ««« Efc»«y Map (Tow.)
                                                                            IDD"
                                                                                   or Efkiency Map
                                                                                 Propdmj Max T*qu« Cunwflt ml
                                                                                      ax Torque Curv*
-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                                      Maux E*ci»ncy Map (To
-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Different useable state-of-charge ranges during the standard driving cycles under normal
temperature conditions have also been selected depending on the powertrain configuration:

       •   10 to 20 percent SOC range for micro, mild, and full HEVs.
       •   65 percent SOC range for PHEVs
       •   90 percent for BEVs.
   Over time, batteries lose some of their power and energy capacity. To maintain similar
performance at the end of life (EOL) compared with the beginning of life (BOL), an oversize
factor was applied while sizing the batteries for power (HEVs) and energy (PHEV). These
factors represent the percentage of power and energy that will not be provided by the battery at
the EOL compared with the initial power and energy given by the manufacturer. The
performance data used to model  the other components are based on normal temperature
operating conditions. The vehicles are sized with a 20 percent power oversize factor for all
hybrid vehicles and energy oversize factors of 30 percent for PHEVs. BEVs 200 AER are not
oversized.

   The performance data used for the energy storage systems (i.e., Voc, Rint...) represent state-
of-the-art technologies. Since most of the current R&D activities focus on battery life and cost
and considering the time for new materials to be introduced into the market, it is expected that
the battery  performance data will remain fairly constant in the near future.

   Vehicle  test data have shown that, for the drive cycles and test conditions considered, battery
cooling does not draw a significant amount of energy, if any at all, for most of the vehicle
powertrain  architectures. The exception is high energy PHEVs and BEVs, for which an
additional constant power draw is used to account for battery  cooling.  The auxiliary loads in
Autonomie vehicle simulations reflect those impacts.

   The energy storage system block models the battery pack as a charge reservoir and an
equivalent circuit. The equivalent circuit accounts for the circuit parameters of the battery pack
as if it were a perfect open-circuit voltage  source in series with an internal resistance and 2 RC
circuits which represent the polarization time  constants. The amount of charge that the energy
storage system can hold is taken as constant,  and the battery is subject to a minimum voltage
limit.  The amount of charge required to replenish the battery after discharge is affected by
coulombic efficiency. A simple single-node thermal model of the battery is  implemented with
parallel-flow air cooling.

   The voltage is calculated at t=0 as Vout = Voc - Rint * I, with Voc = open-circuit voltage,
Rint = internal resistance (two separate sets of values for charge and discharge), and I = internal
battery current (accounts for coulombic efficiencies).

5.4.2.10.2     Fuel Cell Systems

   The fuel cell system is modeled to represent the hydrogen  consumption as a function of the
produced power as shown in Figure 5.225. The system's peak efficiency is 60 percent, including
the balance of plant, and represents normal temperature operating conditions. The system's
specific power is 659 W/kg.

   The hydrogen storage technology selected  is a high-pressure tank with a  specific weight of
0.04 kg H2/kg, sized to provide a 320-mile range on the FTP  drive cycle.
                                             5-524

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
                       70
                       60
                     0)

                     I 40
                     09
                     O
                     LL
                       30
                       20
                                  20      40      60      80
                                        PCS Net Power, %

                            Figure 5.225 Fuel Cell System Efficiency
                                                     100
5.4.2.11
Light-weighting
   In the NHTSA analysis, light-weighting assumptions are associated with the glider weight. Its
secondary effect (such as downsizing) will be taken into account as part of the vehicle sizing
algorithm. The glider percentage mass reduction values selected for the model are:

       •  0 percent (reference vehicle)
       •  5 percent reduction
       •  7. Spercent reduction
       •  10 percent reduction
       •  15 percent reduction
       •  20 percent reduction
   Only the baseline vehicles and the vehicles with high levels of mass reduction (10, 15 and 20
percent) are sized to meet the vehicle technical  specifications. Vehicles with lower levels of mass
reduction (5 and 7.5 percent) inherit sizing characteristics (i.e. engine power) from their
respective baseline.
5.4.2.12
Rolling Resistance
   The following rolling resistance reduction values were selected for the NHTSA CAFE
analysis:

       •  0 percent (reference vehicle)
       •  10 percent reduction
       •  20 percent reduction
   These values represent a reduction in the coefficient of rolling resistance and were chosen to
bound the possible rolling resistance improvements expected in future vehicles. No sizing is
performed on this dimension.
5.4.2.13
Aerodynamic
                                              5-525

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
   The following aerodynamic reduction values were selected by NHTSA for the CAFE
analysis:

       •  0 percent (reference vehicle)
       •  10 percent reduction
       •  20 percent reduction
   These values represent a reduction in drag coefficient (Cd) and were chosen to bound the
possible rolling resistance improvements expected in future vehicles. No powertrain sizing is
performed on this dimension. The reference values were selected after an analysis of the current
vehicle characteristics and will be updated based on new information.
5.4.2.14      Accessory Loads

   Electrical and mechanical accessory base loads are assumed constant over the drive cycles,
with a value of 240 W for conventional, HEV and blended PHEV powertrains. For EREV
PHEVs and BEVs, a value of 460W is used. Derived from data from Argonne's Advanced
Powertrain Research Facility, these values are used to represent the average accessory load
consumed during the standard urban FTP and EPA's Highway Fuel Economy Test (HWFET)
drive-cycle testing on a dynamometer. Only the base load accessories are assumed during the
simulations,  similar to the dynamometer test procedure.

5.4.2.15      Driver

   The driver model is based on a look-ahead controller for drive cycle simulations. No
anticipation is imposed (0 sec anticipated time) during sizing for acceleration testing, in order to
provide realistic vehicle performances.

5.4.2.16      Electrified Powertrains

   Interest in electric drive vehicle technologies is growing, and the number of electrified vehicle
options available from OEMs is rapidly increasing. This growth represents a shift of focus from
market entry and environmental drivers to mainstream, customer-committed development.
ANL's  assumptions for electrified vehicles are based on the latest assumptions provided by DOE
and OEM experts for the 2020 production timeframe.  ANL is considering additional modeling
based on recent input from DOE and other experts.

   Hybrid vehicles combine at least two energy sources, such as an internal combustion engine
or fuel  cell system with an energy storage system. Electric drive vehicles have the potential to
reduce  energy consumption in several ways, including the following:

       •   Regenerative braking: A regenerative brake is an energy mechanism that reduces the
          vehicle's speed by converting some of its kinetic energy into a storable form of
          energy for future use instead of dissipating it as heat, as with a conventional friction
          brake. Regenerative braking can also reduce brake wear and the resulting fine
          particulate dust.
       •   Engine shutoff under various driving conditions (e.g., vehicle stopped, low power
          demand).
       •   Engine downsizing, which may be possible to accommodate  an average load (not a
          peak load), would reduce the engine and powertrain weight. Higher torque at low
                                            5-526

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
          speed from the electric machine also allows the vehicle to achieve the same
          performance as conventional vehicles with a lower vehicle specific power (W/kg).
       •  Optimal component operating conditions: For example, the engine can be operated
          close to its best efficiency line.
       •  Accessory electrification allows parasitic loads to run on as-needed basis.
       •  The energy storage systems of PHEVs and battery electric vehicles can also be
          recharged, further improving fuel displacement.
   However, vehicle  electrification also have disadvantages that could affect energy
consumption, including increased vehicle weight due to additional components.

   Two major types of hybrids have been considered for transportation applications: electrical
and hydraulic. Since Hydraulic Hybrid Vehicles have been studied almost exclusively for
medium- and heavy-duty applications, only HEVs have been considered in the present study.

   HEVs combine electric and mechanical power devices. The main components of HEVs that
differentiate them from conventional vehicles are the  electric machine (motor and generator),
energy storage (e.g., battery or ultra-capacitors), and power electronics. The electric machine
absorbs braking energy, stores it in the energy storage system, and uses it to meet acceleration
and peak power demands.

5.4.2.16.1     Electrified Powertrain Configurations

   The various HEV  powertrain configurations can be classified on the basis of their
hybridization degree, as shown in Figure 5.226. The hybridization degree is defined as the
percentage of total power that can be delivered electrically. The higher the hybridization degree,
the greater is the ability to propel the vehicle using electrical energy.
                                             5-527

-------
                                Technology Cost, Effectiveness, and Lead-Time Assessment
                         Full Power Assist
                         Electric Only Mode
                         Full Regenerative
                            Braking
                        Medium Power Assist,
                         Operating Strategy
                           Optimization
                          Limited Motor
                             Assist
                         Limited/Medium
                           Regenerative
                            Braking
                                                                             FJ  I

                                                                        h I Micro HEV
                                                                 I  j
                       Figure 5.226 Electric Drive Configuration Capabilities

   A number of different powertrain architectures have been considered and introduced in the
market for different applications. These architectures are usually classified into three categories:
series, parallel, and power split. The following sections describe some of the possible powertrain
configurations for each architecture.
5.4.2.16.2
Parallel Hybrid Vehicle
   In a parallel configuration, the vehicle can be directly propelled by either electrical or
mechanical power. Direct connection between the power sources and the wheels leads to lower
powertrain losses compared to the pure series configuration. However, since all of the
components' speeds are linked to the vehicle's speed, the engine cannot routinely be operated
close to its best efficiency curve.

   Several subcategories exist within the parallel configuration:

       •  MHEV: A small electric machine is used to turn the engine off when the vehicle is
          stopped.

       •  Starter-alternator: This configuration is based on a small electric machine (usually 5
          to 15 kW) located between the engine and the transmission. Because of the low
          electric-machine power, this configuration is mostly focused on  reducing
          consumption by eliminating idling. While some energy can be recuperated through
          regenerative braking, most of the negative electric-machine torque available is
          usually used to absorb the engine's negative torque.  Since the electric machine speed
          is linked to the engine, the vehicle cannot operate in electric mode other than for
          extremely low speeds  (e.g., creep). In addition, the electric machine  is used to smooth
          the engine torque by providing power during high transient events to reduce
                                               5-528

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
          emissions. The electric machine can be connected to the engine either through a belt
          or directly on the crankshaft.

       •  Pre-transmission: This configuration has an electric machine in between the engine
          and the transmission. The electric machine power ranges from 20 to 50kW for light
          duty applications, which allows the driver to propel the vehicle in electric-only mode
          as well as recover energy through regenerative braking. The pre-transmission
          configuration can take advantage of different gear ratios that allow the electric
          machine to operate at higher efficiency and provide high torque for a longer operating
          range. This configuration allows operation in electric mode during low and medium
          power demands, in addition to the ICE on/off operation. The main challenge for these
          configurations is being able to maintain a good drive  quality because of the engine
          on/off feature and the high  component inertia during shifting events.

       •  Post-transmission: This configuration shares most of the same capabilities as the pre-
          transmission. The main difference is the location of the electric machine, which in
          this case is after the transmission.  The post-transmission configuration has the
          advantage of maximizing the regenerative energy path by avoiding transmission
          losses, but the electric machine torque must be higher because it cannot take
          advantage of the transmission torque multiplication.
5.4.2.16.3     Power Split Hybrid Vehicle

   As shown in Figure 5.227, power split hybrids combine the best aspects of both series and
parallel hybrids to create an extremely efficient system.  The most common configuration, called
an input split, is composed of a power split device (planetary gear transmission), two electric
machines and an engine. Within this architecture, all these elements can operate differently.
Indeed, the engine is not always on and the electricity from the generator may go directly to the
wheels to help propel the vehicle, or go through an inverter to be stored in the battery. The
operational phases for an input split configuration are the following:

       •  During vehicle launch, when  driving, or when the state of charge of the battery is
          high enough, the ICE is not as efficient as electric drive, so the ICE is turned off and
          the electric machine alone propels the vehicle.
       •  During normal  operation, the ICE output power is split, with part going to drive the
          vehicle and part used to generate electricity.  The electricity goes either to the electric
          machine, which assists in propelling the vehicle, or to charge the energy storage
          system. The generator also acts as a starter for the engine.
       •  During full-throttle acceleration, the ICE and electric machine both power the
          vehicle, with the energy storage device (e.g., battery) providing extra energy.
       •  During deceleration or braking, the electric machine acts as a generator, transforming
          the kinetic energy of the wheels into electricity to charge the energy storage system.
                                              5-529

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
                                Generator
Converter     Battery
                        Figure 5.227 Power Split Hybrid Electric Vehicle

   Several variations of the power split have been implemented, including single-mode and
multi-mode power splits. Examples of single-mode power split hybrids include the Toyota Prius
and Ford Fusion Hybrid.  An example of a multi-mode power split hybrid is the General Motors
Chevrolet Tahoe. It should be noted that there are possible tradeoffs between complexity and
energy consumption benefits for multi-mode systems.616

5.4.2.16.3.1   Voltec Genl Plug-in Hybrid Vehicle

   PHEVs differ from HEVs in their ability to recharge the energy storage system through the
electric grid. PHEVs energy storage systems have usually a higher total energy compared to
HEVs and they also use a larger portion of it (e.g., when most HEVs use 10 to 30 percent of their
total battery energy, PHEVs use 60 percent or more of their total energy). Since the vehicle is
designed to have a high capacity energy storage, electrochemical batteries are usually used for
this application. All the HEV configurations described above can be used as PHEVs. In most
cases, because of the desire to propel the vehicle using electrical energy from the energy storage
system, the electric machine power is greater for a PHEV compared to an HEV.

   ANL used the Gen 1 VOLTEC configuration from General Motors in its simulation to
represent a PHEV. Argonne is currently working on developing new vehicles models and sizing
algorithms for the three new powertrain configurations recently introduced by  GM so that those
options can be considered in the next round of simulations in Autonomie.
   The VOLTEC GEN1 configuration from General Motors allows different operating modes
(e.g. series and parallel, parallel and power split). The VOLTEC GEN1 powertrain architecture,
also called the EREV (Extended Range Electric Vehicle), provides four modes of operating,
including two that are unique and maximize the powertrain efficiency and performance. The
electric transaxle has been specially designed to enable patented operating modes, both to
improve the vehicle's electric driving range when operating as a BEV and to reduce energy
consumption when extending the range by operating with an ICE. The EREV powertrain
introduces a unique two-electric machine electric-vehicle driving mode that allows both the
driving electric machine and the generator to provide tractive effort while simultaneously
reducing electric machine speeds and the total associated electric machine losses. For HEV
                                            5-530

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
operation, the EREV transaxle uses the same hardware that enables one-electric machine and
two-electric machine operation to provide both the completely decoupled action of a pure series
hybrid and a more efficient flow of power with decoupled action for driving under light load and
at high vehicle speed.

   It is important to note that many different variations exist within each configuration (i.e.,
power-split configurations can be single-mode, two-mode, three-mode, etc.) and between
configurations (i.e., several configurations are considered to be a mix of series, parallel and/or
power-split). Overall, several hundred configurations are possible for electric-drive vehicles. It is
also not uncommon for a specific OEM to use multiple powertrain configurations across its
electrified vehicle line up. Recent presentations from General Motors highlighted the fact that,
while sharing multiple components, the powertrains from the upcoming Gen2 Volt, Cadillac
CIS and Malibu were all different.

   In more detail, the Voltec Genl system has four different operating modes, as shown in
Figure 5.228:

   During EV operation:

      •  One-electric machine EV: The single-speed EV drive power-flow, which provides
          more tractive effort at lower driving speeds
      •  Two-electric machine EV:  The output power split EV drive power flow, which has
          greater efficiency than one-electric machine EV at higher speeds and lower loads
   During extended-range (ER) operation:

      •  One-electric machine ER (series): The series ER power flow, which provides more
          tractive effort at lower driving speeds
      •  Combined two-electric machine ER (split): The output power split ER power-flow,
          which has greater efficiency than series at higher speeds and lighter loads
   A vehicle-level control strategy was developed on the basis of vehicle test data to properly
select each of the operating modes. The logic developed for the power split mode is similar to the
one for the input split configuration discussed previously.

   For the two-level EV mode, an algorithm has been developed to minimize the losses of both
electric machines at every sample time on the basis of each component's efficiency map. For the
series mode, the combination of the engine and electric machine losses is also minimized at
every sample time. It is important to note that the engine is not operated at its best efficiency
point, but rather along its best efficiency line for drive quality and efficiency reasons.
                                             5-531

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
 B_ECTRICCRIVING
 Low3»«J
B_ECTRICCRIVING
HghSpMd
(2-M«Of)

 EXTENDED RANGE DPI VING
 Low^eed
 (IMotorSenes)
BCTBMCBD RANGE DRIVI NG
Hgh Speed
(2-Mct
-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
  Fuel Cell
  System
                                                                                 Chassis
                                    DC/DC
Accessory
                      Figure 5.229 Series Fuel Cell Hybrid Electric Vehicle

   Several variations of the series configuration have been considered. One of the important
considerations in the design of a series HEV is related to the use of a single gear ratio versus a
two-speed transmission. Using a single gear ratio usually leads to low maximum vehicle speed
and poor performance at high speed due to the low electric machine torque in that operating
regime. When applications require better performance at high speeds, a two-speed transmission
is considered.

5.4.2.16.5     Powertrain Electrification Selection

   The selection of hybridization degree and powertrain configuration is complex, since
numerous options exist. On the basis of current production vehicles  as well as anticipated near-
future trends, the following powertrain configurations were selected for the modeling analysis to
match Volpe's requests:

       •   12-V micro-hybrid electric vehicle (micro-HEV/start-stop system, no regen braking).
       •   Belt-integrated starter generator
       •   Crank-integrated starter generator
       •   Full hybrid electric vehicle, single-mode power split configuration, fixed ratio
       •   Full hybrid electric vehicle, Pre-Transmission configuration, 6-speed DCT.
       •   PFIEV, Voltec extended-range electric vehicle  (EREV) configuration with 30 AER on
          the FTP cycle
       •   PHEV, Voltec EREV configuration, with 50 AER on the FTP drive cycle
       •   Battery electric vehicle, with 200 AER on the FTP drive  cycle
       •   Fuel cell FIEV, series configuration, with 320-mile range on the FTP drive cycle
   Note that the AER values are based on unadjusted electrical consumptions on the UDDS
driving cycle. Recent announcements by automakers indicate 200 plus mile label ranges are
likely.  If this is the case, UDDS driving cycle AERs will be closer to 250 miles and if so ANL
will update its assumptions for future simulation modeling.
                                             5-533

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.17
Drive Cycles and Vehicle Simulation Conditions
   Simulated test procedures followed the current recommendations of the EPA, with the two-
cycle test based on the FTP and HFET drive cycles. Combined values are calculated on the basis
of a 55 percent city and 45 percent highway cycle using the standard test procedure.

   Autonomie includes some temperature models for some powertrains and component
technologies, but considering the wider range of options to be considered as part of the study, all
the component performance data and controllers are assumed to be operating under warm
conditions. As a result, the additional energy consumption due to the FTP cold start has been
calculated in post-processing by applying a fuel consumption penalties depending on the
assumed warmup strategy. A constant value of 15 percent across all technology options has been
applied based on a combination of Argonne APRF test data and analysis of the latest EPA
vehicle certifications data as shown in Figure 5.230. No cold start penalty was applied for BEVs.
  Figure 5.230 Cold Start Penalty between Bag 1 and 3 on the FTP Cycle Based on 2016 EPA Certification
                                         Data
5.4.2.18
Vehicle Sizing Process
   To compare different vehicle technology-configuration-powertrain combinations, all vehicles
to be studied were sized to meet the same requirements:

       •  Initial vehicle movement to 60 mph <= 9 sec ±0.1 sec
       •  Maximum grade (gradeability) of 6 percent at 65 mph at Gross Vehicle Weight
          (GVW)
       •  Maximum vehicle speed >100 mph
   These requirements are a good representation of the current American automotive market and
of American drivers' expectations.  The relationship between curb weight and GVW for current
technology-configuration-powertrain combinations was modeled and forms the basis for
estimating the GVWs of future vehicle scenarios. The following equation has been used to
estimate the GVW of future technologies:
                                            5-534

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   GVW (kg) = 1.25 x vehicle test weight (kg) + 193

   To compare different vehicle technology-configuration-powertrain combinations, all selected
vehicles to be sized are designed to meet the same requirements. Note that not all vehicles are
sized but the baseline vehicle (MRO, AEROO, ROLLO) and higher mass reduction level vehicles
(MRS, 4, 5 with AEROO, ROLLO).

   Improperly sizing the components will lead to differences in energy  consumption and will
influence the effectiveness results. On this basis, we have developed several automated sizing
algorithms to provide a fair comparison between technologies. Algorithms have been defined
depending on the powertrain (e.g., conventional, power split, series, electric) and the application
(e.g., HEV, PHEV).

   All algorithms are based on the same concept: the vehicle is built from the bottom up,
meaning each component assumption (e.g.,  specific power, efficiency) is taken into account to
define the entire set of vehicle attributes (e.g., weight). This process is always iterative in the
sense that the main component characteristics (e.g., maximum power, vehicle weight) are
modified until all vehicle technical specifications are met. The transmission gear  span or ratios
are currently not modified to be optimized with specific engine technologies as this might also
lead  to overestimating the effectiveness impact of technologies. On average, the algorithm takes
between five and 10 iterations to converge.  Figure 5.231  to Figure 5.236 shows the iterative
process for each powertrain.

   A conventional vehicle is mainly defined by its internal combustion  engine; its ability to
realize a cycle or acceleration performance  is directly linked to its power density. Therefore, the
sizing algorithm focuses on calculating the  mechanical power needed to meet the requirements.
Figure 5.231 illustrates the steps in the sizing process. To begin the sizing process, a default
vehicle is  created. A simulation is then performed to determine the engine peak power  and
vehicle mass:

   First, the desired power is estimated to meet the grade-ability and acceleration performance
requirements, and engine power is updated  with the maximum value.

   Second, the sizing enters in an acceleration loop to check the performance run initial vehicle
movement (IVM) up to 60  mph, and the IVM to 60 mph  is recorded. The definition of IVM is
that the vehicle must move 1 ft (1/3 m) before the clock starts to record the performance time.
This metric provides a more consistent result and removes phenomena that are difficult to model
at initial acceleration—such as tire and clutch slip—from consideration.

   Finally, the vehicle is run on acceleration performance for passing with its updated
parameters. At the end, the time to reach the target (i.e., 0-60 mph and  50-80 mph) are
compared with the simulated data. This is the main condition to exit the routine.
                                             5-535

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
              Initialize Variables
                 J L
           Use grade, perfo estimation to
            initialize power of engine
               Compute mass
           Run acceleration simulation
                (0-60 mph)
               Engine power=
          max(grade power, accele power)
             Update vehicle mass
Passing acceleration performance loop
          (50-80mph)
                                            Run passing acceleration
                                             simulation (50-80 mph)
     Acceleration performance loop
             (0-60mph)
                                               Engine power =
                                          maxfgrade power, accele power,
                                               passing power)
                                              Update vehicle mass
                                          Compute Values using Equations

                                          Run Simulation

                                          Tune Variable
                       Figure 5.231 Conventional Powertrain Sizing Algorithm

   For the hybrid electric vehicles, engine power is sized to meet 70 percent of peak power
required to meet the VTS. The battery power is mainly determined to capture all the regenerative
energy from the urban dynamometer driving schedule. The electric machine power is sized to
meet the grade-ability and performance requirements. Figure 5.232 shows the iterative process
used to calculate data for a single power split FtEV.

   The following procedure is used:

       •   Battery power is sized to recuperate 100 percent energy through regenerative braking
           on HDDS.
       •   Electric machine (EMI) power is sized to recuperate 100 percent energy through
           regenerative braking on HDDS and to meet the acceleration performance
           requirement.
       •   Electric machine (EM2) power is sized as following:
       •   EM2 peak power is sized to start engine at the top of vehicle speed on HDDS
       •   EM2 peak power is sized to control engine at the zero of vehicle speed for
           acceleration performance
       •   EM2 continuous power is sized to control engine at maximum grade (i.e., engine
           power fraction going through electro-mechanical power path)
                                                5-536

-------
                                  Technology Cost, Effectiveness, and Lead-Time Assessment
             :
   initialize Variables
:
               Use grade, perfo, regen
             estimation to initialize power of
             engine, battery, electric motors
            	(EM1.2)
                 Compute mass

Run acceleration simulation
    (0-60 mph}
                  EMlpower =
             max(regen power, accele power)
                             *
                Update vehicle mass
            up
            Update power of engine, battery,
              EM 2 for gracfe, perfo, regen
                Update vehicle mass
                                   Passing acceleration performance loop
                                              (50-80mph)
                                              Run passing acceleration
                                              simulation (SO-SO mph)
       Acceleration performance loop
               (0-60mph)
                                                 EMI power =
                                            maxfregen power, accele power)
                                               Update vehicle mass
                                            Update power of engine, battery,
                                             EM2 for ^ade, perfo, regen
                                               Update vehicfe mass
                                                                    Compute Values using Equations

                                                                    Run Simulation

                                                                    Tune Variable
                    Figure 5.232 Split Hybrid Electric Powertrain Sizing Algorithm

   The main algorithm for the single power split powertrain is as follows, and the iterative
process is shown in Figure 5.233:

        •   Battery energy is sized to meet the all-electric range (AER) requirements on UDDS
           based on unadjusted values. Using the full history of the range attained by the vehicle
           from each sizing run, the desired range, and the current battery energy, a new
           estimate was made for the desired battery energy.
        •   Battery and EMI powers are sized to be able to follow the UDDS cycle in electric-
           only mode (this control is only used for the sizing; a blended approach is used to
           evaluate consumptions) or to meet the acceleration performance requirements.
        •   Vehicle weight is a function of the engine peak power, electric machines peak power,
           and battery  energy.
        •   Electric machine (EM2) power is sized the  same way as for a single power split HEV.
                                                   5-537

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
:o nitutoe powers and capacey
                                                                                Compute Values using Equation

                                                                                Run Simulation

                                                                                Tune Variable

                                                                                Update Value*using Simulation

       Update pow Of engine, battery.
        EM2 for grade, per'o, regen
         Update vehide mass
                     /    rj ^  S5ssrn\
                     "    I snvJatoniSO-BO mph)
        Acceleration
      performance loop
         fO-60mph)
                            to pa
                          \
Y-



1



V,















» ^

-






                                        Total Iterated* IK
                                       weaweteiatwnfi prams
                                        uteleratton OK wo/
                       Passing acceleration
                        performance loop
                          (50-80mph)
                                          Loop for following the
                                           UDD5 in EV mode
                Figure 5.233 Split Plug-in Hybrid Electric Powertrain Sizing Algorithm

   The main algorithm for the series-split powertrain is as follows, and the iterative process is
shown in Figure 5.234:
           Battery energy is sized to meet AER on HDDS based on unadjusted values.
           Battery and EMI powers are sized to be able to follow the aggressive US06 drive
           cycle (duty cycle with aggressive highway driving) in electric-only mode or to meet
           the acceleration performance requirements.
           Vehicle weight is a function of the engine peak power, electric machines peak power,
           and battery energy.
           Electric machine (EM2) power is sized to endure the engine peak power as a
           generator and kick on the engine at top speed on the HDDS.
                                                 5-538

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
                                                                               Run Simulation

                                                                               Tune Variable

                                                                               update Values using Simulation Result
                                                                              Yes  Are Hceteraton 8, passing
                                                                                  acceleration OK wo/
                                                                                 tunngof EMI power?
Use grade, perfo. range estimation
 to Irtttialile powers and capacity
                                                                  Passing acceleration
                                                                  performance loop
                                                                    (50-80mph)
     Loop for following the
       US06 in EV mode
                            EV range loop
                                                Acceleration
                                               performance loop
                                                 (0-60mph)

                 Figure 5.234 Series-Split Hybrid Electric Powertrain Sizing Algorithm

   The main algorithm for the single gear BEV powertrain is as follows, and the iterative process
is shown in Figure 5.235:

       •   Battery energy is sized to meet AER on UDDS based on unadjusted values.
       •   Battery and EMI powers are sized to be able to follow the aggressive US06 drive
           cycle (duty cycle with aggressive highway driving) or to meet the acceleration
           performance requirements.
       •   Vehicle weight is a function of the electric machine peak power and battery energy.
   To be able to maintain the same performance at the end of life as at the beginning of life, an
oversize factor is applied while sizing the batteries for both energy (these oversizing factors
influence the weight only).
                                                 5-539

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
                                                Acceleration
                                              performance loop
                                                 (0-60mph)
                    Figure 5.235  Battery Electric Powertrain Sizing Algorithm
                                                                            ^^^^^^ Yes  Are ace ete rations passing
                                                                            Stop   *	   acceleration OK wo/
                                                                                      tuning of EM power?
                                                                                        Compute Values using Equations

                                                                                        Run Simulation

                                                                                        Tune Variable

                                                                                        Update Values using Simulation Results
              _^X

Loop for following the
      US06
                                                 Acceleration
                                               performance loop
                                                  (0-60mph)
            Figure 5.236  Fuel Cell Series Hybrid Electric Powertrain Sizing Algorithm
                                                   5-540

-------
                               Technology Cost, Effectiveness, and Lead-Time Assessment
   Since each powertrain and application is different, the rules are specific:

       •  For HEVs, the electric-machine and battery powers are determined in order to capture
          all of the regenerative energy from an FTP cycle. The engine and the generator are
          then sized to meet the gradeability and performance (initial vehicle movement to 60
          mph) requirements.
       •  For PHEV30 and 50s, the main electric-machine and battery powers are sized to be
          able to follow the aggressive US06 drive cycle (duty cycle with aggressive highway
          driving) in electric-only mode. The battery's usable energy is defined to follow the
          FTP drive cycle for 50 miles, depending on the requirements. The genset (engine +
          generator) or the fuel cell systems are sized to meet the gradeability requirements.
       •  For BEVs, the electric machine and energy storage systems  are sized to meet all of
          the vehicle technical specifications.
   The micro-HEV, BISG, and CISG have sizing results very similar to their conventional
counterparts as they all use the same sizing rule except for the electric machine and energy
storage systems.

   Once the vehicles were sized to meet the same vehicle technical specifications,  they were
simulated following the appropriate standard driving cycles. It is important to properly store
individual results as structured data because they will be reused to support database generation
(see Section 11).

5.4.2.19      Autonomie Outputs

   Once a simulation is complete, the results are  stored in a folder which contains the results for
one combination and characterizes one branch/path of the tree. Figure 5.237  shows the folder
organization for each individual simulation. Folders can contain up to five directories, depending
on the vehicle technology and the type of run performed. Results are divided into directories
representing the cycle or procedure simulated. For example, the combined procedure for
conventional vehicles has two parts separating the FTP and HFET run,  and the PFLEV procedure
has four parts separating the FTP and HFET runs as well as the charge-sustaining and charge-
depleting modes. The last directory is the sizing structure (performance test).
                                             5-541

-------
                                                         Technology Cost, Effectiveness,  and  Lead-Time Assessment
     V   « VOLPE_2013_ result! * 02 - Simulation results > bacth2 - 08212013 t
Orgar
           A Open
 ;- Favorites
 • Desktop
  4 Downloads
 ji^ Recent Places

.., Libraries
 3 Documents
 J. Music
 H «*»<*
 B Video,

* Computer
 ^ vms_fyl3_gpra (W,
 ft OS (C:)
 V ES_Shared(H:)
 ^ Cluster (\\VMS-01J
 tj* VMS Share (P:)
 ^? amoawad (\\wa1es1
 V VMS_FY11_VOLPE_
 ^ Cluster (\\vms-rf!:)
          |   Midsize_CONV.DI20L.5AU_MRO.AEROO_ROLLO_&aidl3_141116    10/20/201311:32 ...  File folder
., Midsize_COIMV_DI20L_5AU_MRO,AEROO_ROLLl_OSl 413,141040     9/7/2013 2M AM
 , Midsize_COIMV_DI20L_5AU_MRO_AEROO_ROLL2_081413_141036     8/22/2013 3;21 PM
t. Midsize_COIW_DT20L_5AU^M(W.AEROO.ROLG.OSU13.141041     3/22/2013 3:21 PM
;. Midsize_COIW.DI20L_5AU_Mra}.AER01_ROLLO_Oai413_141042     9/7/2013 2:44 AM
  Midsae_COW,DI20L_5AU^Mra_AER01_ROLLl_OS1413_141044     8/25/2013 7;42 PM
_;, Midsiie_CONV_!JI20L_5AU_MK)_AER01_ROLL2_081413_141044     8/22/2013 3:22 PM
                                                                      File folder
                                                                      Filefolder
                                                                      Filefolder
                                                                      Filefolder
                                                                      Fitefolder
                  Organize f
                                Bui
                    i" Favorites
                    • Desktop
                    fL Downloads
                    '-   Recent Places

                     Librarie:-
                    _^' Documents
                    j,  Music
                    h.=! Pictures
                    B Videos

                  '*• Computer
                    ^ vms_fy!3_gpra
                          V1 ES.Shared (H:)
                          4* Cluster O.WMS-
                          V VMS Share (P:)
                          ^ amoawad (\\WE
                                 bisg_DI20LJAU^MRO_ROLLO_AEROO_au
                                 cisg_DI20L_5AU_MRO_ROLLO_AEROO_au
                                 conv_DI20L_5AU_MRO_ROLLO_AEROO_au
                                 micro_DI20L_5AU_MRO_ROLLO_AEROO_au
Date modified
9/7/2013 2:43 AM
9/7/2013 2:43 AM
9/7/2013 2:43 AM
9/7/2013 2:43 AM
Type
Filefolder
Filefolder
Filefolder
Filefolder
                             1 Cluster O
                               .  unsized
                              HI bisg_D12<
                              H cisg_DI2(
                              ,	, conv_Dli
                                 micro.Dl
                              Q'- Mids
                              Q " Midsize,
                              g run_bisg.
                              ,	, run_cisg.
                              i	j run_com
                              ,	j run_mici
                              g sizing_bi
                              ,	, sizing_ci;
                              __, sizing_cc
                              ,	, sizing_m
                                  20 items     Offline status:
                                           Offline availability:
                                                                              « Mids.,.  ^  bisg_...
                                                                                                                                :~_C>U'C'i_5A.i'_''
                                                               Organize ^      ,^| Open      Burn
                                                                                                          New folder
                                                        -    a    ft
                                                                              Name
                                                                                                                               Date modified
                                                                                                                                                      Type
                                                                           Combined Procedure (FTP*HWFET)_Partl
                                                                       Ji  Combined Procedure (FTP<-HWFET)_Part2   9/7/2013 2:43 AM
                                                                           sizing
                                                                                                                                                      Filefolder
                                                                                                                                                      File folder
                                                                    Combined  Prc
                                                                    Filefolder
                    Mid;iie_CONV_DOOL_5AU_MR3_AER02_ROLU_08U13_141241
 ti
Midsize_CQNV_DI20L_5AU_MRQ_AEROO_..
Filefolde.
                         odified: 10,^0/2013 11:32 PM
                        ailBbilrtji Not available
                                                                Offline status: Onlins
                                                                                                                         ^ Combin.,.
                                                                                               Organize ^
                                                                                                                 Burn
                                                                                                                           New folder
                                                                                                                                                   «s  -     a
                                                                                                 ,   F,
                                                                                                IS Lib
                                                                                                  HD
                                                                                                  -h N  '
                      Name                                             Date moc
                       _ O0.a_run                                      8/19/2013
                      *  data, mat                                     8/19/2013
                       	, info.txt                                       3,'19/2013
                       l*li par_bisg_midsize_au_2wd_VOLPE.mdl        8/19/2013
                       _, simulation.a_result                           8/19/2013
                                                                                                           5 items      Offline status:  Online
                                                                                                                    Offline availability:  Not available
                                            Figure 5.237  Organization of Simulation Results
                                                                                    5-542

-------
                              Technology Cost, Effectiveness, and Lead-Time Assessment
5.4.2.20      Individual Vehicle Simulation Quality Check

   Once the individual simulations are completed, at the results are analyzed both a high level
(i.e., vehicle energy consumption) and a low level (i.e., time-based engine power) through
Autonomie graphical user interface. An algorithm is also used to automatically flag any potential
issues within a simulation (i.e., too many shifting events on a specific cycle).
   An exhaustive list of parameters are extracted and checked for each vehicle simulation,
including:

       •   Trace
       •   Vehicle Weight
       •   Engine Percentage ON
       •   Engine Number of Starts
       •   Engine/Fuel Cell Average Efficiency
       •   Engine/Fuel Cell Power
       •   Engine Speed
       •   Electric Machine Average Efficiency
       •   Electric Machine Power
       •   Electric Machine Speed
       •   Electric Machine Max Current
       •   Number of Shifts
       •   Time Fraction in Top Gear
       •   Battery SOC
       •   HEV Delta SOC
       •   Percentage Regeneration Recovered
       •   Electric Consumption
       •   Fuel Economy ratios
   Distribution plots are generated as part of the report for visual perspectives (Figure 5.238).
                                            5-543

-------
                   Technology Cost, Effectiveness, and Lead-Time Assessment
30
               Distribution of Fuel Consumption for Eng: Eng13 OHC
                Incremental percentage compared with: Eng12ftOHC
                          Standard Deviation O.9:
                             Number of occurences - all tech. combined
                             Average value = 3.2
                             Baseline vehicle (ANL) = 4.1
                             VOLPE result = 3.6
                            - Density of data
20
10
  0.6
              1.6     2     2.5    3     3.6    4    4.6
                  Fuel Consumption Improvement (INC), %
                                                             6.5
           Figure 5.238  Example of QA/QC Distribution Plot
                                    5-544

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
References
1 U.S. EPA, October 2014, Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy
Trends: 1975 Through 2014, Report EPA-420-R -14-023a.
2 Stanton, D.W. "Light Duty Efficient, Clean Combustion." Final Report by Cummins, Inc., to the U.S. Department
of Energy, Report No. DE-FC26-07NT43279, June 3, 2011. Mt£^/!BS£SMijffi/iC!!^^
3 Yu, C., Park, K., Han, S., and Kim, W., "Development of Theta II 2.4L GDI Engine for High Power & Low
Emission," SAE Technical Paper 2009-01-1486, 2009, doi: 10.4271/2009-01-1486.
4 Saeki, T., Tsuchiya, T, Iwahashi, K., Abe, S. "Development of V6 3.5-Liter 2GR-FSE Engine." Toyota Technical
Review, Volume 55, No. 1, pp 94-99, November 2006. Link?
5 Ikoma, T., Abe, S., Sonoda, Y., Suzuki, H. et al., "Development of V-6 3.5-liter Engine Adopting New Direct
Injection System," SAE Technical Paper 2006-01-1259, 2006, doi: 10.4271/2006-01-1259.
6 Yamaguchi, J. "Lexus Gives V6 Dual Injection." SAE Automotive Engineering International, January 2006, pp 17-
20.
7 U.S. EPA. "Annual Certification Test Results & Data." Light Duty Vehicle Data - Certified Vehicle Test Result
Report Data, MY2010-2016. https://www3.epa.gov/otaq/crttst.htm
8 Ford Motor Company. 2016.  "More Torque and Better Boost: 2017 Ford F-150 to Debut with All-New 3.5-Liter
Ecoboost Engine and 10-Speed Transmission."
https://media.ford.com/content/fordmedia/fna/us/en/news/2016/05/03/2017-ford-fl50-more-torque-better-boost.pdf,
last accessed July 5, 2016.
9 Ford Motor Company. 2015.  "Ford Marks Production Milestone As 5-Millionth EcoBoost-Equipped Vehicle Rolls
Off Assembly Line." https://media.ford.com/content/fordmedia/fna/us/en/news/2015/03/17/ford-marks-production-
milestone-as-5-millionth-ecoboost-equipped.pdf, last accessed July 5, 2016.
10 Volkswagen. 2015. http://www.volkswagen.co.uk/technology/petrol/active-cylinder-technology-act, last accessed
July 5, 2016.
11 Eichler, F., Demmelbauer-Ebner, W., Theobald, J., Stiebels, B., Hoffmeyer, H., Kreft, M.  "The New EA211 TSI®
evo from Volkswagen." 37. Internationales Wiener Motorensymposium 2016.
12 Schamel, A., Scheidt, M., Weber, C. Faust, H. Is Cylinder Deactivation a Viable Option for a Downsized 3-
Cylinder Engine? Vienna Motor Symposium, 2015.
13 Wilcutts, M., Switkes, J., Shost, M. and Tripathi, A., "Design and Benefits of Dynamic Skip Fire Strategies for
Cylinder Deactivated Engines," SAE Int. J. Engines 6(1):2013, doi: 10.4271/2013-01-0359.
14 Ernst, R., Fridfeldt, R., Lamb, S., Lloyd-Thomas, D., Phlips, P., Russell, R., Zenner, T. "The New  3 Cylinder l.OL
Gasoline Direct Injection Turbo Engine from Ford. 20th Aachen Colloquium - Automobile and Engine Technology,
2011.
15 Wada, Y., Nakano, K., Mochizuki, K., and Hata, R., "Development of a New 1.5L 14 Turbocharged Gasoline
Direct Injection Engine," SAE Technical Paper 2016-01-1020, 2016, doi: 10.4271/2016-01-1020.
16 Nakano, K., Wada, Y., Jono, M., Narihiro, S.  "New In-Line 4-Cylinder Gasoline Direct Injection Turbocharged
Downsizing Engine." Honda R&D Technical Review, April 2016, pp 139-146.
17 National Instruments. Subsystems Required to Control Low Temperature Combustion Engines. Jun 11, 2014.
http://www.ni.com/white-paper/13516/en/pdf, last accessed August 25, 2015.
18 Takahashi, D., Nakata, K., Yoshihara, Y.,  Ohta, Y. et al., "Combustion Development to Achieve Engine Thermal
Efficiency of 40% for Hybrid Vehicles," SAE Technical Paper 2015-01-1254, 2015, doi: 10.4271/2015-01-1254.
19 Yamada, T., Adachi,  S., Nakata, K., Kurauchi, T. et al. "Economy with Superior Thermal Efficient Combustion
(ESTEC)," SAE Technical Paper 2014-01-1192 doi: 10.4271/2014-01-1192.
20 Lee, S., Schenk, C., and McDonald, J., "Air Flow Optimization and Calibration in High-Compression-Ratio
Naturally Aspirated SI Engines with Cooled-EGR," SAE Technical Paper 2016-01-0565, 2016, doi:  10.4271/2016-
01-0565.
21 Ellies, B., Schenk, C., and Dekraker, P., "Benchmarking and Hardware-in-the-Loop Operation of a 2014 M 2.0L
13:1 Compression Ratio Engine," SAE Technical Paper 2016-01-1007, 2016, doi: 10.4271/2016-01-1007.
22 Souhaite, P., Mokhtari, S. "Combustion system design of the new PSA Peugeot Citroen EB TURBO PURE TECH
engine," Proceedings: Internationaler Motorenkongress 2014, DOI: 10.1007/978-3-658-05016-0_5.
23 Budack, R., Kuhn, M., Wurms, R., Heiduk, T., "Optimization of the Combustion Process as Demonstrated on the
New Audi 2.01 TFSI," 24th Aachen Colloquium Automobile and Engine Technology 2015.
                                                   5-545

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
24 Wurms, R., Budack, R., Grigo, M., Mendl, G., Heiduk, T., Knirsch, S.  "The new Audi 2.01 Engine with
innovative Rightsizing," 36. Internationales Wiener Motorensymposium 2015.
25 Hatano, I, Fukushima, H., Sasaki, Y., Nishimori, K., Tabuchi, T., Ishihara, Y. "The New 1.6L 2-Stage Turbo
Diesel Engine for HONDA CR-V." 24th Aachen Colloquium - Automobile and Engine Technology 2015.
26 Steinparzer, F., Nefischer, P., Hiemesch, D., Kaufmann, M.,  Steinmayr, T. "The New Six-Cylinder Diesel Engines
from the BMW In-Line Engine Module." 24th Aachen Colloquium - Automobile and Engine Technology 2015.
27 Eder, T., Weller, R., Spengel, C., B6hm,J., Herwig, H., Sass, H.  Tiessen, I, Knauel, P. "Launch of the New
Engine Family at Mercedes-Benz." 24th Aachen Colloquium -  Automobile and Engine Technology 2015.
28 Koeberlein, D. "Cummins SuperTruck Program - Technology and System Level Demonstration of Highly
Efficient and Clean, Diesel Powered Class 8 Trucks." U.S. DOE Vehicle Technologies Office Annual Merit Review
and Peer Evaluation, June 9-12, 2015.
29 Gibble, J. "Volvo SuperTruck - Powertrain Technologies for Efficiency Improvement." U.S. DOE Vehicle
Technologies Office Annual Merit Review and Peer Evaluation, June 9-12, 2015.
30 Ltickert, P., Arndt, S., Duvinage, F., Kemmner, M., Binz, R., Storz, O., Reusch, M., Braun, T., Ellwanger, S.
"The New Mercedes-Benz 4-Cylinder Diesel Engine OM654 - The Innovative Base Engine of the New Diesel
Generation." 24th Aachen Colloquium - Automobile and Engine Technology 2015.
31 Busch, R., Jennes, J., Mtiller, J., Kriiger, M., Naber, D., Kauss, H. "Emission and Fuel Consumption Optimized
Turbo Charging of Passenger Car Diesel Engines." Vienna Motor Symposium, 2015.
32 Ruth, M. "ATP-LD; Cummins Next Generation Tier 2 Bin 2  Diesel Engine." .S. DOE Vehicle Technologies
Office Annual Merit Review and Peer Evaluation, May 14-17,  2013.
33 Ruth, M. "ATP-LD; Cummins Next Generation Tier 2 Bin 2  Diesel Engine." .S. DOE Vehicle Technologies
Office Annual Merit Review and Peer Evaluation, June 9-12, 2015.
34 U.S. DOE, April 8, 2015. Cummins Improving Pick-Up Truck Engine Efficiency with DOE and Nissan.
Accessed on 2/3/2016 at the following URL: http://energy.gov/eere/success-stories/articles/cummins-improving-
pick-truck-engine-efficiency-doe-and-nissan.
35 Ogihara, H., "Research Into Surface Improvement for Low Friction Pistons," SAE Technical Paper 2005-01-1647,
2005, doi: 10.4271/2005-01-1647.
36 Kim, Y., Kim, S.J., Lee, J., Lim, D. "Nanodiamond Reinforced PTFE Composite Coatings." MTZ 76 (2) pp 32-35,
February 2015.
37 Hanke, W., Ando, H., Fahr, M., Voigt, M. "Friction Reduction in Power Cylinder Systems for Passenger Car
Diesel Engines" MTZ 75 (2) pp 26-31, February 2014.
38 Landerl, C., Ruelicke, M., Durst, B., Mattes, W. "The New BMW Inline 6-Cylinder Gasoline Engine with
TwinPower Turbo, Direct Injection, and VALVETRONIC in the New BMW 7 Series." 24th Aachen Colloquium -
Automobile and Engine Technology 2015.
39 NeuBer, H.K. "The Car of the Future will continue to Fascinate People." Oral Presentation, Vienna Motor
Symposium, 2015.
40 Ragot, P. and Rebbert, M., "Investigations of Crank Offset and It's Influence on Piston and Piston Ring Friction
Behavior Based on Simulation and Testing," SAE Technical Paper 2007-01-1248, 2007, doi: 10.4271/2007-01-
1248.
41 Toyota Motor Corporation. 2009 Toyota Prius Service Manual. Toyota Technical Information System.
https://techinfo.toyota.com/.
42 Confer, K., Kirwan, J., and Engineer, N., "Development and  Vehicle Demonstration of a Systems-Level Approach
to Fuel Economy Improvement Technologies," SAE Technical Paper 2013-01-0280, 2013, doi: 10.4271/2013-01-
0280.
43 U.S. EPA, October 2014, Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy
Trends: 1975 Through 2014, Report EPA-420-R -14-023a, p. 74.
44 "Once-promising dual-clutch transmissions lose favor in U.S.," Automotive News, December 7, 2015
http://www.autonews.com/article/20151207/OEM06/312079988/once-promising-dual-clutch-transmissions-lose-
favor-in-u.s.
45 Newman, K., Kargul, J., and Barba, D., "Benchmarking and Modeling of a Conventional Mid-Size Car Using
ALPHA," SAE Technical Paper 2015-01-1140, 2015, doi: 10.4271/2015-01-1140.
46 Newman, K., Kargul, J., and Barba, D., "Development and Testing of an Automatic Transmission Shift Schedule
Algorithm for Vehicle Simulation," SAE Int. J. Engines 8(3):1417-1427, 2015, doi: 10.4271/2015-01-1142.
47 Moskalik, Andrew, Barba, Dan, and Kargul, John, "Investigating the Effect of Advanced Automatic
Transmissions on Fuel Consumption Using Vehicle Testing and Modeling," to be presented at the 2016 SAE
International Congress, April 2016.
                                                   j-j4o

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
48 Gall, I, "2011 Dodge Charger V6," http://www.caranddriver.com/reviews/2011-dodge-charger-v6-test-review,
(July 2011), accessed July 2016.
49 Gall, J., "2012 Dodge Charger SXT V6," http://www.caranddriver.com/reviews/2012-dodge-charger-sxt-v6-test-
review (January 2012), accessed July 2016.
50 N. Kim, A. Rousseau, H. Lohse-Bush, "Advanced Automatic Transmission Model Validation Using
Dynamometer Test Data," SAE 2014-01-1778, SAE World Congress, Detroit, April 2014.
51 N. Kim, H. Lohse-Bush, A. Rousseau, "Development of a model of the dual clutch transmission in Autonomie
and validation with dynamometer test data, International Journal of Automotive Technologies, March 2014, Volume
15, Jssucl, pp 263-271.
52 H.Naunheimer, et al., "Automotive Transmissions - Fundamentals, Selection, Design and applications," Springer
publications.
53 Dick, A., Greiner, J., Locher, A., and Jauch, F., "Optimization Potential for a State of the Art 8-Speed AT," SAE
International Journal of Passenger Cars - Mechanical Systems, 6(2):899-907, 2013, doi: 10.4271/2013-01-1272.
54 Aoki, T., Kato, H., Kato, N., and Masaru, M., "The World's First Transverse 8-Speed Automatic Transmission,"
SAE Technical Paper 2013-01-1274, 2013, doi: 10.4271/2013-01-1274.
55 Christoph Dorr, "The New Automatic Transmission 9G-TRONIC," presented at the 2014 Car Training Institute
Transmission Symposium, Rochester MI.
56 Gaertner, L. and Ebenhoch, M., "The ZF Automatic Transmission 9HP48 Transmission System, Design and
Mechanical Parts," SAE International Journal of Passenger Cars - Mechanical Systems,  6(2):908-917, 2013, doi:
10.4271/2013-01-1276.
57 Shidore, N. et. al. 2014. "Impact of Advanced Technologies onEngine Targets." Project VSS128, DOE Merit
Review, June.
58 Eckl, B., and D. Lexa. 2012. How Many Gears do the Markets Need? GETRAG. International CTI Symposium,
Berlin, Germany, December.
59 Darrell Robinette, 2014, "A DFSS Approach to Determine Automatic Transmission Gearing Content for
Powertrain-Vehicle System Integration, SAE International Journal of Passenger Cars - Mechanical Systems 7 (3).
60 Juergen Greiner, Martin Grumbach, Albert Dick, and Cristophe Sasse, 2015, "Advancement in NVH- and Fuel-
Saving Transmission and  Driveline Technologies," SAE technical paper 2015-01-1087
61NAS, 2011, p. 62 footnote.
62 NAS (2015), Prepublication Copy, p. 5-16.
63Wikimedia Commons. https://commons.wikimedia.org/wiki/File:ZF_Stufenautomatgetriebe_8HP70.jpg Photo
credit: Stefan Krause.
64 U.S. EPA, October 2014, Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy
Trends: 1975 Through  2014, Report EPA-420-R -14-023a, p. 75.
65 Start of Volume Production: New Generation of the ZF 8-Speed "Automatic Transmission in the BMW 5 Series,"
August 21, 2014, https://www.zf.com/corporate/en_de/magazine/magazin_artikel_viewpage_22067944.html.
66 Jens Meiners, "2016 Mini Cooper Clubman Revealed: Another Bigger, Four-Door Mini," Car and Driver, June
2015, http://www.caranddriver.com/news/2016-mini-clubman-revealed-news.
67 Gaertner, L. and Ebenhoch, M., "The ZF Automatic Transmission 9HP48 Transmission System, Design and
Mechanical Parts," SAE Int. J. Passeng. Cars - Mech. Syst. 6(2):908-917, 2013, doi: 10.4271/2013-01-1276.
68 "Land Rover To Demonstrate Latest Technical Innovation With The World's First 9-Speed Automatic
Transmission," Land Rover Media Centre, February 27, 2013, http://newsroom.jaguarlandrover.com/en-in/land-
rover/news/2013/02/rr_rre_9-speed_transmission_270213/.
69 Daimler. 2013. New Nine-Speed Automatic Transmission Debuts in the Mercedes-Benz E350 Blue Tec: Premier
of the new 9G-Tronic. Daimler, July 24. http://media.daimler.com/dcmedia/0-921-1553299-l-1618134-l-0-l-0-0-0-
0-1549054-0-l-0-0-0-0-0.html.
70 Motor Authority: Technology Preview: We Drive Honda's 10-Speed Automatic Transmission,
http://www.motorauthority.eom/news/l 100878_technology-preview-we-drive-hondas-10-speed-automatic-
transmission.
71 ZF, "Fuel Saving and Minimizing CO2 Emissions: 6% Lower Fuel Consumption,"
http://www.zf.com/corporate/en_de/products/innovations/8hp_automatic_transmissions/lower_consumption/lower_
consumption, html.
72 Dick, A., Greiner, J., Locher, A., and Jauch, F., "Optimization Potential for a State of the Art 8-Speed AT," SAE
Int. J. Passeng. Cars-Mech. Syst. 6(2):899-907, 2013, doi: 10.4271/2013-01-1272.


                                                    5-547

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
73 The New Generation of 8-Speed Automatic Transmission, ZF
http://www.zf.com/corporate/en_de/products/innovations/8hp_automatic_transmissions/8hp_automatic_transmissio
ahtml.
74 Greiner, I, Grumbach, M, Dick, A., and Sasse, C., "Advancement in NVH- and Fuel-Saving Transmission and
Driveline Technologies," SAE Technical Paper 2015-01-1087, 2015, doi: 10.4271/2015-01-1087.
75 Christoph Dorr, "The New Automatic Transmission 9G-TRONIC from Mercedes- Benz," presented at the 2014
CTI Symposium, Plymouth, MI.
76 Greiner, J. and Grumbach, M., "Automatic Transmission Systems Beyond 2020 - Challenges and Competition,"
SAE Technical Paper 2013-01-1273, 2013, doi: 10.4271/2013-01-1273.
77 Driveline News, Jan 22 2014, "BMW and Mini Strategy Revealed," http://www.drivelinenews.com/transmission-
insigh^mw-and-mini-transmission-strategy -revealed/.
78 Markus Nell, "BMW's flexible powertrain family with a new generation of transverse automatic transmissions,"
presented at the 2014 Car Training Institute Transmission Symposium, Rochester, MI.
79 Christoph Dorr, "The New Automatic Transmission 9G-TRONIC," presented at the 2014 Car Training Institute
Transmission Symposium, Rochester, MI.
80 Darrell Robinette and Daniel Wehrwein, "Automatic Transmission Technology Selection Using Energy Analysis"
presented at the CTI Symposium 9th International 2015 Automotive Transmissions, HEV and EV Drives.
81 Juergen Greiner, Martin Grumbach, Albert Dick, and Cristophe Sasse, 2015, "Advancement in NVH- and Fuel-
Saving Transmission and Driveline Technologies," SAE technical paper 2015-01-1087
82 Darrell Robinette, 2014, "A DFSS Approach to Determine Automatic  Transmission Gearing Content for
Powertrain-Vehicle System Integration, SAE International Journal of Passenger Cars  - Mechanical Systems 7 (3).
83 Hans Greimel, "ZF CEO: We're not chasing 10-speeds," Automotive News, November 23, 2014,
http://www.autonews.eom/article/20141123/OEM10/311249990/zf-ceo:-were-not-chasing-10-speeds.
84 NAS (2015), Prepublication Copy, p. 5-9.
85 U.S. EPA, October 2014, Light-Duty Automotive  Technology, Carbon Dioxide Emissions, and Fuel Economy
Trends: 1975 Through 2014, Report EPA-420-R -14-023a, p. 80.
86Wikimedia Commons, https://commons.wikimedia.org/wiki/File:Dual-clutch_transmission.svg, Credit: Xavax.
87 Carney, D. 2014. Honda's new 8-speed DCT uses a Torque Converter. SAE Automotive Engineering Magazine,
August 6.
88 NAS (2015), Prepublication Copy, p. 5-7.
89 NRC (2015), Prepublication Copy, p. 5-7.
90 Eckl, B.," DCT in the American Market: Transferring Customer Perceptions into Product Refinements,"
presented at the 2014 Car Training Institute Transmission Symposium, Rochester, MI.
91 Eckl, B., and D. Lexa. 2012. How Many Gears do the Markets Need? GETRAG. International Car Training
Institute Transmission Symposium, Berlin, Germany, December.
92Wikimedia Commons, https://commons.wikimedia.org/wiki/File:Toyota_Super_CVT-i_01.JPG, Photo credit:
Hatsukari715.
93Wikimedia Commons, https://commons.wikimedia.org/wiki/File:Pivgetriebe.png, Credit: Btideler Naumann
94 Shinji Morihiro, "Fuel Economy Improvement by Transmission" presented at the CTI Symposium 8th
International 2014 Automotive Transmissions, HEV and EV Drives.
95 Masayoshi Nakasaki and Yoshikazu Oota, "Key Technologies Supporting Belt-type CVT Evolution," presented at
the 2014 Car Training Institute Transmission Symposium,  Rochester MI.
96 Maruyama, F., Kojima, M., and Kanda, T., "Development of New CVT for Compact Car," SAE Technical Paper
2015-01-1091,  2015, doi:  10.4271/2015-01-1091.
97 Hakamagi, J., Kono, T., Habuchi,  R., Nishimura, N. et al., "Development of New Continuously Variable
Transmission for 2.0-Liter Class Vehicles," SAE Technical Paper 2015-01-1101, 2015, doi: 10.4271/2015-01-1101.
98 Shimokawa, Y., "Technology Development to Improve JATCO CVT8 Efficiency," SAE Technical Paper 2013-
01-0364, 2013, doi: 10.4271/2013-01-0364.
99 Lindsay Brooke, "JATCO's next-gen CVTs bring high ratio spreads, more efficiency," Automotive Engineering
Magazine, April 23, 2012, http://articles.sae.org/10947/.
100 Naotoshi, P. "Development of a New Generation CVT with Auxiliary Gear Box,"  SAE Technical Paper 2016-01-
1109, 2016, doi: 10.4271/2016-01-1109.
101 Naotoshi, P. "Development of a New Generation CVT with Auxiliary Gear Box,"  SAE Technical Paper 2016-01-
1109, 2016, doi: 10.4271/2016-01-1109.
102 Shimokawa, Y., "Technology Development to Improve JATCO CVTS Efficiency," SAE Technical Paper 2013-
01-0364, 2013, doi: 10.4271/2013-01-0364.

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
103 Maruyama, F., Kojima, M, and Kanda, T., "Development of New CVT for Compact Car," SAE Technical Paper
2015-01-1091, 2015, doi: 10.4271/2015-01-1091.
104 Inukai, K., Shibahara, A., Uchino, T., Keiichi, N. et al., "Development of High-Efficiency New CVT for Midsize
Vehicle," SAE Technical Paper 2013-01-0365, 2013, doi: 10.4271/2013-01-0365.
105 Hakamagi, I, Kono, T., Habuchi, R., Nishimura, N. et al., "Development of New Continuously Variable
Transmission for 2.0-Liter Class Vehicles," SAE Technical Paper 2015-01-1101, 2015, doi: 10.4271/2015-01-1101.
106 Don Sherman, Dec 2013, "The Unsinkable CVT: How the "Gearless" Transmission Is Getting Its Groove Back,"
Car and Driver, http://www.caranddriver.com/features/how-cvt-transmissions-are-getting-their-groove-back-
feature.
107 Shimokawa, Y., "Technology Development to Improve JATCO CVT8 Efficiency," SAE Technical Paper 2013-
01-0364, 2013, doi: 10.4271/2013-01-0364.
108 Maruyama, F., Kojima, M., and Kanda, T., "Development of New CVT for Compact Car," SAE Technical Paper
2015-01-1091, 2015, doi: 10.4271/2015-01-1091.
109 Inukai, K., Shibahara, A., Uchino, T., Keiichi, N. et al., "Development of High-Efficiency New CVT for Midsize
Vehicle," SAE Technical Paper 2013-01-0365, 2013, doi: 10.4271/2013-01-0365.
110 Hakamagi, I, Kono, T., Habuchi, R., Nishimura, N. et al., "Development of New Continuously Variable
Transmission for 2.0-Liter Class Vehicles," SAE Technical Paper 2015-01-1101, 2015, doi: 10.4271/2015-01-1101.
111 Mamiko Inoue, "Advanced CVT control to achieve both fuel economy and drivability," presented at the 2015 Car
Training Institute Transmission Symposium, Novi, MI.
112 Masayoshi Nakasaki and Yoshikazu Oota, "Key Technologies Supporting Belt-type CVT Evolution," presented
at the 2014 Car Training Institute Transmission Symposium, Rochester, MI.
113 Dana Holding Corp. 2014. Dana Advances Development of VariGlide™ Continuously Variable Planetary
Technology. PRNewswire, May 19.  http://www.prnewswire.com/news-releases/dana-advances-development-of-
variglide-continuously-variable-planetary-technology-259791981 .html.
114 Dick, A., J. Greiner, A. Locher, andF. Jauch. 2013. Optimization Potential for a State of the Art 8-Speed
AT. SAE 2013-01-1272.
115 Dick, A., Greiner, J., Locher, A., and Jauch, F., "Optimization Potential for a State of the Art 8-Speed AT," SAE
Int. J. Passeng. Cars-Mech. Syst. 6(2):899-907, 2013, doi: 10.4271/2013-01-1272.
116 Aoki, T., Kato, H., Kato, N., and Masaru, M., "The World's First Transverse 8-Speed Automatic Transmission,"
SAE Technical Paper 2013-01-1274, 2013, doi: 10.4271/2013-01-1274.
117 Christoph Dorr, "The New Automatic Transmission 9G-TRONIC," presented at the 2014 Car Training Institute
Transmission Symposium, Rochester, MI.
118 Christoph Dorr, "The New Automatic Transmission 9G-TRONIC," presented at the 2014 Car Training Institute
Transmission Symposium, Rochester, MI.
119 Martin, K. 2012. Transmission Efficiency Developments. SAE Transmission and Driveline Symposium:
Competition for the Future,  October 17-18. Detroit, Michigan,  [as cited in NAS (2015), Prepublication Copy, p. 5-
22.].
120 NSK Europe. 2014. New Low-Friction TM-Seal for Automotive Transmissions,
llSjx/ZHWHJlfejJi^^
121 Christoph Dorr, "The New Automatic Transmission 9G-TRONIC," presented at the 2014 Car Training Institute
Transmission Symposium, Rochester, MI.
122 Noles, J. 2013. Development of Transmission Fluids Delivering Improved Fuel Efficiency by Mapping
Transmission Response to Viscosity and Additive Changes. Presentation at the SAE Transmission & Driveline
Symposium, Troy, Michigan, October 16-17.[as cited in NAS (2015), Prepublication Copy, p. 5-25.].
123 NAS (2015), Prepublication Copy, p. 5-28.
124 NAS (2015), Prepublication Copy, p. 5-27.
125 Shimokawa, Y., "Technology Development to Improve JATCO CVT8 Efficiency," SAE Technical Paper 2013-
01-0364, 2013, doi: 10.4271/2013-01-0364.
126 Ando, T., Yagasaki, T., Ichijo, S., Sakagami, K. et al., "Improvement of Transmission Efficiency in CVT
Shifting Mechanism Using Metal Pushing V-Belt," SAE Int. J. Engines 8(3):1391-1397, 2015, doi: 10.4271/2015-
01-1103.
127 Greiner, J., Grumbach, M., Dick, A., and Sasse, C., "Advancement inNVH- and Fuel-Saving Transmission and
Driveline Technologies," SAE Technical Paper 2015-01-1087, 2015, doi: 10.4271/2015-01-1087.
128 Markus Nell, "BMW's flexible powertrain family with a new generation of transverse automatic transmissions,"
presented at the 2014 Car Training Institute Transmission Symposium, Rochester, MI.

                                                    5-549

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
129 Markus Nell, "BMW's flexible powertrain family with a new generation of transverse automatic transmissions,"
presented at the 2014 Car Training Institute Transmission Symposium, Rochester, MI.
130 Albert Dick, Juergen Greiner, Anton Locher, and Friedemann Jauch, "Optimization Potential for a State of the
Ary 8-Speed AT," SAEInternationalJournal of Passenger Cars -Mechanical Systems 6(2):2013, doi:
10.4271/2013-01-1272.
131 Darrell Robinette and Daniel Wehrwein, "Utilizing Energy Analysis Methods to Select Transmission
Technologies and Optimize Powertrain-Vehicle System Fuel Consumption," presented at the SAE 2014
Transmission and Driveline Symposium.
132 Markus Nell, "BMW's flexible powertrain family with a new generation of transverse automatic transmissions,"
presented at the 2014 Car Training Institute Transmission Symposium, Rochester, MI
133 Wikimedia Commons, https://commons.wikimedia.org/wiki/File:Bauma_2007_ZF_Drehmomentwandler.jpg,
Photo credit: Aconcagua.
134 Weissler, Paul. 2011. "2012 MazdaS Skyactiv achieves 40 mpg without stop/start." Automotive Engineering
Magazine, October 28.
135 Markus Nell, "BMW's flexible powertrain family with a new generation of transverse automatic transmissions,"
presented at the 2014 Car Training Institute Transmission Symposium, Rochester, MI
136 See Docket ID EPA-HQ-OAR-2010-0799, item number EPA-HQ-OAR-2010-0799-12013.
137 Nelson, P. et al., "Modeling the Performance and Cost of Lithium-Ion Batteries for Electric Drive Vehicles,"
Second Edition, Argonne National Laboratory, ANL-12/55 (December 2012).
138 See EPA Docket EPA-HQ-OAR-2015-0827, Microsoft Excel attachment to Docket Item titled "Data and Charts
for Selected Figures in Draft TAR Section 5.2 and 5.3."
139 See Table 3.5-25 of RIA, 2017-2025 FRM.
140 Howell, D., "Electric Drive Vehicle Battery Status and Trends," Vehicle Technologies Office, Department of
Energy. Presented at SAE Government-Industry Meeting, Washington, D.C., January 21, 2016.
141 Table 3.5-25 of RIA, 2017-2025 FRM.
142Nykvist, B.  andNilsson, M.; "Rapidly Falling Costs of Battery Packs for Electric Vehicles," Nature Climate
Change, March 2015; doi: 10.1038/NCLIMATE2564.
143 Green Car Congress, "GM and LG expand their relationship on Bolt EV; 12 components, including PEEM,"
October 20, 2015. Retrieved from http://www.greencarcongress.com/2015/10/20151020-gmlg.html.
144 Siemens, "Siemens and Valeo join forces for global leadership in powertrains for electric cars," Press Release,
April 18, 2016. Retrieved April 29,  2016 from http://www.siemens.com/press/PR2016040250COEN.
145 Schultz,  Jay W. and Huard, Steve,  "Comparing AC Induction with Permanent Magnet motors in hybrid vehicles
and the impact on the value proposition." White Paper, Parker Hannifin, 2013. Retrieved from
http://www.parkermotion.com/whitepages/Comparing_AC_and_PM_motors.pdf on Oct 14, 2015.
146 Widmer, James D. et al., "Electric vehicle traction motors without rare earth magnets," Sustainable Materials and
Technologies 3 (2015) 7-13.  MJiyM6dMM]ltJ^
147 Jenkins,  J., "A Closer Look at Switched Reluctance Motors," Charged Magazine, Cot/Nov 2012, pp. 26-28.
148 Ruoff, C., "A Closer Look at Torque Ripple," Charged Magazine, Jul/August 2015, pp. 22-29.
149 "pjrst look at ail-new Voltec propulsion system for 2G Volt; the only thing in common is a shipping cap," Green
Car Congress, October 29, 2014. Retrieved May 2, 2016 from
http://www.greencarcongress.com/2014/10/20141029-voltec.html.
150 "Chevrolet Introduces All-New 2016 Volt," General Motors Press Release, January 12, 2015.
http://media.gm.com/media/us/en/gm/news.detail.html/content/Pages/news/us/en/2015/Jan/naias/chevrolet/volt/0112
-volt-2016-intro.html.
151 Morris, C., "2016 Chevy Volt: GM's top electrification engineers on designing the all-new EREV,"
https://chargedevs.com/features/2016-chevy-volt-gms-top-electrification-engineers-on-designing-the-all-new-erev/.
152 Anwar, M. et al., "Power Dense and Robust Traction Power Inverter for the Second-Generation Chevrolet Volt
Extended-Range EV," SAE Int. J. Alt. Power. 4(1): 145-152, 2015, doi: 10.4271/2015-01-1201.
153 Grewe, T., General Motors, "GM RWD PHEV Propulsion System for the Cadillac CT6 Sedan," SAE paper
2016-01-1159, presented at 2016 SAE Hybrid and Electric Vehicle Technologies Symposium, February 10, 2016.
154 Grewe, T., General Motors, "Chevrolet Malibu Hybrid Propulsion System," presented at 2016 SAE Hybrid and
Electric Vehicle Technologies Symposium, February 10, 2016.
155 "Toyota Unveils Advanced Technologies in All-New Prius,"



                                                    5-550

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
156 Cole, J., "Exclusive: GM Exec Says Spark EV's 400 Ib-ft of Torque No Misprint," May 2, 2013 (source
confirmed by conversation with P. Savagian, February 2016). At http://insideevs.com/gm-general-says-spark-evs-
4001b-ft-of-torque-no-misprint/.
157 Whaling,  C., "Market Analysis of Wideband Gap Devices in Car Power Electronics." Retrieved March 9, 2016
from http://tec.ieee.org/2014/10/31/market-analysis-wideband-gap-devices-car-power-electronics/
158 Kimura, T. et al., "High-power-density Inverter Technology for Hybrid and Electric Vehicle Applications,"
Hitachi Review Vol. 63 No. 2 (2014). Retrieved on March 31, 2016 from
http://www.hitachi.com/rev/pdf/2014/r2014  02 106.pdf.
159 Goreham, I, "Toyota reveals breakthrough that will increase future Prius MPG by 10%," Torque News.
Retrieved May 2, 2016 from http://www.torquenews.com/1083/tovota-reveals-breakthrough-will-increase-future-
prius-mpg-10#sthash. i!xVsn43. dpuf.
leo Be& F_; "A Novei Design Methodology for a 1.5 KW DC/DC Converter in EV and Hybrid EV Applications,"
International Science Index, Electrical and Computer Engineering Vol:8, No:9,  2014.
161 Al Sakka, M. et al., "DC/DC Converters for Electric Vehicles,"  in Electric Vehicles - Modelling and Simulations,
Dr. Seref Soylu (Ed.), ISBN: 978-953-307-477-1, InTech, 2011. Available from: http://cdn.intechopen.com/pdfs-
wm/19583.pdf.
162 Safoutin, M., Cherry, I, McDonald, J., and Lee, S., "Effect of Current and SOC on Round-Trip Energy
Efficiency of a Lithium-Iron Phosphate (LiFePO4) Battery Pack," SAE Technical Paper 2015-01-1186, 2015,
doi:10.4271/2015-01-1186.
163 Idaho National Laboratory, "Steady State  Vehicle Charging Fact Sheet: 2015 Nissan Leaf," INL/EXT-15-34055.
Downloaded Jan. 26, 2016 from http://avt.inel.gov/pdf/fsev/SteadyStateLoadCharacterization2015Leaf.pdf.
164 Conlon, B., Blohm, T., Harpster, M., Holmes, A. et al., "The Next Generation "Voltec" Extended Range EV
Propulsion System," SAE Int. J. Alt. Power.4(2):248-259, 2015, doi: 10.4271/2015-01-1152.
165 Raghavan, A., "SENSOR: Smart Embedded Network of Sensors with Optical Readout," PARC and LG Chem
Power, presented at AABC 2014, February 2014.
166 http://www.uscar.0rg/guest/partnership/l/us-drive.
167 U.S. DRIVE Partnership, "US DRIVE Electrical and Electronics Technical TeamRoadmap," June 2013. Down
loaded on Dec. 23, 2015 from
http://wwwl.eere.energy.gov/vehiclesandfuels/pdfs/program/eett_roadmap June2013.pdf.
168 U.S. DRIVE Partnership, "US DRIVE Electrical and Electronics Technical Team Roadmap," June 2013, pp. 7-8.
169 Slenzak, J., "Next Generation Electrification Products: Focus on Integration  and Cost Reduction," Bosch, The
Battery Show 2015, Novi MI, September 15, 2015.
170  cf. U.S. DRIVE Electrical and Electronics Technical Team Roadmap, p. 10.
171 Kane, M., "Deep Dive: Chevrolet Bolt Battery Pack, Motor And More." Retrieved April 18, 2016 from
http://insideevs.com/deep-dive-chevrolet-bolt-battery-pack-motor-and-more/.
172 EPA, "Regulatory Impact Analysis: Final  Rulemaking for 2017-2025 Light-Duty Vehicle  Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards," 2012, p. 3-52. See Docket ID EPA-HQ-
OAR-2010-0799, docket item number EPA-HQ-OAR-2010-0799-12013.
173 EPA, "Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends:  1975
through 2015," December 2015, p 72.
174 EPA, "Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends:  1975
through 2015," December 2015, p 75.
175 EPA, "Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends:  1975
through 2015," December 2015, p 48.
176 Consumer Reports News, "Stop idling! Stop-start systems have  great promise
for saving fuel," June 29, 2012. Retrieved from http://www.consumerreports.org/cro/news/2012/06/stop-idling-stop-
start-svstems-have-great-promise-for-saving-fuel/index.htm.
177 Martinez, M., "More cars getting stop-start despite driver resistance," The Detroit News, October 8, 2014.
Retrieved from http://www.detroitnews.com/storv^usiness/autos/2014/10/07/autos-stop-start-driver-
resistance/16893597/.
178 Murray, C., "GM Chooses Micro- Over Mild-Hybrid in New Malibu," Design News, October 16, 2013.
179 Green Car Congress, "Schaeffler demo vehicle hits 2020 CAFE targets without electrification; 15% cut in fuel
consumption; 48V hybrid system for 2025," January 27, 2014. Retrieved from
http://www.greencarcongress.com/2014/01/20140127-schaeffler.html.


                                                    5-551

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
180 Halvorson, B.," Ultracapacitor Resistance Breaking Down Among Automakers?," Green Car Reports, June 2,
2015. Retrieved from http://www.greencarreports.com/news/1098550_ultracapacitor-resistance-breaking-down-
among-automakers.
181 Bragman, A., "2014 Chevrolet Malibu: Frist Drive," Cars.com, October 11, 2013. Retrieved from
https://www.cars.com/articles/2013/10/2014-chevrolet-malibu-first-drive/
182 Ford Motor Company, "70 Percent of Ford Lineup to Have Auto Start-Stop by 2017; Fuel Economy Plans
Accelerate," Press Release, December 12, 2013.
183 M. Everett, "Advancing Ultracapacitor Applications in Industrial and Transportation Applications," 2015 AABC,
Detroit.
184 Mazda, "Brake Energy Regeneration System: i-ELOOP continuously recovers kinetic energy as the vehicle
decelerates and reuses it as electricity," retrieved on April 1, 2016 from
http://www.mazda.com/en/innovation/technology/env/i-eloop/.
185 Blessing, U., "GETRAG 48V: How much hybridization is possible with the new vehicle power?." 14th VDI
Congress, 2014.
ise FEy5 "In_market Application of Start-Stop Systems in European Market," Final Report, P26844-01/ Al/ Ol/
61605, December 2011. Retrieved from
http://www.theicct.org/sites/default/files/FEV_LDV%20EU%20Technology%20Cost%20Analysis_StartStop%20O
verview.pdf.
187 EPA, "Regulatory Impact Analysis: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards," 2012, p. 3-52. See Docket ID EPA-HQ-
OAR-2010-0799, docket item number EPA-HQ-OAR-2010-0799-12013.
188 H-D Systems, "Future U.S. Trends in the Adoption of Light-Duty Automotive Technologies," Integrated Final
Report, Prepared for: American Petroleum Institute, September 2013, p.  60.
189 German, I, "Hybrid Vehicles: Technology Development and Cost Reduction," International Council on Clean
Transportation (ICCT), Technical Brief No. 1, July 2015.
190 Graf, F. et al., "AES: An Approach to an Integrated 12V/48V Energy Storage Solution," Proceedings of the 24th
Aachen Colloquium Automobile and Engine Technology 2015, pp. 1077-1090.
191 National Academy of Sciences, "Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles," 2015, Prepublication Copy, p. 4-39.
192 Bilek, M., "2013 Chevrolet Malibu - Refreshed for 2013, Chevy's midsize offering gains refinement and
features," DriveChicago.com. Retrieved May 5, 2016 from http://www.drivechicago.com/reviews/2013-chevrolet-
malibu-refreshed-for-2013-chevys-midsize-offering-gains-refinement-and-features_2424
193 Edmunds.com, "Full Edmunds Expert Review:  2013 Chevrolet Malibu." Retrieved on May 5, 2016 from
http://www.edmunds.com/chevrolet/malibu/2013/review/
194 General Motors, "2013  Malibu Eco Is The Most Fuel-Efficient Malibu Ever," Press Release, December 12, 2011.
http://media.gm.com/media/us/en/chevrolet/news.detail.html/content/Pages/news/us/en/201 l/Dec/1212_malibu/121
2_eassist.html.
195 Gross, O., "Defining Energy Storage System Requirements Based upon Passenger Vehicle Fleet Needs," 2015
Battery Congress.
196 General Motors, "CMC Introduces 2016 Sierra with eAssist," Retrieved March 9, 2016  from
http://www.gm.com/mol/m-2016-feb-0225-sierra-eAssist.html.
197 General Motors, "Chevrolet Introduces 2016 Silverado with eAssist," Retrieved March  11, 2016 from
http://www.gm.com/mol/m-2016-feb-0225-silverado-eAssist.html.
198 FEV, "Light-Duty Vehicle Technology Cost Analysis: 2013 Chevrolet Malibu ECO with eAssist BAS
Technology Study," Report FEV P311264, Contract No. EP-C-12-014, WA 1-9 (January 2014).

199 Howard, B., "Honda Fit: 86 mpg from the next  hyper-efficient hybrid," Extreme Tech, August 8, 2013. Retrieved
from http://www.extremetech.com/extreme/163462-honda-fit-86-mpg-from-the-next-hyper-efficient-hybrid.
200 Truett, R. et al., "Electric turbocharger eliminates lag, Valeo says," Automotive News, August 3, 2014. Retrieved
from http://www.autonews.com/article/20140803/OEM10/308049992/electric-turbocharger-eliminates-lag-valeo-
says.
201 Hanlon, M., "Controlled Power Technologies shows 48V electric supercharger," Gizmag, October 4, 2011.
Retrieved from http://www.gizmag.com/controlled-power-technologies-48v-electric-vtes-supercharger/20037/.

202 Fiat Chrysler Automobiles, "Business Plan Update 2014-2018," January 27, 2016. Retrieved from
http://www.fcagroup.com/en-

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
US/investor_relations/events_presentations/quarterly_results_presentations/FCA_2014_18	Business_Plan_Update.
pdf.
203 Sanchez, E., "Ram, Jeep Hybrids Revealed in FCA Future Business Plan," Truck Trend Network, January 27,
2016. Retrieved from http://www.trucktrend.com/news/1601-ram-jeep-hybrids-revealed-in-fca-ruture-business-
plan/.
204 Irwin, J., "Supplier Schaeffler in 48V Mild-Hybrid Vanguard," Wards Auto, January 15, 2016. Retrieved from
http://wardsauto.com/technologv/supplier-schaeffler-48v-mild-hvbrid-vanguard.
205 Voelcker, J., "2016 Hyundai Tucson Shows Two Different Hybrid Concepts In Geneva," Green Car Reports,
March 17, 2015. Retrieved from http://www.greencarreports.com/news/1097268 2016-hyundai-tucson-shows-two-
different-hybrid-concepts-in-geneva.
206 Duren, A., A123 Systems, "48V Battery System Design for Mild Hybrid Applications," SAE Hybrid & Electric
Vehicle Technologies  Symposium, February 2016.
207 Kok, D., Ford Motor Company, "Power of Choice: The Role of Hybrid Vehicle Technology in Meeting
Customer Expectations," The Battery Show 2014, September 18, 2014.
208 Harrop, P., "48V vehicle systems becoming significant," IDTechEx.com, August 12, 2015. Retrieved from
http://www.electricvehiclesresearch.com/articles/8266/48v-vehicle-systems-becoming-significant
209 Wiesenberger, J., Continental, "The Evolving EV and Hybrid Roadmap," The Battery Show 2015, Novi,
Michigan, September 15, 2015.
210 Stenzel, U., "48V Mild Hybrid Systems: Market Needs and Technical Solutions," AVL Powertrain Engineering,
AVL UK Expo 2014. Retrieved March 7, 2016 from
https://www.avl.com/documents/10138/1379144/AVL_UK_Expo_48V_Mild_Hybrid_Systems.pdf.
211 Brooke, L., "Ford accelerates research on 48-V mild hybrid systems," SAE International, February 12, 2015.
Retrieved from http://articles.sae.org/13908/.
212 Bosch, "The hybrid for everyone: Bosch's 48-volt system makes sense even in compact vehicles," Press Release,
September 4, 2015. Retrieved April 1, 2016 from http://www.bosch-
presse.de/presseforum/details.htm?txtID=7384&locale=en.
213 MacRae, C., "February 2015 management briefing: 48V mild hybrids (3)," retrieved from http://www.just-
auto.com/analysis/48v-mild-hybrids-3_idl56298.aspx.
214 Berman, B., "Hybrid Car Affordability Leaps Forward with P2 Technology," April 19, 2011. Retrieved from
http://www.hybridcars.com/hybrid-car-affordability-leaps-forward-p2-technology-29761/.
215 EPA, "Regulatory Impact Analysis: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards," 2012, p. 3-52. See Docket ID EPA-HQ-
OAR-2010-0799, docket item number EPA-HQ-OAR-2010-0799-12013.
216 Gehm, R., "Inside Honda's new two-motor 50-mpg Accord Hybrid," SAE International, December 18, 2013.
Retrieved from http://articles.sae.org/12666/.
217 Sessions, R., "2015 Toyota Camry SE Hybrid," Car and Driver, January 2015. Retrieved from
http://www.caranddriver.com/reviews/2015-toyota-camry-hybrid-test-review.
218 Green Car Congress, "Toyota details powertrain advances in Gen4 Prius; available E-Four system for all-wheel
drive (not for US)," October 13, 2015. Retrieved from http://www.greencarcongress.com/2015/10/20151013-
prius.html.
219 FEV, "Light Duty Technology Cost Analysis, Power-Split and P2 Hybrid Electric Vehicle Case Studies," Report
FEV07-069-303, October 10, 2011. https://www3.epa.gov/otaq/climate/documents/420rll015.pdf.
220 German, J., "Hybrid Vehicles Technology Development and Cost Reduction," Technology Brief, No. 1, July
2015,ICCT,p. 6.
221 Table 3.5-25 of RIA, 2017-2025 FRM.
222 National Academy  of Sciences (NAS), National Research Council (NRC), "Overcoming Barriers to Deployment
of Plug-in Electric Vehicles," NAS Committee on Overcoming Barriers to Electric-Vehicle Deployment, 2015.
223 New Qualified Plug-in Electric Drive Motor Vehicles, 26 U.S.C. § 30D.
224 See https://www.fueleconomy.gov/feg/taxphevb.shtml.
225 Cobb, J., "How Long Does the 2017 Chevy Bolt Have Before Federal Credits Begin Fading Away?"
HybridCars.com, January 13, 2016. Retrieved on January 27, 2016 from http://www.hybridcars.com/how-long-does-
the-2017-chevy-bolt-have-before-federal-credits-begin-fading-away/.

226 Shelton,  S., "US-Bound BMW 330e and X5 xDrive40e PHEVs Bow In LA," HybridCars.com, November 20,
2015. Retrieved from http://www.hybridcars.com/us-bound-bmw-330e-and-x5-xdrive40e-phevs-bow-in-la/.

                                                    5-553

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
227 Herman, B., "New Target Date for Outlander PHEV Introduction: Spring 2016", Retrieved April 25, 2016 from
http://www.plugincars.com/new-target-date-outlander-phev-introduction-spring-2016-130848.html.
228 Halvorson, B., "2017 Chrysler Pacifica Hybrid: More Details On 30-Mile Plug-In," Jan. 20, 2016,
http://www.greencarreports.eom/news/l 101960_2017-chrysler-pacifica-hybrid-more-details-on-30-mile-plug-in.
229 Cole, I, "Upcoming Honda PHEV To Have 40 Miles Of Electric Range," http://insideevs.com/upcoming-honda-
phev-40-miles-electric-range/.
230 General Motors, "The Results are In: More Range for the 2016 Volt," Press Release, August 4, 2015.
http://media.chevrolet.com/media/us/en/gm/news.detail.html/content/Pages/news/us/en/2015/aug/0804-volt-
range.html.
231 California Air Resources Board, "CCR Section 1962.2: 2018 and Subsequent Model Year Requirements," July
2014. Downloaded on May 26, 2016 from hfl|>://yra{wj^
232 World Bank, "The China New Energy Vehicles Program: Challenges and Opportunities," Report 61259, April
2011. Available at http://www-
wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/2011/04/20/000356161_20110420054039/Rend
ered/PDF/612590WPOPRTM01BOX358342B01PUBLICll.pdf.
233 International Council on Clean Transportation, "Renewed and Enhanced Subsidies for Energy-Efficient Vehicles
in China," Policy Update, November 2013. Available at
http://www.theicct.org/sites/default/files/publications/ICCTupdate_CNsubsidies_nov2013.pdf.
234 Anderman, M., "xEV Market Drivers and Trends; the Role of Regulations, Incentives, and Technology,"
presented at 2015 Advanced Automotive Batteries Conference, June 17, 2015.
235 Cadillac, "Cadillac CT6 Plugln Sets New Electric Luxury EV Performance Benchmark," Press Release, April 19,
2015. Retrieved on 10/23/2015 from
http://media.cadillac.com/media/cn/en/cadillac/news.detail.html/content/Pages/news/cn/en/2015/april/0419_PHEV.h
tml.
236 Cole, J., "Washington Expands $3,100 Electric Vehicle Incentive Program." Retrieved on April 19, 2016 from
http://insideevs.com/washington-launches-3100-electric-vehicle-incentive-program/.
237 Grewe, T., "Generation Two Voltec Drive System," presented at SAE Hybrid and Electric Vehicle Symposium,
February 2015.
238 See Table 3.5-25 of RIA, 2017-2025 FRM.
239 Green Car Congress, "Navigant Research Leaderboard puts LG Chem as leader for
Li-ion batteries for transportation," November 25, 2015. Retrieved from
http://www.greencarcongress.com/2015/ll/20151125-navigant.html.
240 Navigant Research,  "Navigant Research Leaderboard Report: Lithium Ion Batteries for Transportation,"
Summary Page. Retrieved from https://www.navigantresearch.com/research/navigant-research-leaderboard-report-
lithium-ion-batteries-for-transportation.
241 Kurylko, D. T., "BMW will boost i3's range," Automotive News, January 18, 2016. Downloaded on Jan. 19,
2016 from http://www.autonews.com/article/20160118/OEM05/301189994/bmw-will-boost-i3s-range.
242 BMW Group, "The new 2017 BMW i3 (94 Ah): More range paired to high-level dynamic performance," Press
Release, May 2, 2016. Retrieved on May 2, 2016 from
https://www.press.bmwgroup.com/usa/article/detail/T0259560EN_US/the-new-2017-bmw-i3-94-ah-more-range-
paired-to-high-level-dynamic-performance.
243 Autoblog, "VW e-Golf will get 30% range boost thanks to improved batteries," January 9, 2016. Downloaded on
Jan. 19, 2016 from http://www.autoblog.com/2016/01/09/volkswagen-e-golf-30-percent-range-increase-new-
batteries/.
244 Ford Motor Company, "Ford Investing $4.5 Billion in electrified Vehicle Solutions, Reimagining How To Create
Future Vehicle User Experiences," Press Release, December 10, 2015. Retrieved on April 19, 2016 from
https ://media.ford. com/content/fordmedia/fna/us/en/news/2015/12/10/ford-investing-4-5 -billion-in-electrified-
vehicle-solutions.pdf.
245 Hyundai Motor America, "2017 Hyundai loniq Model Lineup Makes U.S. Debut at New York International Auto
Show," Press Release, March 23, 2016. Retrieved on May 5, 2016 from
http://www.hyundainews.com/us/en/print/45159.
246 Tesla Motors website, Model 3 reservation page. Retrieved on April  19, 2016 from
https://www.teslamotors.com/model3.
247 Tesla Motors Press Release, "Tesla Model S Sales Exceed Target," March 31, 2013.
http://www.teslamotors.com/de_AT/blog/tesla-model-s-sales-exceed-target.

                                                     5-554

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
248 Konekamp, A., "The Tesla Model S Battery: A Battery Pack Analysis Study;" AVL List GmbH, 2015 Advanced
Automotive Battery Conference, Detroit MI June 15-19, 2015.
249 Rugh, I, Hovland, V. and Andersen, S., "Significant Fuel Savings and Emissions Reductions by Improving
Vehicle Air Conditioning," National Renewable Energy Laboratory, NREL/CP-5400-62232, presented at the 15th
Annual Earth Technologies Forum and Mobile Air Conditioning Summit, April 15, 2004.
250 Kim, N. et al., "Modeling Electrified Vehicles Under Different Thermal Conditions," Argonne National
Laboratory, presented at Society of Automotive Engineers Thermal Management Systems Symposium, September
29-October 1, 2015.
251 Chaney, L., et al., "Climate Control Load Reduction Strategies for Electric Drive Vehicles in Cold Weather,"
National Renewable Energy Laboratory, presented at Society of Automotive Engineers Thermal Management
Systems Symposium, September 29-October 1, 2015.
252 Chen, K., Bozeman, J., Wang, M., Ghosh, D. et al.,  "Energy Efficiency Impact of Localized Cooling/Heating for
Electric Vehicle," SAE Technical Paper 2015-01-0352, 2015, doi: 10.4271/2015-01-0352.
253 Green Car Congress, "Kia using SK Innovation NCM Li-ion cells in Soul EV," February 24, 2014. Retrieved
from http://www.greencarcongress.com/2014/02/20140224-kia. html.
254 Howell, D., "U.S. DOE Electric Drive Vehicle Battery R&D Impacts, Progress, and Plans," AABC 2015, Detroit
MI, June 2015.
255 Weissler, P., "Electronics diverge in engineering Ford's hybrid C-Max and plug-in Energi," SAE International,
January 9, 2013. Retrieved May 5, 2016 from http://articles.sae.org/11705/
256 Kessen, Jeff, and Duren, Angela,  "System Design Solutions for 48V Lithium-Ion Batteries," AABC 2015,
AABTAM Session 2, June 18, 2015.
257 Nuspl, G. et al., "Developing Battery Materials for Next Generation Applications," The Battery Show 2015, Novi
MI, 2015.
258 Oury, A., General Motors, "2016  Chevrolet Malibu  Hybrid Battery Pack," presented at AABC 2015, Detroit
Michigan, June 18, 2015.
259 Voelcker, J., "2016 Toyota Prius: A Few Details On Engine, Hybrid System Released," Green Car Reports,
October 12, 2015. Retrieved on May 5, 2016 from http://www.greencarreports.eom/news/l 100436_2016-toyota-
prius-a-few-details-on-engine-hybrid-sy stem-released.
260 Sion Power, "Sion Power's Licerion High Energy  Batteries," The Battery Show 2015, Novi, MI, September 2015.
261 PR Newswire, "Anesco and OXIS to Release Lithium Sulfur Battery Storage by 2016," Press Release, July  14,
2015. Retrieved from http://www.prnewswire.com/news-releases/anesco-and-oxis-to-release-lithium-sulfur-battery-
storage-by-2016-514802521 .html.
262 Morris, C., "OXIS Energy achieves 300 Wh/kg with lithium-sulfur cell," Charged Magazine, Nov/Dec 2014, p.
21. Retrieved from https://chargedevs.com/newswire/oxis-energy-achieves-specific-energy-of-300-whkg-with-
lithium-sulfur-cell/.
263 Ruoff, C., "Tesla Tweaks its Battery Chemistry: A Closer Look at Silicon Anode Development," Charged
Magazine, Jul/Aug 2015, p. 30.
264 Sawers, P., "Dyson acquires SaktiS for $90M to help commercialize 'breakthrough' solid-state battery tech."
Retrieved on April 1, 2016 from http://venturebeat.com/2015/10/19/dyson-acquires-sakti3-for-90m-to-help-
commercialize-breakthrough-solid-state-battery-tech/.
265 Vaughan, A and Carrington, D., "Dyson developing an electric car, according to government documents," The
Guardian, March 23, 2016. Retrieved on April  1,  2016  from
http://www.theguardian.com/environment/2016/mar/23/dyson-developing-electric-car-government-documents.
266 Seeo, "Bosch Has Groundbreaking Battery Technology for Electric  Vehicles." Retrieved on April 19, 2016 from
http://www.seeo.com/news/bosch-has-groundbreaking-battery-technology-for-electric-vehicles/
267 Business Wire, "Samsung SDI to  Supply Cylindrical EV Batteries to JAC Motors," November 23, 2015.
Retrieved from http://www.businesswire.com/news/home/20151123005451/en/Samsung-SDI-Supply-Cylindrical-
EV-Batteries-JAC
268 Anderman, M., "Battery Packs of Modern xEVs: A  Comprehensive Engineering Assessment: Extract," slide 18,
Total Battery Consulting, 2016.
269 Element Energy, "Cost and Performance of EV Batteries: Final Report for The Committee on Climate Change,"
March 21, 2012. Retrieved from http://www.element-energy.co.uk/wordpress/wp-content/uploads/2012/06/CCC-
battery-cost_-Element-Energy-report_March2012_Finalbis.pdf.
270 Faguy, Peter, "Overview of the DOE Advanced Battery R&D Program," presented at the 2015 U.S. DOE Vehicle
Technologies Office Annual Merit Review and Peer Evaluation, June 8, 2015.

                                                    5-555

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
271 Yeow, K. et al, "3D Thermal Analysis of Li-ion Battery Cells with Various Geometries and Cooling Conditions
Using Abaqus," 2012 SIMULIA Community Conference.
272 Brooke, L., "Battery guru: Future of 18650 cells unclear beyond Tesla S," SAE International, February 17, 2014.
Retrieved on May 5, 2016 from http://articles.sae.org/12833/.
273 Brooke, L., "GM unveils more efficient 2016 Volt powertrain," October 29, 2014. Retrieved on May 5, 2016
from http://articles.sae.org/13666/.
274 Nissan, "2016 Nissan LEAF Press Kit: Overview." Retrieved on May 5, 2016 from http://nissannews.com/en-
US/nissan/usa/presskits/us-2016-nissan-leaf-press-kit.
275 Kwon, Jason; "Samsung's Automotive Battery Strategy," Samsung SDI, AABC 2015, June 18, 2015.
276 Kia Motors, "Advanced battery pack for new Soul EV," Press Release, February 24, 2014. Retrieved April 4,
2016 from https://press.kia.com/eu/press/products/14_02_25_22%20-%20geneva%20soul%20ev%20batteries/.
277 Kia Motors America, "Kia Emergency Response Guide: Soul EV," September 2014.
278 Green Car Congress, "New 2016 Nissan LEAF with available 30 kWh pack for 107 milerange," September 10,
2015. Retrieved from http://www.greencarcongress.com/2015/09/20150910-leaf.html.
279 Schmitt, B., "2nd gen Leaf expected 2018: 60kWh NMC battery, 300 mile range, autonomous, CFRP." Retrieved
January 27, 2016 from http://dailykanban.com/2015/10/2nd-gen-leaf-expected-2018-60kwh-nmc-battery-300-mile-
range-autonomous-cfrp/.
280 Halvorson, B., "Nissan's 60-kWh, 200-Mile Battery Pack: What We Know So Far," Green Car Reports,
November 5, 2015.  Retrieved from http://www.greencarreports.com/news/1100775_nissans-60-kwh-200-mile-
battery-pack-what-we-know-so-far.
281 General Motors, "Drive Unit and Battery at the Heart of Chevrolet Bolt EV," Press Release, Jan. 11, 2016.
Retrieved on January 27, 2016  from
http://media.chevrolet.com/media/us/en/chevrolet/home.detail.html/content/Pages/news/us/en/2016/Jan/naias/chevy/
0111-bolt-du.html.
282 Markel, T., National Renewable Energy  Laboratory, "Plug-In HEV Vehicle Design Options and Expectations,"
NREL/PR-540-40630, presented at ZEV Technology Symposium, California Air Resources Board, Sacramento CA,
September 27, 2006.
283 McGrath, P., "Advanced Management and Protection of Energy Storage Devices (AMPED)," presented at The
Battery Show 2015, Novi MI, September 17, 2015.
284 Brown, Carlton;  "Generation 2 Lithium-Ion Battery Systems: Technology Trends and KPIs," Robert Bosch
Battery Systems, 2015 Advanced Automotive Battery Conference, Detroit MI June 15-19, 2015.
285 "Informal Testing of 2013 Volkswagen Jetta Hybrid Battery Usage," Memo to Docket EPA-HQ-OAR-2015-
0827, June 28, 2016.
286 Smith, K. et al., "Predictive  Models of Li-ion Battery Lifetime," National Renewable Energy Laboratory, 2015
Advanced Automotive Battery  Conference, Detroit MI June 15-19, 2015.
287 "2016 Chevrolet Volt Battery," 2015 Advanced Automotive Battery Conference, Detroit MI June 18, 2015.
288 ANL Advanced Powertrain  Research Facility (APRF), http://www.anl.gov/energy-systems/group/downloadable-
dy namometer-database.
289 Noland, D.," Life With Tesla Model S: Battery Degradation Update," Green Car Reports, October 26, 2015.
Retrieved from http://www.greencarreports.eom/news/l 100603_life-with-tesla-model-s-battery-degradation-
update/page-2.
290 Idaho National Laboratory,  "Vehicle Testing - Light Duty - BEV" (website). See https://avt.inl.gov/vehicle-
type/bev.
291 MyKiaSoulEV.com, topic on High Voltage Battery System. See
http://www.mykiasoulev.com/forum/viewtopic.php?p= 1923.
292 BMW Group, "BMW i3 Technical Data." See
http://www.bmw.com/com/en/newvehicles/i/i3/2015/showroom/technical_data.html.
293 Idaho National Laboratory,  "Battery Pack Laboratory Testing Results, 2014 BMW i3 EV - VTN 5486," retrieved
fromhttps://avt.inl.gov/sites/default/files/pdf/fsev/batteryi5486.pdf.
294 Idaho National Laboratory,  "Battery Pack Laboratory Testing Results, 2014 BMW i3 EV - VTN 5626," retrieved
fromhttps://avt.inl.gov/sites/default/files/pdf/fsev/batteryi5626.pdf.
295 Idaho National Laboratory,  "Battery Pack Laboratory Testing Results, 2014 BMW i3 EV - VTN 5655," retrieved
fromhttps://avt.inl.gov/sites/default/files/pdf/fsev/batteryi5655.pdf.
296 Idaho National Laboratory,  "Battery Pack Laboratory Testing Results, 2014 BMW i3 EV - VTN 5658," retrieved
fromhttps://avt.inl.gov/sites/default/files/pdf/fsev/batteryi5658.pdf.

                                                    5-556

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
297 General Motors, "Brownstown Battery Assembly Expands Capabilities, Will build battery system for 2015
Chevrolet Spark EV," Press Release, May 14, 2014. Retrieved from
http://media.chevrolet.com/media/us/en/chevrolet/news.detail.html/content/Pages/news/us/en/2014/May/0514-
brownstowahtml.
298 Uchida, T., "Analysis on Customers' Usage Data of Electric Vehicles," Honda R&D Co. Ltd. Automobile R&D
Center, 2015 Advanced Automotive Battery Conference, Detroit MI, June 15-19, 2015.
299 Lee, I, "Integrated Cabin and Battery Thermal Management System," Hanon Systems, presented at SAE 2015
Thermal Management Systems Symposium, Troy, MI, Sept. 30, 2015.
300 Wawzyniak, M., "Battery Thermal Management Architectures and Components," MAHLE Behr GmbH & Co
KG, presented at SAE 2015 Thermal Management Systems Symposium, Troy, MI, Sept. 30, 2015.
301 Jeckel, A.; Beste, F., and Glossman, T., "Thermal Management of the High Voltage Battery," Daimler AG,
presented at SAE 2015 Thermal Management Systems Symposium, Troy, MI, Sept. 30, 2015.
302 Gross, O. and Clark, S., "Battery Thermal Management in xEVs: Session Introduction," Fiat Chrysler
Automobiles, presented at AABC 2015, Detroit MI, June 2015.
303 Bower, G., "Chevy Bolt 200 Mile EV Battery Cooling and Gearbox Details," InsideEVs.com, January 18, 2016.
Retrieved from http://insideevs.com/chevy-bolt-200-mile-ev-battery-cooling-and-gearbox-details-bower/.
304 Doggett, S., "Leafs Batteries 'More Primitive' Than Those  in Tesla's First Prototype, Musk Says," Edmunds Auto
Observer, August 10, 2010. Retrieved from http://www.edmunds.com/autoobserver-archive/2010/08/leafs-batteries-
more-primitive-than-those-in-teslas-first-prototype-musk-says.htm.l.
305 Siry, D., "In Race to Market, Nissan's Electric Car Takes Shortcuts," Wired.com, January 22, 2010. Retrieved
from http://www.wired.com/2010/0 l/nissan-leaf-2./.
306 Moloughney, T., "No Active Thermal Management: Did Nissan Make the Right Call?" Retrieved from
http://www.plugincars.com/no-active-thermal-management-did-nissan-make-right-call.html.
307 Halvorson, B., "Nissan Leaf Likely To Get 'Hot Weather' Battery In Next Year," Green Car Reports, November
26, 2013. Retrieved from http://www.greencarreports.com/news/1088715_nissan-leaf-likely-to-get-hot-weather-
battery-in-next-year.
308 Wood, S., "Leveraging Automotive Lithium-Ion Technology for Commercial and Industrial Applications,"
Johnson Controls, presented at AABC 2015, Detroit MI, June 2015.
309 Porsche, "Porsche Mission E," Press Release, September 14, 2015. Retrieved from
http://newsroom.porsche.com/en/products/iaa-2015-porsche-mission-e-mobility-all-electrically-concept-car-
11391.html.
310 See Nelson, P. et al. "Modeling the Performance and Cost of Lithium-Ion Batteries for Electric Drive Vehicles,"
p. 19 and p. 23.
311 Wu, Y., "Assembly Processes for Lithium-Ion Batteries, 14.1.3. Coating Process," mLithium-Ion Batteries:
Fundamentals and Applications, Volume 4 of Electrochemical Energy Storage and Conversion, CRC Press (2015),
p. 483.
312 Gallagher, K., "PHEV and EV Battery Performance and Cost Assessment," Argonne National Laboratory, 2015
U.S. Department of Energy Vehicle Technologies Office Annual Merit Review and Peer Evaluation Meeting, June
2015.
313 Kia Motors, "Advanced Battery Pack for Kia Soul EV," Press Release, February 24, 2014. Retrieved on January
27, 2016 from http://www.kianewscenter.com/news/global-news/advanced-battery-pack-for-kia-soul-ev-
/s/45548fc8-7279-4da3-81b2-d30df557420c.
314 Kane, M., "Samsung SDI Presents Batteries That Enable 370 Miles (600 km) Of Range At 2016 NAIAS,"
retrieved January 27, 2016 from http://insideevs.com/samsung-sdi-presents-batteries-that-enable-370-miles-600-km-
of-range-at-2016-naias/.
315 Samsung SDI, "Samsung SDI Aims to Drive Battery Business Growth in North American Automobile Market
With Its Top Technology Leadership," Press Release, Business Wire, January 11, 2016. Retrieved January 27, 2016
from http://www.businesswire.com/news/home/20160111005560/en/Samsung-SDI-Aims-Drive-Battery-Business-
Growth.
316 Rauwald, C., "VW to Add Flat Batteries to Boost Sales of Electric Vehicles," Bloomberg Business, December
13, 2015. Retrieved on January 27, 2016 from http://www.bloomberg.com/news/articles/2015-12-13/vw-
developing-new-technology-to-integrate-flat-batteries-in-cars-ii4eee5d.
317 Nelson, Paul A.; Ahmed, Shabbir; Gallagher, Kevin G., and Dees, Dennis W., "Cost Savings for Manufacturing
Lithium Batteries in a Flexible Plant," Pre-publication draft, Argonne National Laboratory, 2015.
318 Corrigan, D., "XALT Energy Lithium-Ion Batteries for Heavy Duty Vehicle and Marine Applications," AABC
2015, June 18, 2015.

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
319 Paul, I, "Impact of Standardized Module Design - Commercial PEV and Second Life ESS Applications," AABC
2015, June 18, 2015.
320 Cobb, I, '"Not a compliance car,' GM says 2017 Chevy Bolt can meet demand of over 50,000 peryear,"
HybridCars.com, January 14, 2016. Downloaded on January 27, 2016 from http://www.hybridcars.com/not-a-
compliance-car-gm-says-2017-chevy-bolt-production-capacity-exceeds-50000-per-year/.
321 Roland Berger, "Consolidation Across the Li-Ion Battery Market Gaining Speed," Press Release, October 19,
2012. Retrieved from http://www.rolandberger.us/news/2012-10-19-Li-ion-Batteries-Bubble.html.
322 Ramsey, M., "Auto Industry's Ranks of Electric-Car Battery Suppliers Narrow," Wall Street Journal, August 20,
2015.
323 Green Car Congress, "Navigant Research ranks LG Chem as top automotive Li-ion battery manufacturer," June
25, 2013. Retrieved from http://www.greencarcongress.com/2013/06/navigant-20130625.html.
324 Bathion, Michael;  "Wanxiang Wins U.S. Approval to Buy Battery Maker A123," Bloomberg News, January 30,
2013. Retrieved Oct. 28, 2015 from http://www.bloomberg.com/news/articles/2013-01-29/wanxiang-wins-cfius-
approval-to-buy-bankrupt-battery-maker-al23.
325 Business Wire, "LG Chem, Panasonic, and Samsung SDI Score Highest in Assessment of Lithium Ion Battery
Manufacturers, According to Navigant Research," Press Release, December 3, 2015. Retrieved from
http://www.businesswire.com/news/home/20151203005014/en/.
326 Ogura, K., "LG Chem, Tesla tie-up could jolt Panasonic," Nikkei Asian Review, October 28,  2015. Retrieved
fromhttp://asia.nikkei.com/Business/Deals/LG-Chem-Tesla-tie-up-could-jolt-Panasonic?page=l
327 Business Wire, "Samsung SDI to Acquire Magna International's Battery Pack Business," Press Release, February
23, 2015. Retrieved on May 6, 2016 from http://www.businesswire.com/news/home/20150223005533/en/Samsung-
SDI-Acquire-Magna-International%E2%80%99s-Battery-Pack.
328 Planner, E. and Landers, P.; "Nissan Considers Shift to LG Chem Batteries," The Wall Street Journal, July 16,
2015. Retrieved on May 6, 2016 fromhttps://www.linkedin.com/pulse/nissan-considers-shift-lg-chem-batteries-
john-e-halajko.
329 Kim, Y., "Tesla approaches LG, Samsung for Model 3 batteries," The Korea Times, May 17,  2016. Retrieved on
May 19, 2016 fromhttp://www.koreatimes.co.kr/www/news/tech/2016/05/133_204903.html.
330 Clarke, G. M., "Lithium Resource Availability -  A Question of Demand," Proceedings 9th Advanced Automotive
Battery Conference (AABC), June 12, 2009, Long Beach CA.
331 Evans, K., "The Future of Electric Vehicles: Setting the Record Straight on Lithium Availability," Journal of
Energy Security, August 27 2009.
332 Boston Consulting Group, "Batteries for Electric Cars: Challenges, Opportunities, and the Outlook to 2020,"
January 7, 2010.
333 Gaines, L. and Cuenca, R., "Costs of Lithium-Ion Batteries for Vehicles," Argonne National Laboratory
ANL/ESD-42, May 2000.
334 Stafford, J., "Tesla And Other Tech Giants Scramble For Lithium As Prices Double," OilPrice.com, April 12,
2016. Retrieved on May 5, 2016 from http://oilprice.com/Energy/Energy-General/Tesla-And-Omer-Tech-Giants-
Scramble-For-Lithium-As-Prices-Double.html.
335 West, J., "Tesla Motors Inc ignites a lithium race among Albemarle, SQM and FMC," Financial Post, April 4,
2016. Retrieved on May 5, 2016 fromhttp:^usiness.financialpost.com/midas-letter/tesla-motors-inc-ignites-a-
lithium-race-among-albemarle-sqm-and-fmc.
336 Carnegie Mellon University,  "Lithium market fluctuations unlikely to significantly impact battery prices," Press
Release, May 3, 2016. Retrieved on May 5, 2016 from
https://engineering.cmu.edu/media/press/2016/05_03_lithium_market_fluctuations.html.
337 Spoken comments by Mark Reuss, Executive Vice President of Global Product Development, GM 2015 Global
Business Conference, Oct 1 2015, slide 52 in 2015_GBC_Combined_PDF_v3.pdf. See Docket Memo "Transcript of
Comments on Bolt EV Battery Cell Cost at GM 2015 Global Business Conference," EPA Docket EPA-HQ-OAR-
2015-0827.
338 General Motors, "General Motors 2015 Global Business Conference," Presentation,  October 1, 2015, slide 52 in
2015_GBC_Combined_PDF_v3 .pdf.
339 Cole, J., "GM: Chevrolet Bolt Arrives In 2016, $145/kWh Cell Cost, Volt Margin Improves $3,500,"
InsideEVs.com, October 2, 2015. Retrieved from http://insideevs.com/gm-chevrolet-bolt-for-2016-145kwh-cell-
cost-volt-margin-improves-3500/.
340 Kalhammer, F. et al., "Status and Prospects for Zero Emissions Vehicle Technology - Report  of the ARE
Independent Expert Panel 2007," Sacramento, California, US: State of California Air Resources  Board, 2007.

                                                     5-558

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
341 U.S. Advanced Battery Consortium, "USABC Goals for Advanced Batteries for EVs - CY 2020
Commercialization," Downloaded on December 28, 2015 from
http://www.uscar.org/commands/files_download.php?files_id=364.
342 Tataria, H. and Lopez, H.A., "EV Battery Development (Envia Systems)," in Section III.A. 1 p. 31 of "Fiscal Year
2013 Annual Progress Report for Energy Storage R&D," U.S. Department of Energy, DOE/EE-1038, February
2014.
343 Keller, G., "DOE's Efforts to Develop Hybrid Powertrain technologies for Heavy-Duty Vehicles," SAE 2015
Hybrid & Electric Vehicle Technologies Symposium, February 12, 2015.
344 Supplementary spreadsheet accompanying Nykvist, B. and Nilsson, M; "Rapidly Falling Costs of Battery Packs
for Electric Vehicles," Nature Climate Change, March 2015; doi: 10.1038/NCLIMATE2564, available at
http://www.nature.com/nclimate/journal/v5/n4/extref/nclimate2564-sl.xlsx.
345 Voelcker, I, "Nissan Leaf New Battery Cost: $5,500  For Replacement With Heat-Resistant Chemistry," Green
Car Reports, June 28, 2014. Retrieved from http://www.greencarreports.com/news/1092983_mssan-leaf-battery-
cost-5500-for-replacement-with-heat-resistant-chemistry.
346 Brockman, B., "Update on Nissan LEAF Battery Replacement" MYNissanLeaf.com, June 27, 2014. Retrieved
from http://www.mynissanleaf.com/viewtopic.php?f=4&t=17168.
347 Cole, J., "Nissan Prices LEAF Battery Replacement at $5,499, New Packs More Heat Durable," InsideEVs.com,
June 27, 2014. Retrieved from http://insideevs.comforeaking-nissan-prices-leaf-battery-replacement-5499-new-
packs-heat-durable/.
348 Voelcker, J., "Nissan Leaf $5,500 Battery Replacement Loses Money, Company Admits," Green Car Reports,
July 24, 2014. Retrieved from http://www.greencarreports.com/news/1093463_nissan-leaf-5500-battery-
replacement-loses-money-company-admits.
349 Cobb, J., "How Long Will An Electric Car's Battery Last?" HybridCars.com, April 30, 2014. Retrieved from
http://www.hybridcars.com/how-long-will-an-evs-battery-last/.
350 NewGMParts.com. See http://stores.revolutionparts.com/newgmparts.com/chevrolet/volt/20979876/2011-
year/base-trim/l-41-14-electric-gas-engine/electrical-cat/electrical-components-scat/?part_name=battery-assy
351 InsideEVs.com, "BMW 13 Battery Module Costs $1,715.60 - 8 Modules Per Car- Total Cost $13,725," January
26, 2015. Retrieved from http://insideevs.conVbmw-i3-battery-module-costs-1715-60-8-modules-per-car-total-cost-
13725/.
352 Tesla Motors,  "Roadster 3.0 Battery Upgrade," retrieved from
http://shop.teslamotors.com/collections/roadster/products/roadster-3-0-upgrade.
353 Ramsey, M., "Tesla Gets Boost From Korean Battery Maker LG Chem," The Wall Street Journal, October 28,
2015. Retrieved from http://www.wsj.com/articles/tesla-gets-boost-from-korean-battery-maker-lg-chem-
1446007554.
354 Smart USA, Battery Assurance Plus Brochure. Retrieved on May 6, 2016 from
http://www.smartusa.com/resources/doc/smart_battery-assurance-plus_SEO.pdf.
355 PluginCars.com, "Smart Electric Drive Review." Retrieved from http://www.plugincars.com/smart-ed
356 White, J., "The Smart, a Very Cheap Electric Car With a Very Expensive Battery," The Wall Street Journal,
August 7, 2013. Retrieved from http://blogs.wsj.com/corporate-intelligence/2013/08/07/the-smart-a-very-cheap-
electric-car-with-a-very-expensive-battery/.
357 Kane, M., "UK Pricing Reveals 30 kWh 2016 Nissan LEAF Costs Just £1,600 More Than 24 kWh Version,"
InsideEVs.com, September 14, 2015. Retrieved from http://insideevs.com/uk-pricing-reveals-30-kwh-2016-nissan-
leaf-costs-just-1600-more-than-24-kwh-version/.
358 Lux Research, "Tesla Motors' Gigafactory Will See More Than 50% Overcapacity in its Li-ion Production,"
Press Release, September 3, 2014. Retrieved from http://www.luxresearchinc.com/news-and-events/press-
releases/read/tesla-motors%E2%80%99-gigafactory-will-see-more-50-overcapacity-its-li.
359 Energy Policy Act of 2005, Pub. L. no.  109-58, 119 Stat 594 (2005). Print.
http://www.gpo.gov/fdsvs/pkg/PLAW-109publ58/pdf/PLAW-109publ58.pdf.
360 Hyundai USA, "Hyundai proudly hands keys to first Tucson Fuel Cell customer at Tustin Hyundai," Press
Release, June 13, 2015. Retrieved fromhttps://www.hyundaiusa.com/about-
hvundai/news/Corporate HYUNDAI PROUDLY HANDS KEYS TO FIRST TUCSON FUEL CELL GUSTO
MER AT TUSTIN HYUNDAI-20140613.aspx.
361 Toyota Motor Corporation, "First Toyota Mirai owners get a jump on the future," Press Release, October 21,
2015. Retrieved from http://www.tovotanewsroom.com/releases/tovota+mirai+owners+jump+future.htm.


                                                    5-559

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
362 Honda Motor Company, "Honda to launch next generation advanced powertrain vehicles by 2018, FCV Concept
makes North American debut," Press Release, January 12, 2015.  Retrieved from
http://world.honda.com/worldnews/2015/4150113FCV-Concept-North-American-Debut/index.html.
363 CARD, "Annual evaluation of fuel cell electric vehicle deployment and hydrogen fuel station network
development," Sacramento: California Air Resources Board, 2015.
364 Directed Technologies Inc (DTI). "Mass-Production Cost Estimation for Automotive Fuel Cell Systems 2010
U.S. DOE Merit Review." Presentation, DOE Annual Merit Review and Peer Evaluation Meeting. US Department
of Energy. Washington, D.C. June 7-11, 2010.
365 Papageorgopoulos, D., "Fuel Cells Program - Plenary Presentation," Presentation, DOE Annual Merit Review
and Peer Evaluation Meeting, US Department of Energy. Arlington, VA. June 8, 2015.
366 Mass production cost estimation of direct H2 PEM fuel cell systems for transportation applications: 2014 update.
Arlington: Brian James, Jennie Moton, and Whitney Colella, 2014.
367 U.S. Department of Energy Hydrogen and Fuel Cells Program, "Record 15015: Fuel Cell System Cost - 2015,"
https://www.hvdrogen.energy. gov/pdfs/15015  fuel cell  system  cost 2015.pdf. 2015.
368 U.S. Department of Energy Hydrogen and Fuel Cells Program, "Record 15015: Fuel Cell System Cost - 2015,"
https://www.hvdrogen.energy. gov/pdfs/15015  fuel cell  system  cost 2015.pdf. 2015.
369 James, B.,  "Fuel cell vehicle and bus cost analysis," Presentation, DOE Annual Merit Review and Peer
Evaluation Meeting, US Department of Energy. Arlington, VA. June 10, 2015.
370 Department of Energy, "Fuel cell technologies office multi-year research, development, and demonstration plan,"
Washington, DC: US Department of Energy, 2012.
371 Mass production cost estimation of direct H2 PEM fuel cell systems for transportation applications: 2014 update.
Arlington: Brian James, Jennie Moton, and Whitney Colella, 2014.
372 Mass production cost estimation of direct H2 PEM fuel cell systems for transportation applications: 2014 update.
Arlington: Brian James, Jennie Moton, and Whitney Colella, 2014.
373 James, B.,  "Hydrogen storage cost  analysis," Presentation, DOE Annual Merit Review and Peer Evaluation
Meeting. US Department of Energy. Arlington, VA. June 9, 2015.
374 James, B.,  "Hydrogen storage cost  analysis," Presentation, DOE Annual Merit Review and Peer Evaluation
Meeting. US Department of Energy. Arlington, VA. June 9, 2015.
375 Department of Energy, "Fuel cell technologies office multi-year research, development, and demonstration plan,"
Washington, DC: US Department of Energy, 2012.
376 Memo to Docket EPA-HQ-OAR-2015-0827, "Reports on Pressure Vessel System Cost Provided by Strategic
Analysis Inc. to  CARD,"  provided to  California Air Resources Board, August 14, 2015.
377 Oak Ridge National Laboratory, "Transition to hydrogen fuel cell vehicles & the potential hydrogen energy
infrastructure  requirement," Oak Ridge: Oak Ridge National Laboratory, 2008.
378 National Research Council, "Transitions to alternative transportation technologies: A focus on hydrogen,"
Washington, DC: National Research Council of the National Academies, 2008.
379 Analyzing the transition to electric drive in California. Knoxville: David Greene, Sangsoo Park, and Changzheng
Liu, 2013.
380 National Research Council, "Transitions to alternative vehicles and fuels," Washington, DC: National Research
Council of the National Academies, 2013.
381 CARD, "Annual evaluation of fuel cell electric vehicle deployment and hydrogen fuel station network
development," Sacramento: California Air Resources Board, 2015.
382 California Fuel Cell Partnership, "A California road map: The commercialization of hydrogen fuel cell vehicles,"
Sacramento: California Fuel Cell Partnership, 2012.
383 CARD, "EMFAC 2014 Volume III- Technical documentation," Sacramento:  California Air Resources Board,
2015.
384IHS Automotive, "Global light vehicle sales summary," Englewood: IHS Automotive, 2015.
385 IHS Global Insight, "Guide to the World Industry Service forecast database," Englewood: IHS Global Insight.
Accessed 28 October 2015.
386 European Automobile Manufacturers Association, "Historical series: 1990-2014: New passenger car registrations
by country," Brussels: ACEA (European Automobile Manufacturers Association), Accessed 28 October 2015.
387 CARD, "Annual evaluation of fuel cell electric vehicle deployment and hydrogen fuel station network
development," Sacramento: California Air Resources Board, 2015.
388 Department of Energy, "Fuel cell technologies office multi-year research, development, and demonstration plan,"
Washington, DC: US Department of Energy, 2012.

                                                    5-560

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
389 Stetson, N., "Hydrogen storage program area plenary presentation." Presentation, DOE Annual Merit Review and
Peer Evaluation Meeting, US Department of Energy. Arlington, VA. June 8, 2015.
390 U.S. Department of Energy Hydrogen and Fuel Cells Program, "Record 15013: Onboard Type IV Compressed
Hydrogen Storage System - Cost and Performance Status 2015,"
https://www.hvdrogen.energy.gov/pdfs/15013 onboard storagejerformance cost.pdf. 2015.
391 Hyundai USA, "Hyundai proudly hands keys to first Tucson Fuel Cell customer at Tustin Hyundai," Press
Release, June 13, 2015. Retrieved fromhttps://www.hyundaiusa.com/about-
hvundai/news/Corporate  HYUNDAI  PROUDLY HANDS KEYS TO FIRST  TUCSON FUEL  CELL GUSTO
MER AT TUSTIN  HYUNDAI-20140613.aspx.
392 Honda Motor Company, "Honda exhibits world premiere of Clarity Fuel Cell, planned production model of its
all-new fuel cell vehicle at 44th Tokyo Motor Show 2015," Press Release, October 28, 2015. Retrieved from
http://world.honda.com/news/2015/4151028eng.html.
393 Konno, N., Mizuno, S., Nakaji, H., and Ishikawa, Y., "Development of compact and high-performance fuel cell
stack," SAE Int. J. Alt. Power. 4(1):123-129, 2015.
394 Honda Motor Company, "Honda exhibits world premiere of Clarity Fuel Cell, planned production model of its
all-new fuel cell vehicle at 44th Tokyo Motor Show 2015," Press Release, October 28, 2015. Retrieved from
http://world.honda.com/news/2015/4151028eng.html.
395 Toyota Motor Corporation, "The Toyota Mirai brings the future to your driveway," Press Release, November 17,
2014. http://tovotanews.pressroom.tovota.com/releases/tovota+mirai+fcv+future+novl7.htm.
396 Stetson, N., "Hydrogen storage program area plenary presentation." Presentation, DOE Annual Merit Review and
Peer Evaluation Meeting, US Department of Energy. Arlington, VA. June 8, 2015.
397 Toyota Motor Corporation, "BMW group and Toyota Motor Corporation agree to further strengthen
collaboration," June 29, 2012. Retrieved from
http://pressroom.tovota.com/releases^mw+group+tovota+motor+corporation+agree+further+strengthen+collaborati
oahtm.
398 Daimler, "The strategic cooperation between Daimler and the Renault-Nissan alliance forms agreement with Ford
to accelerate commercialization of fuel cell electric vehicle technology,"  Press Release, January 28, 2013. Retrieved
from http://www.daimler.com/dccom/0-5-7171-l-1569731-l-0-0-0-0-0-12037-0-0-0-0-0-0-0-0.html.
399 General Motors Corporation,  "GM, Honda to collaborate on next-generation fuel  cell technologies," Press
Release, July 2, 2013. Retrieved from
http://media.gm.com/media/us/en/gm/news.detail.html/content/Pages/news/us/en/2013/Jul/0702-gm-honda.html.
400 Toyota Motor Corporation, "Lexus LF-LC flagship concept revealed at Tokyo Motor Show," Press  Release,
October 28, 2015. Retrieved from http://pressroom.tovota.com/releases/lexus+lf-
fc+concept+tokvo+motor+show.htm.
401 BMW Group,  "BMW  Group innovation days 2015: Drive technologies of the future," Press Release, July 2,
2015. Retrieved from https://www.press.bmwgroup.com/global/pressDetail.html?title=bmw-group-innovation-davs-
2015-drive-technologies-of-the-future&outputChannelId=6&id=T0223224EN&left  menu  item=node_5237.
402 Audi USA, "The Audi A7 sportback h-tron Quattro," Press Release, November 19, 2014. Retrieved from
http://www.audiusa.com/newsroom/news/press-releases/2014/ll/audi-a7-sportback-h-tron-quattro.
403 NESCAUM, "8 state alliance releases plan to put 3.3 million zero-emission vehicles on the road," May 29, 2014.
Retrieved from http://www.nescaum.org/topics/zero-emission-vehicles/press-release-8-state-alliance-releases-plan-
to-put-3-3-million-zero-emission-vehicles-on-the-road.
404 CARD, "International  alliance on zero-emission vehicles grows to 11  partners," September 29, 2015. Retrieved
fromhttp://www.arb.ca.gov/newsrel/newsrelease.php?id=761.
405 National Research Council Canada, "Full-Scale Wind Tunnel Investigation of Light-Duty Vehicle
Aerodynamics," NR AL-2014-0017, March 2014.
406 Larose, G., Belluz, L., Whittal, I., Belzile, M. et al., "Evaluation of the Aerodynamics of Drag Reduction
Technologies for Light-duty Vehicles: a Comprehensive Wind Tunnel Study," SAE Int. J. Passenger Cars - Mech.
Syst. 9(2):2016, doi:  10.4271/2016-01-1613.
407 National Research Council Canada "Evaluation of the aerodynamics of new drag reduction technologies for
light-duty vehicles: results of Phase III" LTR-AL-2015-0061, February 23, 2016.
408 U.S. EPA, "EPA Staff Informal Survey of Aerodynamic Technologies Represented at the
2015 North American International Auto Show (NAIAS)," memo to Docket EPA-HQ-OAR-2015-0827, May 16,
2016.
409 Arai, M.,  "Development of the Aerodynamics of the New Nissan Murano," Society of Automotive Engineering
(SAE) Technical Paper #2015-01-1542, 2015.
                                                    5-561

-------
                                    Technology Cost, Effectiveness, and Lead-Time Assessment
410 Madsen, A., "2015 Acura TLX Body Structure Review," Honda R&D Americas, Inc., 2015 Great Designs in
Steel Conference, Livonia, MI, May 13, 2015.
411 Udy, I, "2015 Ford F-150 Is Most Aerodynamic F-Series Ever," Motor Trend, August 26, 2014. Retrieved from
http://www.motortrend.com/news/2015-ford-f-150-is-most-aerodynamic-f-series-ever/.
412 National Research Council Canada, "Evaluation of the aerodynamics of new drag reduction technologies for
light-duty vehicles: results of Phase II," LTR-AL-2015-0273, Issued for Client's Comments, March 22, 2015.
413 California Air Resources Board, "Technical Analysis of Vehicle Load-Reduction Potential for Advanced Clean
Cars," Request for Proposal No. 13-313, July 30, 2013.
414 CONTROLTEC LLC, "Technical Analysis of Vehicle Load Reduction Potential for Advanced Clean Cars
(Contract 13-313, Final Report, prepared for CARB and CA EPA, April 29, 2015.
415 Chappell, L., "For mpg gains, tiremakers deliver - consumers shrug," Automotive News, December 1, 2014.
416 Automotive World, "From radial to radical - where next for light vehicle tyres?," March 2015.
417 Bridgestone,  "Bridgestone's 'ologic technology' hits the road," Press Release, January 8, 2014. Retrieved on
March 25, 2016  from http://www.bridgestone.eu/corporate/press-releases/2014/01/bridgestones-ologic-technology-
hits-the-road/.
418 Donley, T., "Improving Vehicle Fuel Efficiency Through Tire Design, Materials, and Reduced Weight," Cooper
Tire & Rubber Co., presentation at 2014 DOE Annual Merit Review, Project VSS083.
419 "Light Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975 Through
2015, Executive Summary," U.S. EPA, EPA-420-S-15-001 December 2015,
http://www3 .epa.gov/fueleconomy/fetrends/1975-2015/420s 15001 .pdf.
420 "CMC Introduces All-New 2017 Acadia," January 12, 2016,
http://media.gmc.com/media/us/en/gmc/home.detail.html/content/Pages/news/us/en/2016/Jan/naias/gmc/0112-
acadia.html.
421 Chrysler's T&C minivan renamed "Pacifica," cuts 250 Ibs. with steel, aluminum, magnesium," January  12, 2016,
"http://www.repairerdrivennews.com/2016/01/12/chryslers-tc-minivan-renamed-pacifica-cuts-250-lbs-with-
aluminum-steel-magnesium/.
422 Ducker Worldwide, "2015 North American Light Vehicle Aluminum Content Study, Executive Summary," June
2014, }itt2://w]iTvxi!^^
423 Composites Forecasts and Consulting, December 2015.
424 Email of Abey Abraham of Ducker Worldwide to Cheryl Caffrey, 2/25/2016.
425 http://www.freep.com/story/money/cars/mark-phelan/2015/06/24/2016-chevrolet-cruze-chevy-
challenges/71265088/.
426 http://www.worldautosteel.org/why-steel/steel-muscle-in-new-vehicles/2016-chevy-malibu-larger-lighter-more-
efficient-with-hss/.
427 "Cadillac XT5's new platform cuts weight- at less cost," SAE Automotive Engineering, April 2016.
428 "The New Audi Q7- Sportiness,  efficiency, premium comfort," December 12, 2014
https://www.audiusa.com/newsroom/news/press-releases/2014/12/the-new-audi-q7-sportiness-efficiency-premium-
comfort.
429 Mg Showcase Issue 12, International Magnesium Association, Spring 2010,
http://www.intlmag.org/showcase/mgshowcasel2_marl0.pdf.
430 "2017 GMC Acadia loses 700 pounds, gains everywhere else," Autoblog, January 12,  2016,
http://www.autoblog.com/2016/01/12/2017-gmc-acadia-detroit-official/.
431 "New GMC Acadia is Sportier, with More Safety Features," Detroit Free Press, Greg Gardner, January  12, 2016,
http://www.freep.com/story/money/cars/general-motors/2016/01/12/new-gmc-acadia-sportier-more-safety-
features/78519228/.
432 "Response to Peer Reviews for 'Light-Duty Truck Weight Reduction Study with Crash Model, Feasibility and
Cost Analysis - Project Number: T8009-130016', September 24, 2015.
433 "Automotive  Aluminum - Part of the Solution," Automotive Megatrends, March 17, 2015,
http://www.drivealuminum.org/research-resources/PDF/Speeches%20and%20Presentations/2015/automotive-
megatrends.
434 "Light-Duty Vehicle Mass Reduction and Cost Analysis -Midsize Crossover Utility Vehicle", EPA-420-R-12-
026, August 2012.

435 General Motors, "General Motors 2015 Global Business Conference," Presentation, October 1, 2015, Slides 43-
45 in document, https://www.gm.com/content/dam/gm/events/docs/5194074-596155-ChartSet-10-l-2015.

                                                     5-562

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
436 "Identifying Real world Barriers to Implementing Lightweighting Technologies and Challenges in Estimating the
Increase in Costs," Center for Automotive Research, Jay Baron, PhD, January 2016
http://www.cargroup.org/?module=Publications&event=View&pubID=128.
437 "Mass Reduction and Cost Analysis - Light Duty Pickup Truck Model years 2020-2025, FEV North America,
Inc., EPA-420-R-15-006, http://www3.epa.gov/otaq/climate/documents/mte/420rl5006.pdf. 2015.
438 Mass Reduction for Light Duty Vehicles for Model Years 2017-2025," Final Report, DOT HS 811 666, U.S.
DOT NHTSA, August 2012, ftp://ftp.nhtsa.dot.gov/CAFE/2017-25_Final/811666.pdf.
439 "Update to Future Midsize Lightweight Vehicle Findings in Response to Manufacturer Review and IIHS Small
Overlap Testing," NHTSA, DOT HS 812 237, February 2016.
http://www.nhtsa.gov/staticfiles/rulemaking/pdf/cafe/812237_LightWeightVehicleReport.pdf.
440 "Evaluating the Structure and Crashworthiness of a 2020 Model-year, Mass Reduced Crossover Vehicle using
FEA Modeling," California Air Resources Board and Lotus Engineering, August 31, 2012.
441 "An Assessment of Mass Reduction Opportunities for a 2017 - 2020 Model Year Vehicle Program," Lotus
Engineering, Inc. for ICCT, March 2010.
442 "Venza Aluminum BIW Concept Study," April 2013, http://www.drivealuminum.org/research-
resources/PDF/Research/2013/venza-biw-full-study.
443" http://energv.gov/sites/prod/files/2015/12/f27/06%20-%20Lightweight%20Materials.pdf.
444 Skszek, T., Conklin, J, Zaluzec, M, Wagner, D.,"Multi-Material Lightweight Vehicles: Mach-II Design," U.S.
DOE, Ford, Vehma, June 17, 2014, http://energy.gov/sites/prod/files/2014/07/fl7/lm088_skszek_2014_o.pdf.
445 Mascarin,A., Hannibal,T., Raghunathan, A., Ivanic, Z., Francfort, J., "Vehicle Lightweighting: 40% and 45%
Weight Savings Analysis: Technical Cost Modeling for Vehicle Lightweighting,"  INL/EXT-14-33863, April 2015,
http://avt.inl.gov/pdf/TechnicalCostModel40and45PercentWeightSavings.pdf.
446 Skszek, T., Zaluzec, M., Conklin, J., and Wagner D., "MMLV: Project Overview," SAE Technical Paper 2015-
01-0407, 2015 doi: 10.4271/2015-01-0407.
447 Bolar, N., Buchler, T., Li, A., and Wallace, J., "MMLV: Vehicle Durability Design, Simulation and Testing,"
SAE Technical Paper 2015-01-1613, 2015, doi: 10.4271/2015-01-1613.
448 Bushi, L., Skszek, T., and Wagner, D., "MMLV: Life Cycle Assessment," SAE Technical Paper 2015-01-1616,
2015,  doi: 10.4271/2015-01-1616.
449 Chen, X., Conklin, J., Carpenter, R.,  Wallace, J. et al., "MMLV: Chassis Design and Component Testing," SAE
Technical Paper 2015-01-1237, 2015, doi: 10.4271/2015-01-1237.
450 Chen, Y., Board, D., Faruque, O., Stancato, C. et al, "MMLV: Crash Safety Performance," SAE Technical Paper
2015-01-1614, 2015, doi: 10.4271/2015-01-1614.
451 Conklin, L., Beals, R., and Brown, Z., "BIW Design and CAE," SAE Technical Paper 2015-01-0408, 2015, doi:
104271/2015-01-0408.
452 Corey, N., Madin, M., and Williams, R., "MMLV: Carbon Fiber Composite Engine Parts," SAE Technical Paper
2015-01-1239, doi:  10.4271/2015-01-1239.
453 Gur, Y., Pan, J., Huber, J., and Wallace, J., "MMLV: NVH Sound Package Development and Full Vehicle
Testing," SAE Technical Paper 2015-01-1615, 2015, doi:  10.4271/2015-01-1615.
454 Jaranson, J., and Ahmed, M., "MMLV: Lightweight Interior Systems Design," SAE Technical Paper 2015-01-
1236,  2015, doi: 10.4271/2015-01-1236.
455 Kearns, J., Park, S., Sabo, J., andMilacic, D., "MMLV: Automatic Transmission Lightweighting," SAE
Technical Paper 2015-01-1240, 2015, doi: 10.4271/2015-01-1240.
456 Maki, C., Byrd, K., McKeough, B., Rentschler, R.  et al., "MMLV: Aluminum Cylinder Block with Bulkhead
Inserts and Aluminum Alloy Connecting Rod," SAE Technical Paper 2015-01-1238, 2015, doi: 10.4271/2015-01-
1238.
457 Plourde, L., Azzouz, M., Wallace, J., and Chellman, M., "MMLV: Door Design and Component Testing," SAE
Technical Paper 2015-01-0409, 2015, doi: 10.4271/2015-01-0409.
458 Schutte, C., "DOE Focuses on Developing Materials to Improve Vehicle Efficiency," SAE Technical Paper
2015-01-0405, 2015, doi: 10.4271/2015-01-0405.
459 Skszek,T., Conklin, J., Wagner, D., Zaluzec.M., "Multi-Material Lightweight Vehicles," Ford Motor Company
and Vehma International, June 11, 2015, presented at 2015 DOE AMR,
http://energy.gov/sites/prod/files/2015/06/f24/lm072_skszek_2015_o.pdf.
460 IBIS Associates, Inc., "Technical Cost Modeling for Vehicle Lightweighting: 40% and 45% Weight Reduction,"
June 11, 2015, http://energy.gov/sites/prod/files/2015/06/f24/lm090_mascarin_2015_o.pdf.
461 Chapter 6 Lightweight Materials within the "Vehicle Technologies Office: 2015 Annual Merit Review Report"
http://energv.gov/sites/prod/files/2015/12/f27/06%20-%20Ligbteeight%20Materials.pdf.

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
462 "Light-Duty Truck Weight Reduction Study with Crash Model, Feasibility and Cost Analysis," T8009-130016,
Department of Transport Canada, September 24, 2015.
463 Brooke, L. "Systems Engineering a new 4x4 benchmark," SAE Automotive Engineering, June 2, 2014
464 Phelps, P., "EcoTrac Disconnecting AWD System," presented at 7th International CTI Symposium North
America 2013, Rochester MI.
465 Pilot Systems, "AWD Component Analysis," Project Report, performed for Transport Canada, Contract T8080-
150132, May 31,2016.
466 Martin, B. et al., "The Innovative driveline of the 9-Speed Jeep Cherokee," presented at 8th International CTI NA
Symposium, May 2014, Rochester, MI.
467 Lee, B., "A Novel Clutch Solution for AWD Disconnect," presented at 9th International CTI Symposium North
America 2015, Rochester, MI.
468 U.S. Environmental Protection Agency, Greenhouse Gas Emission Standards for Light Duty Vehicles,
Manufacturer Performance Report for the 2014 Model Year, EPA-420-R-15-026, December 2015.
469 Sciance, F. et al., "Developing the AC17 Efficiency Test for Mobile Air Conditioners," SAE Technical Paper
2013-01-0569, April 2013.
470 See Greenhouse Gas Emission Standards for Light-Duty Vehicles: Manufacturer Performance Report for the
2014 Model Year, EPA-420-R-15-026, December 2015.
471 Minnesota Pollution Control Agency, Mobile  Air Conditioner Leakage Rates, www.pca.state.mn.us.
472 http://www.greencarcongress.com/2015/10/20151020-mbco2.html.
473 Andersen et al., 2015. "Secondary Loop Motor Vehicle Air Conditioning Systems (SL—MACs). Using Low-
Global Warming Potential (GWP) Refrigerants in Leak-Tight Systems In Climates with High Fuel Prices and Long,
Hot and Humid Cooling Seasons. Building on the Previous Success of Delphi, Fiat, General Motors, Volvo, Red
Dot, SAE Cooperative Research Projects, And Other Engineering Groups." MACS Briefing, 2015.
474 77 FR 62832, October 15, 2012.
475 77 FR 62839, October 15, 2012.
476 40 CFR 86.1869-12.
477 77 FR 62835-62837.
47840CFR86.1869-12(b).
47940CFR86.1869-12(c).
48040CFR86.1869-12(d).
481 For a description of each technology and the derivation of the pre-defined credit levels see Chapter 5 of "Joint
Technical Support Document: Final Rulemaking for 2017-2025 Light-duty Vehicle Greenhouse Gas Emission
Standards and Corporate Average Fuel Economy," EPA-420-R-12-901, August 2012.
482 40 CFR 86.1869-12 (b)(4).
483 U.S. Environmental Protection Agency, Greenhouse Gas Emission Standards for Light Duty Vehicles,
Manufacturer Performance Report for the 2014 Model Year, EPA-420-R-15-026, December 2015. Manufacturers
were required to report MY2015 by March 2016. The data is under EPA review and will be released publicly in the
MY2015 compliance report.
484 "EPA Decision Document: Mercedes-Benz Off-cycle Credits forMYs 2012-2016," U.S. EPA-420-R-14-025,
Office of Transportation and Air Quality,  September 2014.  See http://www.epa.gov/otaq/regs/ld-
hwy/greenhouse/documents/420r!4025.pdf.
485 "EPA Decision Document: Off-cycle Credits for Fiat Chrysler Automobiles, Ford Motor Company, and General
Motors Corporation," U.S. EPA-420-R-15-014, Office of Transportation and Air Quality, September 2015. See
http://www3.epa.gov/otaq/regs/ld-hwy/greenhouse/documents/420rl5014.pdf.
486 National Academies of Science. 2011. Assessment of Fuel Economy Technologies for Light-Duty Vehicles.
National Academies Press, ISBN 978-0-309-15607-3, http://www.nap.edu/catalog.php?record_id=12924, p.25
487 NAS, 2011, p. 62 footnote.
488 National Academies of Science. 2015.  Cost, Effectiveness and Deployment of Fuel Economy
Technologies for Light-Duty Vehicles. National Academies Press, ISBN 978-0-309-37388-3,
                        p.281.
489 NAS, 2011, p. 46.
490 Energy and Environmental Analysis, Inc. 2007.  Update for Advanced Technologies to Improve Fuel Economy of
Light Duty Vehicles. Prepared for U.S. Department of Energy. August. Arlington, Va (EPA-HQ-OAR-2014-0827-
0722).
491 NAS, 2011, p. 62 footnote.

                                                    5-564

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
492 National Academies of Sciences. "Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles," The National Academies Press, 2015 (EPA-HQ-OAR-2015-0827-DRAFT-0398)
493 https://www3 .epa.gov/otaq/climate/mte.htnrfepa-publications.
494 Nelson, P.A. Gallagher, K.G., Bloom, I., and Dees, D.W., "Modeling the Performance and Cost of Lithium-Ion
Batteries for Electric Drive Vehicles," Second Edition, Argonne National Laboratory, ANL-12/55 (December 2012).
495 Gallagher, K., Shabbir, A., Nelson, P., and Dees, D., "PHEV and EV Battery Performance and Cost Assessment,"
Argonne National Laboratory, presented at the 2015 U.S. DOE Vehicle Technologies Office Annual Merit Review
and Peer Evaluation, June 9, 2015.
496 "Learning Curves used in Developing Technology Costs in the Draft TAR," Memorandum from Todd Sherwood
to Air Docket EPA-HQ-OAR_20 15-0827, June 8, 2016.
497 "Learning Curves used in Developing Technology Costs in the Draft TAR," Memorandum from Todd Sherwood
to Air Docket EPA-HQ-OAR_20 15-0827, June 8, 2016.
498 R-pj international, "Automobile Industry Retail Price Equivalent and Indirect Cost Multipliers," February 2009;
EPA-420-R-09-003 ; http://www.epa.gov/otaq/ld-hwy/420r09003 .pdf.
499 Rogozhin, A., et al., "Using indirect cost multipliers to estimate the total cost of adding new technology in the
automobile industry," International Journal of Production Economics (2009), doi:10.1016/j.ijpe.2009. 11.031.
soo Rogozhin, A., et. al., "Using indirect cost multipliers to estimate the total cost of adding new technology in the
automobile industry," International Journal of Production Economics (2009); "Documentation of the Development
of Indirect Cost Multipliers for Three Automotive Technologies," Helfand, G., and Sherwood, T., Memorandum
dated August 2009; "Heavy Duty Truck Retail Price Equivalent and Indirect Cost Multipliers," Draft Report
prepared by RTI International and Transportation Research Institute, University of Michigan, July 2010.
501 Vyas, A., D. Santini and R. Cuenca; "Comparison of indirect Cost Multipliers for Vehicle Manufacturing;" April
2000.
502 "Joint Technical Support Document: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards," EPA-420-R-12-901, August 2012.
503 National Academies of Sciences. "Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles," The National Academies Press, 2015 (EPA-HQ-OAR-2015-0827-DRAFT-0398).
504 See https://www.epa.gov/otaq/climate/mte.htm.
505 See https://www.epa.gov/otaq/climate/data-testing.htm.
506 Newman, K., Kargul, J., and Barba, D., "Benchmarking and Modeling of a  Conventional Mid-Size Car Using
ALPHA," SAE Technical Paper 2015-01-1140, 2015, doi: 10.4271/2015-01-1 140.
so? EPA-HQ-OAR-2015-0827-0899.
508 Newman, K., Kargul, J., and Barba, D., "Benchmarking and Modeling of a Conventional Mid-Size Car Using
ALPHA," SAE Technical Paper 2015-01-1 140, 2015, doi:10.4271/2015-01-1140.
509 See https://www.epa.gov/otaq/climate/alpha.htm.
510 Newman, K., Kargul, J., and Barba, D., "Development and Testing of an Automatic Transmission Shift Schedule
Algorithm for Vehicle Simulation," SAE Int. J. Engines 8(3):1417-1427, 2015, doi: 10.4271/2015-01-1 142.
511 Lee, S., Lee, B., McDonald, J. and Nam, E., "Modeling and Validation of Lithium-Ion Automotive Battery
Packs," SAE Technical Paper 2013-01-1539, 2013, doi:10.4271/2013-01-1539.
512 Lee, S., Cherry, J., Lee, B., McDonald, J., Safoutin, M., "HIL Development and Validation of Lithium Ion
Battery Packs," SAE Technical Paper 2014-01-1863, 2014, doi: 10.4271/2014-01-1863.
513 Newman, K., Kargul, J., and Barba, D., "Development and Testing of an Automatic Transmission Shift Schedule
Algorithm for Vehicle Simulation," SAE Int. J. Engines 8(3):1417-1427, 2015, doi: 10.4271/2015-01-1 142.
514 Newman, K., Doorlag, M., and Barba, D., "Modeling of a Conventional Mid-Size Car with CVT Using ALPHA
and Comparable Powertrain Technologies," SAE Technical Paper 2016-01-1 141, 2016, doi: 10.4271/2016-01-1 141.
515 Heywood, J. B.,1988, Internal Combustion Engine Fundamentals, McGraw Hill, ISBN 007028637, p. 676.
516 EPA Test Car List data files, Wm^^^LSBMMMml^M^Mm. Accessed April 11, 2016.
517 Argonne National Labs Downloadable Dynamometer Database,
                                                                  Accessed April 12, 2016.
518 Newman, K., Kargul, J., and Barba, D., "Benchmarking and Modeling of a Conventional Mid-Size Car Using
ALPHA," SAE Technical Paper 2015-01-1140, 2015, doi: 10.4271/2015-01-1140.
519 EPA-HQ-OAR-20 15-0827-DRAFT-094 1 .
520 Lee, S., Schenk, C., and McDonald, J., "Air Flow Optimization and Calibration in High-Compression-Ratio
Naturally Aspirated SI Engines with Cooled-EGR," SAE Technical Paper 2016-01-0565, 2016, doi: 10.4271/2016-
01-0565.

                                                   5-565

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
521 Ellies, B., Schenk, C., and Dekraker, P., "Benchmarking and Hardware-in-the-Loop Operation of a 2014
MAZDA
SkyActiv 2.0L 13:1 Compression Ratio Engine," SAE Technical Paper 2016-01-1007, 2016, doi:10.4271/2016-01-
1007.
522 Kargul, I, Moskalik, A., Barba, D., Newman, K. et al., "Estimating GHG Reduction from Combinations of
Current Best-Available and Future Powertrain and Vehicle Technologies for a Midsized Car Using EPA's ALPHA
Model," SAE Technical Paper 2016-01-0910, 2016, doi: 10.4271/2016-01-0910.
523 Patton, et al. "Aggregating Technologies for reduced Fuel Consumption: A Review of the Technical Content in
the 2002 National Research Council Report on cafe." SAE 2002-01-0628.  Society of Automotive Engineers, 2002.
524 National Academies of Science. 2015. Cost, Effectiveness and Deployment of Fuel Economy
Technologies for Light-Duty Vehicles. National Academies Press, ISBN 978-0-309-37388-3,
http://www.nap.edu/21744, p.2.
525 Schamel, A.,  Scheidt, M., Weber, C. Faust, H. Is Cylinder Deactivation a Viable Option for a Downsized 3-
Cylinder Engine? Vienna Motor Symposium, 2015.
526 Wilcutts, M., Switkes, I, Shost, M. and Tripathi, A., "Design and Benefits of Dynamic  Skip Fire Strategies for
Cylinder Deactivated Engines," SAE Int. J. Engines 6(1):2013, doi: 10.4271/2013-01-0359.
527 Stuhldreher, M., "Fuel Efficiency Mapping of a 2014 6-Cylinder GM EcoTec 4.3L Engine with Cylinder
Deactivation," SAE Technical Paper 2016-01-0662, 2016, doi: 10.4271/2016-01-0662.
SAE Technical Paper 2016-01-0662, 2016, doi: 10.4271/2016-01-0662.
528 Alger, T., Gingrich, J., Mangold, B., and Roberts, C., "A Continuous Discharge Ignition System for EGR Limit
Extension in SI Engines," SAE Int. J. Engines 4(l):677-692, 2011, doi: 10.4271/2011-01-0661.
529 Alger, T., Mangold, B., Roberts, C., and Gingrich, J., "The Interaction of Fuel Anti-Knock Index and Cooled
EGR on Engine Performance and Efficiency," SAE Int. J. Engines 5(3):1229-1241, 2012, doi: 10.4271/2012-01-
1149.
530 U.S. EPA, "Computer Simulation of Light-Duty Vehicle Technologies for Greenhouse Gas Emission Reduction
in the 2020-2025 Timeframe." Prepared for EPA by Ricardo, Inc. and Systems Research and Applications
Corporation under EPA Contract No. EP-C-11-007/Work Assignment No. 0-12. EPA Report No. 420-R-l 1-020,
December 2011.
531 Cruff, L., Kaiser, M., Krause, S., Harris, R. et al., "EBDI® - Application of a Fully Flexible High BMEP
Downsized Spark Ignited Engine," SAE Technical Paper 2010-01-0587, 2010, doi: 10.4271/2010-01-0587.
532 Kaiser, M., Krueger, U., Harris, R., and Cruff, L., ""Doing More with Less" - The Fuel Economy Benefits of
Cooled EGR on  a Direct Injected Spark Ignited Boosted Engine," SAE Technical Paper 2010-01-0589,  2010,
doi:10.4271/2010-01-0589.
533 King, J., Heaney, M. Saward, J., Fraser, A., Criddle, M. Cheng, T., Morris, G., Bloore, P. "HyBoost: An
Intelligently Electrified Optimised Downsized Gasoline Engine Concep.t' SAE-China and  FISITA (eds.),
Proceedings of the FISITA 2012 World Automotive Congress, Lecture Notes in Electrical Engineering  191,
DOI: 10.1007/978-3-642-33777-2_39.
534 King, J., Bocker, O. "Multiple Injection and Boosting Benefits for Improved Fuel Consumption on a Spray
Guided Direct Injection Gasoline Engine." SAE-China and FISITA (eds.), Proceedings of the FISITA 2012 World
Automotive Congress, Lecture Notes in Electrical Engineering 189, DOI: 10.1007/978-3-642-33841-0_18.
535 Wada, Y., Nakano, K., Mochizuki, K., and Hata, R., "Development of a New 1.5L 14 Turbocharged Gasoline
Direct Injection Engine,"  SAE Technical Paper 2016-01-1020, 2016, doi:10.4271/2016-01-1020.
536 Nakano, K., Wada, Y., Jono, M., Narihiro, S. "New In-Line 4-Cylinder Gasoline Direct Injection Turbocharged
Downsizing Engine." Honda R&D Technical Review, April 2016, pp 139-146.
537 Budack, R., Kuhn, M., Wurms, R., Heiduk, T., "Optimization of the Combustion Process as Demonstrated on the
New Audi 2.01 TFSI," 24th Aachen Colloquium Automobile and Engine Technology 2015.
538 Wurms, R., Budack, R., Grigo, M., Mendl, G., Heiduk, T., Knirsch, S. "The new Audi 2.01 Engine with
innovative Rightsizing," 36. Internationales Wiener Motorensymposium 2015.
539 Eichler, F., Demmelbauer-Ebner, W., Theobald, J., Stiebels, B., Hoffmeyer, H., Kreft, M. "The New EA211
TSI® evo from Volkswagen." 37. Internationales Wiener Motorensymposium 2016.
540 "National Academies of Science. "2015. Cost, Effectiveness and Deployment of Fuel Economy Technologies
for Light-Duty Vehicles." ISBN 978-0-309-37388-3 p. 5-52 (Finding 5.3)
http://www.nap.edu/catalog.php?record_id=21744.
541 Newman, K., Kargul, J., and Barba, D., "Benchmarking and Modeling of a Conventional Mid-Size Car Using
ALPHA," SAE Technical Paper 2015-01-1140, 2015, doi: 10.4271/2015-01-1140.

                                                    5-566

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
542 Moskalik, A., Hula, A., Barba, D., and Kargul, I, "Investigating the Effect of Advanced Automatic
Transmissions on Fuel Consumption Using Vehicle Testing and Modeling," SAE Technical Paper 2016-01-1142,
2016.
543 "Light-duty Vehicle Technology Cost Analysis, Advanced 8-Speed Transmissions (Revised Final Report)," EPA-
420-R-13-007.
544 Lee, S., Lee, B., McDonald, I, Sanchez, L., et al, "Modeling and Validation of Power-Split and P2 Parallel
Hybrid Electric Vehicles," SAE Technical Paper 2013-01-1470, doi:10:4271/2013-01-1470
545 California Air Resources Board, "Request for Proposal 15CAR018, Advanced Strong Hybrid and Plug-In Hybrid
Engineering Evaluation and Cost Analysis," May 9, 2016.
546 Nelson, P. A.  Gallagher, K.G., Bloom, I., and Dees, D.W., "Modeling the Performance and Cost of Lithium-Ion
Batteries for Electric Drive Vehicles," Second Edition, Argonne National Laboratory, ANL-12/55 (December 2012).
547 See EPA Docket EPA-HQ-OAR-2015-0827, Microsoft Excel attachment to Docket Item titled "Data and Charts
for Selected Figures in Draft TAR Section 5.2 and 5.3."
548 See 40 CFR 600.116-12(a)(6), 40 CFR 600.210-12(d)(3), and 76 FR 39478.
549 See http://www.uscar.0rg/guest/partnership/l/us-drive.
550 US DRIVE Partnership, "US DRIVE Electrical and Electronics Technical Team Roadmap," June  2013.
downloaded on Dec. 23, 2015 from
http://wwwl.eere.energy.gov/vehiclesandfuels/pdfs/program/eett_roadmap June2013.pdf.
551 Slenzak, I, "Next Generation Electrification Products: Focus on Integration and Cost Reduction," Bosch,The
Battery Show 2015, Novi MI, September  15, 2015.
552 cf. U.S. DRIVE Electrical and Electronics Technical Team Roadmap, p 10.
553 See http://www.A2Macl.com/.
554 General Motors Press Release, "Drive Unit and Battery at the Heart of Chevrolet Bolt EV," January  11, 2016,
downloaded on Jan. 21, 2016 from
http://media.chevrolet.com/media/us/en/chevrolet/home.detail.html/content/Pages/news/us/en/2016/Jan/naias/chevy/
0111-bolt-du.html.
555 Anderman, M., "Battery Packs of Modern xEVs: A Comprehensive Engineering Assessment: Extract," Total
Battery Consulting, 2016, slides  19 and 30.
556 Malliaris, A.C., Hsia, H., and Gould, H.; "Concise description of auto fuel economy and performance in recent
model years," Society of Automotive Engineers,  Paper 760045 (1976).
557 U.S.  EPA, Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975
Through 2015, Report EPA-420-R-15-016, December 2015, p. 35.
558 MacKenzie, D., Hey wood, J.  2012. Acceleration performance trends and the evolving relationship among
power, weight, and acceleration in U.S. Light-duty vehicles: A linear regression analysis. Transportation Research
Board, Paper No. 12-1475, TRB 91st Annual Meeting, Washington, DC, January 2012.
559 Cole, J., "Exclusive: GM Exec Says Spark EV's 400 Ib-ft. of Torque No Misprint," May 2, 2013.  At
http://insideevs.com/gm-general-says-spark-evs-4001b-ft-of-torque-no-misprint/.
560 Nelson, P. A., Santini, D.J., Barnes, J. "Factors Determining the Manufacturing Costs of Lithium-Ion Batteries
for PHEVs," 24th World Battery, Hybrid and Fuel Cell Electric Vehicle Symposium and Exposition EVS-24,
Stavenger, Norway, May  13-16, 2009 (www.evs24.org).
561 Santini, D.J., Gallagher, K.G., and Nelson, P. A. "Modeling of Manufacturing Costs of Lithium-Ion Batteries for
HEVs, PHEVs, and EVs," Paper to be presented at the 25th World Battery, Hybrid and Fuel Cell Electric Vehicle
Symposium and Exposition, EVS-25, Shenzhen, China, November 5-9, 2010 (www.evs25.org). Advance draft
provided by D.J.  Santini, Argonne National Laboratory,  August 24, 2010.
562 See Docket itemEPA-HQ-OAR-2010-0799-11913.
563ICF International, Peer Review Report of the Draft Report "Modeling the Cost and Performance of Lithium-Ion
Batteries for Electric-Drive Vehicles," March 31, 2011. See Docket itemEPA-HQ-OAR-2010-0799-1080.
564 See Docket items EPA-HQ-OAR-2010-0799-1078 and EPA-HQ-OAR-2010-0799-11914.
ses Argonne National Laboratory, "Changes to BatPaC for Version 3.0," available at
http://www.cse.anl.gov/batpac/files/Changes%20to%20BatPaC%20for%20Version%203%2021Dec2015.pdf.
see Argonne National Laboratory, "BatPaC (Battery Performance and Cost)," BatPaC Version 3.0 Beta
17Dec2015.xlsx.  A copy is available in Docket EPA-HQ-OAR-2015-0827.
567 Van Bellinghen, T., "Next Gen Battery Chemistry: Is High Nickel Really the Best Option?" The Battery Show
2015, Novi, MI.
568 Wise, R., BASF, "Designing Cathode Materials for Next Generation Electric Vehicles," The Battery Show 2015,
Novi, MI, September 15, 2015.
                                                    J-JO/

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
569 Anderman, M, "Battery Packs of Modern xEVs: A Comprehensive Engineering Assessment: Extract," Total
Battery Consulting, 2016, slide 52.
570 FEV, "Light Duty Technology Cost Analysis, Power-Split and P2 Hybrid Electric Vehicle Case Studies," Report
FEV07-069-303, October 10, 2011.,"
571 "Identifying Real World Barriers to Implementing Lightweighting Technologies and Challenges in Estimating
the Increase in Costs," Center for Automotive Research, January 2016.
http://www.cargroup.org/?module=Publications&event=View&pubID=128.
572 "Light-Duty Vehicle Mass Reduction and Cost analysis - Midsize Crossover Utility Vehicle," FEV, EPA-420-R-
12-026, August 2012.
573 "Honda's Study & Report on the Study Commissioned by NHTSA "Mass Reduction for Light-Duty Vehicles for
Model Years 2017-2025," Honda, DTNH22-11-C-00193, NHTSA Mass Size Safety Workshop May 13-14, 2013.
574 "Update to Future Midsize Lightweight Vehicle Findings in Response to Manufacturer Review and IIHS Small-
Overlap Testing," NHTSA, DOT HS 812 237, February 2016.
575 "Evaluating the Structure and Crashworthiness of a 2020 Model-Year, Mass Reduced Crossover Vehicle Using
FEA Modeling," Lotus Engineering Inc. for ARB, August 31, 2012.
576 IBIS Associates, Inc., "Vehicle Lightweighting: Mass Reduction Spectrum Analysis and Process Cost Modeling"
Project ID #LM090, Presented at 2016 DOE AMR, June 7,  2016.
577General Motors, "General Motors 2015 Global Business Conference," Presentation, October 1, 2015, Slides 43-45
in document, https://www.gm.com/content/dam/gm/events/docs/5194074-596155-ChartSet-10-l-2015.
578 Conklin, J., Beals, R., and Brown, Z., "BIW Design and  CAE," SAE Technical paper 2015-01-0408, 2015,
doi:10.4271/2015-01-0408.
579 "Evaluating the Structure and Crashworthiness of a 2020 Model-Year, Mass-Reduced Crossover Vehicle Using
FEA Modeling," Lotus Engineering for California Air Resources Board, August 31, 2012.
https://www3.epa.gov/otaq/climate/documents/final-arb-phase2-rpt-rl.pdf.
580 "Venza Aluminum BIW Concept Study," April 2013, http://www.drivealuminum.org/research-
resources/PDF/Research/2013/venza-biw-full-study.
581 Source: Kelkar et al: "Automobile Bodies: Can Aluminum Be an Economical Alternative to Steel?," August 2001
Issue of JOM., 53 (8) (2001)pp.  28-32, http://www.tms.org/pubs/journals/JOM/0108/Kelkar-0108.html.
582 "Chevrolet Silverado," Wikipedia, https://en.wikipedia.org/wiki/Chevrolet_Silverado.
583 "Mass Reduction and Cost Analysis - Light Duty Pickup Truck Model Years 2020-2025," EPA  420rl5006, FEV
North America for EPA, June 8, 2015.
584 "Technical Analysis of Vehicle Load Reduction Potential for Advanced Clean Cars," Controltec, LLC, for
CARB, April 29, 2015.
585 "AWD Component Analysis," Pilot Systems for Transport Canada, 2016.
586 "Corporate Average Fuel Economy for MY2017-MY2025 Passenger Cars and Light Trucks," U.S. DOT
NHTSA,  August 2012, http://www.nhtsa.gov/staticfiles/rulemaking/pdf/cafe/FRIA 2017-2025.pdf.
587 "Update to Future Midsize Lightweight Vehicle Findings in Response to Manufacturer Review and IIHS Small-
Overlap Testing," NHTSA, DOT HS 812 237, February 2016,
http://www.nhtsa.gov/staticfiles/rulemaking/pdf/cafe/812237_LightWeightVehicleReport.pdf.
588 "Light-Duty Truck Weight Reduction Study with Crash Model, Feasibility and Cost Analysis," T8009-130016,
Department of Transport Canada, September 24, 2015.
589 National Academies of Sciences. "Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles," The National Academies Press, 2015, ISBN 978-0-309-37388-3.
590 Heywood, J., MacKenzie, D., Bonde Akerlind, I., Bastani, P., Berry, I., Bhatt, K., Chao, A., Chow, E., Karplus,
V., Ketih, D., Khusid, M., Nishimura, E., Zoepf, S., "On the Road toward 2050: Potential for Substantial Reductions
in Light-Duty Vehicle Energy Use and Greenhouse Gas Emissions," November 2015.
591 FEV, "Light-Duty Technology Cost Analysis, Report on Additional Case Studies," Report FEV 07-069-203 Rev.
D, Contract No. EPA-420-R-13-008, April 2013.
592 Model Years 2012-2016: Final Rule and Final TSD, http://www.nhtsa.gov/Laws+&+Regulations/CAFE+-
+Fuel+Economy/Model+Years+2012-2016:+Final+Rule.
593 FEV, "Light-Duty Vehicle Technology Cost Analysis, Advanced 8-Speed Transmissions Revised Final Report,"
Report FEV 07-069-303, Contract No. EP-C-07-069, WA 3-3 (April 2013).
594 SAE, "Vehicle System Voltage Initial Recommendations." http://standards.sae.org/i2232 199906/
595 FEV, "Light-Duty Vehicle Technology Cost Analysis - European Vehicle Market, Additional Case Studies
(Phase 2)."
http://www.theicct.org/sites/default/files/FEV LDV%20EU%2QTechnologv%20Cost%20Analysis Phase2.pdf

-------
                                   Technology Cost, Effectiveness, and Lead-Time Assessment
596(Ref:DOTHS811666).
597 (Ref: http://www.nhtsa.gov/staticfiles/rulemaking/pdf/MSS/4-Thomas-Honda Report.pdf).
598 R-pj international, "Automobile Industry Retail Price Equivalent and Indirect Cost Multipliers," February 2009;
EPA-420-R-09-003; http://www.epa.gov/otaq/ld-hwy/420r09003.pdf.
599 Rogozhin, A., et al., "Using indirect cost multipliers to estimate the total cost of adding new technology in the
automobile industry," International Journal of Production Economics (2009), doi:10.1016/j.ijpe.2009.11.031.
600 80 FR 40137.
601 Rogozhin, A., et. al., "Using indirect cost multipliers to estimate the total cost of adding new technology in the
automobile industry," International Journal of Production Economics (2009); "Documentation of the Development
of Indirect Cost Multipliers for Three Automotive Technologies," Helfand, G., and Sherwood, T., Memorandum
dated August 2009; "Heavy Duty Truck Retail Price Equivalent and Indirect Cost Multipliers," Draft Report
prepared by RTI International and Transportation Research Institute, University of Michigan, July 2010.
602 [Website] http://www.nhtsa.gov/Laws+&+Regulations/C AFE+-
+Fuel+Economv/CAFE+Compliance+and+Effects+Modeling+Svstem:+The+Volpe+Model
603 [Website] https://www.volpe.dot.gov/policv-planning-environment/corporate-average-fuel-economy/increasing-
fuel-economy-vehicle-fleets.
604 S. Halbach, P. Sharer, P. Pagerit, C. Folkerts, A. Rousseau, "Model Architecture, Methods, and Interfaces for
Efficient Math-Based design and Simulation of Automotive Control Systems," SAE 2010-01-0241, SAE World
Congress, Detroit, April, 2010.
605 N. Kim, J. Jeong, A. Rousseau,  and H. Lohse-Busch, "Control Analysis and Thermal Model Development of
PHEV," SAE 2015-01-1157,  SAE  World Congress, Detroit, AprillS; N. Kim, A. Rousseau, and H. Lohse-Busch,
"Advanced Automatic Transmission Model Validation Using Dynamometer Test Data," SAE 2014-01-1778, SAE
World Congress, Detroit, Aprl4.
; D. Lee, A. Rousseau, E. Rask, "Development and Validation of the Ford Focus BEV Vehicle Model," 2014-01-
1809, SAE World Congress, Detroit, Aprl4; N. Kim, N. Kim, A. Rousseau, M. Duoba, "Validating Volt PHEV
Model with Dynamometer Test Data using Autonomie," SAE 2013-01-1458, SAE World Congress, Detroit, Aprl3.
; N. Kim, A. Rousseau, E. Rask, "Autonomie Model Validation with Test Data for 2010 Toyota Prius," SAE 2012-
01-1040, SAE World Congress, Detroit, Aprl2; Karbowski, D., Rousseau, A, Pagerit, S., Sharer, P., "Plug-in
Vehicle Control Strategy: From Global Optimization to Real Time Application," 22th International Electric Vehicle
Symposium (EVS22), Yokohama,  (October 2006).
606 Karbowski, D., Kwon, J., Kim,N., Rousseau, A., "Instantaneously Optimized Controller for a Multimode Hybrid
Electric Vehicle," SAE paper 2010-01-0816,  SAE World Congress, Detroit, April 2010.
eov p ]\feison K. Amine, A. Rousseau, H. Yomoto (EnerDel Corp.), "Advanced lithium-ion batteries for plug-in
hybrid-electric vehicles," 23rd International Electric Vehicle Symposium (EVS23), Anaheim, CA, (Dec. 2007); D.
Karbowski, C. Haliburton,  A. Rousseau, "Impact of component size on plug-in hybrid vehicles energy consumption
using global optimization," 23rd International Electric  Vehicle Symposium (EVS23), Anaheim, CA, (Dec. 2007).
608 Karbowski, D., Kwon, J., Kim,N., Rousseau, A., "Instantaneously Optimized Controller for a Multimode Hybrid
Electric Vehicle," SAE paper 2010-01-0816,  SAE World Congress, Detroit, April 2010, P. Sharer, A. Rousseau, D.
Karbowski, S. Pagerit, "Plug-in Hybrid Electric Vehicle Control Strategy: Comparison between EV and Charge-
Depleting Options," SAE paper 2008-01-0460, SAE World Congress, Detroit (April 2008), and A. Rousseau, N.
Shidore, R. Carlson, D. Karbowski, "Impact of Battery Characteristics on PHEV Fuel Economy," AABC08.
609 Delorme et al. 2008, Rousseau,  A, Sharer, P, Pagerit, S., Das, S., "Trade-off between Fuel Economy and Cost for
Advanced Vehicle Configurations" 20th International Electric Vehicle Symposium (EVS20), Monaco (April 2005),
Amgad Elgowainy, Andrew Burnham, Michael Wang, John Molburg, and Aymeric Rousseau, "Well-To-Wheels
Energy Use and Greenhouse Gas Emissions of Plug-in Hybrid Electric Vehicles" SAE 2009-01-1309, SAE World
Congress, Detroit, April 2009.
610 Vijayagopal, R., Kwon, J., Rousseau, A., Maloney,  P., "Maximizing Net Present Value of a  Series PHEV by
Optimizing Battery Size and Vehicle Control Parameters" SAE 2010-01-2310, SAE Convergence Conference,
Detroit (October 2010).
611 www.autonomie.net.
612 A list of the vehicles that have been tested at the APRF can be found under http://www.anl.gov/energy-
svstems/group/downloadable-dynamometer-database. http://www.anl.gov/energv-svstems/group/downloadable-
dynamometer-database.
613 For more detailed information on instrumentation and facility capabilities, please refer to the "chassis
Dynamometer Testing Reference Document" (https://anl.app.box.com/s/5tlld40tjhhhtoj2tgOn4y3fkwdbs4m3).
614 Available publicly at www.anl.gov/D3.

-------
                                 Technology Cost, Effectiveness, and Lead-Time Assessment
615 EPA Trends Report, "Light Duty Automotive Technology, Carbon Dioxide Emissions and Fuel Economy
Trends; 1975 Through 2013," EPA-402-R-13-011, December, 2013.
616 Kim, N. Kwon, J., Rousseau, A. Trade-off between Multi-mode Powertrain Complexity and Fuel Consumption.
EVS-25 Shenzhen, China, Nov. 5-9, 2010. Available at: lMmUM^M^^MMMMAoc^&%3^
                                                5-570

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                        GHG Emissions

Table of Contents

Chapter 6:   Assessment of Consumer Acceptance of Technologies that Reduce Fuel
Consumption and GHG Emissions	6-1
  6.1    Introduction	6-1
  6.2    Effects of the Standards on Vehicle Sales	6-1
     6.2.1   Overview of Vehicle Market	6-1
     6.2.2   Consumer Vehicle Choice Modeling and Recent Research	6-2
       6.2.2.1  EPA's Efforts in Developing and Assessing a Consumer Vehicle Choice Model
               6-3
  6.3    Conceptual Framework for Evaluating Consumer Impacts	6-5
  6.4    Consumer Response to Vehicles Subject to the Standards	6-9
     6.4.1   Recent New Vehicles	6-9
       6.4.1.1  Sales	6-9
       6.4.1.2  Evaluations from Professional Auto Reviewers	6-10
       6.4.1.3  Consumer Responses to New Vehicles	6-12
     6.4.2   MY2022-25 Vehicles	6-13
  6.5    Impacts of the Standards on Vehicle Affordability	6-16
     6.5.1   Effects on Lower-Income Households	6-16
     6.5.2   Effects on the Used Vehicle Market	6-16
     6.5.3   Effects on Access to Credit	6-19
     6.5.4   Effects on Low-Priced Cars	6-20
     6.5.5   Conclusion	6-22

Table of Figures
Figure 6.1 Gross Domestic Product Per Capita and Vehicle Production, 2005-2015	6-2
Figure 6.2 Used and New Car Consumer Price Index, 2013=100 (2013$)	6-18
Figure 6.3 Number of <$ 15,000 Car Models Available, from Ward's Automotive Data	6-21
Figure 6.4 Minimum MSRP of All Car Models Available, from Ward's Automotive Data	6-22


Table of Tables
Table 6.1 Efficiency Technology's Positive, Negative, or Neutral Evaluations by Auto Reviews	6-11

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                         GHG Emissions

Chapter 6:  Assessment of Consumer Acceptance of Technologies that Reduce
Fuel Consumption and GHG Emissions

6.1    Introduction

   As part of the midterm evaluation, the agencies committed that, in this Draft TAR, they would
examine "Costs, availability, and consumer acceptance of technologies to ensure compliance
with the standards, such as vehicle batteries and power electronics, mass reduction, and
anticipated trends in these costs."1  Technologies and costs are examined in Chapter 5 of this
document; this chapter reviews consumer acceptance of the technologies being used to meet the
standards. With the program in effect since MY2012, this chapter focuses on the evidence to
date on consumer acceptance of vehicles subject to the standards.

   Chapter 6.2 discusses one potential measure of consumer acceptance, the effects of the
standards on vehicle sales; as discussed there, it is difficult, if not impossible, to disentangle the
effects of the standards on vehicle sales from the effects of macroeconomic or other conditions
on sales. Chapter 6.3 discusses possible reasons why fuel efficient technologies may not be
adopted absent the standards, in spite of the observation that fuel savings outweigh upfront costs.
Chapter 6.4 discusses preliminary results of an EPA-led analysis of how professional auto
reviewers assess the GHG-reducing technologies; in general, the reviews are positive. Finally,
Chapter 6.5 reviews evidence related to the effects of the standards on the affordability of new
and used vehicles, and suggests the difficulty of identifying and measuring such effects.

6.2    Effects of the Standards on Vehicle Sales

6.2.1  Overview of Vehicle Market

   Chapter 3 examines trends in the light-duty vehicle market since the National Program
standards went into effect in MY2012.A As that chapter shows, vehicle sales have been close to
record levels. At the same time that GHG emissions have been  dropping, vehicle footprint has
increased slightly, horsepower has increased, and weight has been roughly constant. The
projections for the car/truck mix used in the 2017-25 rulemaking are close to those being realized
through MY2014 (see Chapter  3.1.4).

   It is difficult, if not impossible, to separate the effects of the  standards on vehicle sales and
other characteristics from the impacts of macroeconomic or other forces on the auto market.
Figure 6.1 graphs light-duty vehicle production6 and gross domestic product (GDP) per capita
from 2005-2015.2 As this figure shows, production in the auto  industry has had a pattern similar
to GDP per capita: production fell with the reduction in economic activity in the 2009 recession,
and has increased as the economy has recovered. The American Automotive Policy Council, in
citing this recovery, notes that "U.S. auto sales increased by double digits from 2010 to 2014,
even though GDP has grown by less than 3 percent each year;"3 it projects sales to reach or
A Note that California's GHG standards began with MY2009 and includes a "deemed to comply" provision with the
  National Program for MY2012 and subsequent, see Section 1.2.3 for further background.
B Vehicle production data represent production volumes delivered for sale in the U.S. market, rather than actual
  sales data. They include vehicles built overseas imported for sale in the U.S., and exclude vehicles built in the
  U.S. for export.
                                              6-1

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                          GHG Emissions
exceed 17 million vehicles each year through 2016, and domestic production to go from 5.8
million vehicles in 2009 to 11.5 million or more vehicles through 2016.  A number of other
factors are also likely to affect new vehicle production and sales, including fuel prices,
demographic factors, and vehicle characteristics including but not limited to fuel economy.
                                                                      55000
                                                                      54000
                                                                      50000
                                                                      49000
                    2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

                          	LDV Production   	Real GDP/Capita, 2013$
           Figure 6.1  Gross Domestic Product Per Capita and Vehicle Production, 2005-2015
           Note: Gross Domestic Product per Capita data are from U.S. Bureau of Economic Analysis,
           Account Code A939RX (Real gross domestic product per capita); LDV production from U.S. EPA
           2015.4 2015 production data are projected, not actual, values.

   The National Program light-duty vehicle standards, which went into effect in MY2012, are
likely to have had some effect on vehicle sales.  We have not identified, however, any sound way
to separately  estimate the effect of the standards on sales. The most solid analysis would involve
the ability to  compare sales in a place not affected by the standards, with sales in a place
identical to the first during the same time period, except where the standards are in effect.
Because the standards are national in scope,  such a comparison is not possible. Alternatively, it
may be possible to examine how sales have changed as the standards have tightened, but it
would be necessary to control for all other factors, such as macroeconomic conditions, that affect
sales. Perhaps all that can be concluded about the effects of the standards on vehicle sales is that
they have clearly not prevented the automobile market from recovering to pre-recession sales
levels (indeed, to record sales levels) through 2015.

  6.2.2   Consumer Vehicle Choice Modeling and Recent Research

   In addition to their effect on overall sales and production, the standards could affect the mix
of vehicles sold.  Consumer vehicle choice models estimate what vehicles consumers buy based
on vehicle and consumer characteristics.  In principle, such models could provide a means of
examining the effects of the standards on both overall vehicle sales and the mix of vehicles sold.
Because the standards are based on the footprints of vehicles, shifts in the mix of vehicles sold
do not necessarily affect automakers' ability to meet the standards, but they could affect total
GHGs emitted. Whitefoot and Skerlos (2012), for example, use a vehicle choice model combined
                                               6-2

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                         GHG Emissions

with producer cost estimates to argue that the footprint-based standard provides some incentive
for automakers to increase the size of vehicles in order to face a less stringent standard, and
higher GHG emissions.5'0 As discussed in Chapter 3, the average footprint of vehicles has
increased slightly since the standards have been implemented. As with sales, this effect is
potentially confounded by a number of factors, such as previous trends, dropping gasoline prices
and increasing consumer income that changes the mix of vehicles purchased.

   In the 2017-25 LDV GHGRIA (Chapter 8.1.2), EPA provided an extensive discussion of
consumer vehicle choice modeling as a way to estimate the effects of GHG/fuel economy
standards on vehicle purchase decisions.6 In that discussion, EPA found that,  despite an
extensive literature of consumer choice models, few researchers have compared estimates of key
model parameters with those of others' models, and there have been  few efforts to test the
forecasting ability of those models. As a start to addressing this gap in the literature, EPA had
commissioned a study of the findings of these models on the role of fuel economy in consumer
vehicle purchases and found highly varied results.7 At the time, EPA concluded that the science
of these models was not adequately developed for use in policy-making.

   Two recent papers have done some work on the predictive abilities of consumer choice
models. Haaf et al. (2014) use data from MY2004-6 vehicles to estimate a number of different
econometric models, and test their predictions against MY2007 and  2010 vehicle sales.8  They
conclude that "the models we construct are fairly poor predictors of future shares." They find
that a "static" model assuming constant market shares - that is, using current-year market shares
rather than a model —  outperformed their estimated models for MY2007, while some attribute-
based models predicted better for MY2010. Raynaert (2014) developed a structural model of
vehicle supply and demand in Europe, using data from 1998-2007; he then compared red sales-
weighted aggregate predictions from the model for MY2011 to actual outcomes.9 He finds close
agreement on aggregate market outcomes: in a period where actual emissions  dropped 14
percent, his estimates for emissions differed from the observed values by 2.3 percent. Weight,
footprint, and the share of diesel also had discrepancies of 3 percent or less; price/income and
horsepower differed by under 10 percent.  He implies, without detailed information, that the
model nevertheless does not predict market shares or total sales very well. These papers leave
questions unanswered about the ability of consumer vehicle choice models to predict sales and
fleet mix.

    6.2.2.1    EPA's Efforts in Developing and Assessing a Consumer Vehicle Choice Model

   As part of its exploration of vehicle choice modeling, EPA commissioned the development of
a vehicle choice model from David Greene and Changzheng Liu of Oak Ridge National
Laboratory (Greene and Liu 2012).10  This model, described in the 2017-2025 RIA (Chapter
8.1.2.8), is designed with a straightforward purpose: to estimate, for a predetermined fleet (the
reference fleet, described in Chapter 4), the effects of changes in only fuel economy and price on
c While the agencies consider the concept of the Whitefoot and Skerlos analysis to have some potential merits, it is
  also important to note that, among other things, the authors assumed different inputs than the agencies actually
  used in the MYs 2012-2016 rule regarding the baseline fleet, the cost and efficacy of potential future
  technologies, and the relationship between vehicle footprint and fuel economy. Changes in any of the underlying
  assumptions is likely to lead to different analytical results, and possibly different implications for agency action.
                                               6-3

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                           GHG Emissions
vehicle sales and class mix.  The model calculates a sales response to a change in the "effective
price" for each vehicle, where the effective price combines any change in up-front cost with a
portion of the future fuel savings (see Greene and Liu 2012 for details).  That portion of future
fuel savings depends on user inputs for factors including the price of fuel, the number of years of
fuel savings that a buyer considers (the payback period), and the discount rate.  It is intended for
use in policy analyses of vehicle GHG/fuel economy regulations, and not to predict changes in
the vehicle market associated with macroeconomic shifts or changes in demographic factors.  As
part of our ongoing study of vehicle choice models, EPA has put the model through a variety of
tests intended to understand it better.11
   One group of tests involved examining the sensitivity of the model  to changes in parameters,
including the role of fuel economy in consumer purchase decisions,  the discount rate, model
elasticities, and the initial vehicle fleet.

   •   First, we examined the effects of a 20 percent improvement in  fuel economy0 for all
       vehicles; in response, total sales increased about 5 percent, with higher  sales increases
       going for some of the larger, less fuel-efficient vehicles. If poor fuel  efficiency would
       otherwise reduce the interest of buyers in those vehicles, then improving their fuel
       economy may disproportionately improve their sales.
   •   Next, we varied the payback period - the number of years of fuel savings that a vehicle
       buyer might consider in the purchase decision - from 1 to 7 years.  Total sales increased
       by less than 1 percent for every additional year of payback period, suggesting that
       modeling results are not highly  sensitive to this parameter.
   •   Similarly, varying the discount rate (used to calculate the value of future fuel savings)
       from 2 to 10 percent changed total sales by less than 1 percent, suggesting insensitivity to
       this parameter as well.
   •   When demand elasticities (percent change in sales in response  to a one  percent change in
       effective price) for all classes in the model are increased by 50 percent, total sales
       increase 7 percent, compared to 5 percent in the baseline case;  if the elasticity of only one
       class is changed, total sales are virtually unaffected, though sales in the class that had the
       elasticity change increased by about 5 percent.
   •   Finally, we experimented with increasing the number of vehicles in the initial fleet by 50
       percent (both uniformly for all vehicles and for one vehicle class at a time), to test
       sensitivity to assumptions about that baseline fleet. The sales response  with a larger fleet
       to the 20 percent change in fuel economy was approximately proportional: just as sales in
       the initial case increased 4.9 percent in response to the changes in fuel economy, sales
       with the larger fleet increased 4.9 percent. Changing the size of individual classes also
       had very little effect on market shares, because they all increased proportionally.
D In the model, sales change in response to an effective price that combines the up-front cost with a share of future
  fuel savings. Increasing fuel economy thus has the opposite effect of increasing price; the former reduces the
  effective price, while the latter increases it. We used the 20 percent increase in fuel economy as a fairly large
  change, especially because it is not offset by any price increase.
                                               6-4

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                        GHG Emissions

   In sum, these tests showed that the results of the model are not highly sensitive to any of these
parameters.  Thus, imprecision in the initial fleet or these other factors is not likely to have a
major effect on the model's predictions.  It also suggests that the results of changing fuel
economy and price in the model may not have large effects on the vehicle fleet. Of course, this
series of tests does not provide insight into whether its predictions are accurate.

   A second exercise examined the model's ability to predict sales. It should be noted that the
model is not intended to predict future sales or fleet mix. To do so would require inclusion of
factors such as macroeconomic conditions and demographic shifts that affect sales; EPA's model
was not designed to include those factors.  As noted above, the model is intended to take as a
given the without-standards fleet, and to estimate the effects of changes in price and fuel
economy on sales and class  shifts,  as a way of focusing specifically on the effects of GHG policy
on the fleet.  For that reason, testing the model by using it to predict sales in a different year is
asking more of the model than the  purposes for which it was intended. We conducted this test,
nevertheless, as an initial attempt to test whether the model's results reflect actual consumer
behavior.

   In this test, we calibrated the model to MY2008 vehicle sales, calculated the difference in
vehicles' fuel economy and  price between MY2008 and MY2010 (another year for which we
had the specific vehicle data needed for this analysis), used the model to estimate responses to
the changes in MY2010 fuel economy and price, and compared the MY2010 predictions to
actual MY2010 sales.  The model did not predict sales or market shares well. The model
predicted an increase in total sales  when actual sales decreased. For market shares, similar to the
near-term results in Haaf et al. (2014), using actual market shares from MY2008 - i.e.,  not using
a model - had better predictions than using the model. These poor predictions are not surprising,
given that MY2010 sales reflect the Great Recession,  a significant factor that the model was not
designed to address. We do not consider these results a demonstration that the model does not
perform well; rather, it indicates the difficulty of testing the predictive abilities of this model as it
is designed.

   At this point, then, EPA does not plan to use this or another vehicle choice model in its
current modeling work. We encourage further research in the validation of these consumer
choice models for policy analysis.

6.3    Conceptual Framework for Evaluating Consumer Impacts

   As discussed in Chapter 12, the agencies estimate that fuel-saving technologies, in addition to
reducing GHG emissions and improving energy security, pay for themselves within a few-year
payback period, and thus save consumers money.  Despite this, development and uptake of
energy efficiency technologies lags behind adoption that might be expected under these
circumstances.  The implication is  that private markets do not provide all the cost-effective
energy-saving technologies identified by engineering analysis. The phenomenon is documented
in many analyses of energy efficiency, and is termed the "energy paradox" or "energy efficiency
gap."12 A number of hypotheses have been raised for the existence of this gap,13 as discussed in
the 2017-25 LD GHG rulemaking. Some arise from market failures,  such as lack of perfect
information. Others point to behaviors on the part of consumers and/or firms that appear not to
be in their own best interest (behavioral anomalies).  Still others point to potential costs of the
                                              6-5

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                        GHG Emissions

standards that are not reflected in EPA analyses.  On the consumer side, these hypotheses
include:

       •  Consumers might lack the information necessary to estimate the value of future fuel
          savings, not have a full understanding of this information even when it is presented,
          or not trust the presented information

       •  Consumers might be "myopic" and hence undervalue future fuel savings in their
          purchasing decisions

       •  Consumers may be accounting for uncertainty in future fuel savings when comparing
          upfront cost to future returns

       •  Consumers may consider fuel economy after other vehicle attributes and, as such, not
          optimize the level of this attribute (instead "satisficing" - that is, selecting a vehicle
          that is acceptable rather than optimal — or selecting vehicles that have some sufficient
          amount of fuel economy)

       •  Consumers might be especially averse to the short-term losses associated with the
          higher prices of energy efficient products relative to the long-term gains of future fuel
          savings (the behavioral phenomenon of "loss aversion")

       •  Consumers might associate higher fuel economy with inexpensive, less well designed
          vehicles

       •  When buying vehicles, consumers may focus on visible attributes that convey status,
          such as size, and pay less attention to attributes such as fuel economy that typically do
          not visibly convey status

       •  Even if consumers have relevant knowledge, selecting a vehicle is a highly complex
          undertaking, involving many vehicle characteristics. In the face of such a
          complicated choice, consumers may use simplified decision rules

       •  Because consumers differ in how much they drive, they may already sort themselves
          into vehicles with different, but individually appropriate, levels of fuel economy in
          ways that an analysis based on an average driver does not identify

       •  Fuel-saving technologies  may impose hidden costs — adverse effects on other vehicle
          attributes

   If consumers are doing a good job of getting their efficient amount of fuel economy, their
willingness to pay for additional fuel savings, revealed in their purchase decisions, should
approximately equal expected future fuel savings. A review of the literature  sponsored by EPA
looked at the range of estimates of the value of fuel economy in consumer purchase decisions in
models of consumer vehicle purchase decisions; it found as many studies with undervaluation of
fuel economy as there were studies with about-right or overvaluation.14 The  studies used in that
review tended to emphasize modeling of vehicle purchase decisions rather than the role of fuel
                                              6-6

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                        GHG Emissions

economy in those decisions. Some recent academic research has looked specifically at the
question of the value of fuel economy.15 Busse et al. (2013) and Sallee et al. (2016) find that
consumers appear to buy fuel economy that does approximate fuel savings; Allcott and Wozny
(2014) find in contrast that the willingness to pay for fuel economy is about 3/4 of the expected
future fuel savings.  Thus, consumers appear to take fuel economy into account when buying
vehicles, but how precisely they do it is not yet clear.

   The 2015 National Academies of Sciences report titled, "Cost, Effectiveness, and Deployment
of Fuel Economy Technologies for Light-Duty Vehicles"16 also reviewed the literature. Among
the studies that NAS reviewed was a 2013 paper by Greene, Evans, and Hiestand, regarding
which the NAS Committee stated, "Four nationwide random sample surveys of 1,000
respondents each, conducted between 2004  and 2013, showed that consumers considered fuel
economy ratings and future fuel prices to be very uncertain. . . .  The surveys also produced
consistent evidence that consumer willingness to pay for fuel savings implies average payback
periods of 2-3 years" (p. 317). Regarding the overall review of the literature conducted by the
NAS Committee, the Committee concludes,

   "How markets actually value increases in new vehicle fuel economy is critical to evaluating
the costs and benefits of fuel economy and GHG standards. Unfortunately, the scientific
literature does not provide a definitive answer at present.  ... In the committee's judgment, there
is a good deal of evidence that the market appears to undervalue fuel economy relative to its
expected present value, but recent work suggests that there could be many reasons underlying
this, and that it may  not be true for all consumers. Given the importance of this question to the
rationale for regulatory standards and their costs and benefits, an improved understanding of
consumer behavior about this issue would be of great value." (p. 318)

   The agencies seek comment on consumer willingness-to-pay for fuel economy, including
considerations of payback periods on the order of 2-3 years, or more,  or less.

   Consumers cannot buy technologies that are not produced; some of the gap  in energy
efficiency may be explained from the producer's side.  Two major themes arise on the producer
side: the role of market structure and business strategy, and the nature of technological invention
and innovation.

   •   Light-duty vehicle production involves significant fixed  costs, and automakers strive to
       differentiate  their products from each other. These observations suggest that automakers,
       rather than meeting the stylized economic model of perfect competition, can act
       strategically  in how they design and market products.  In this context, the fuel economy
       of a vehicle can become a factor in product differentiation rather than a decision based
       solely on cost-effectiveness of a fuel-saving technology.17 Product differentiation carves
       out corners of the market for different automobile brands. For instance, automakers may
       emphasize luxury characteristics in some vehicles to attract people with preferences for
       those characteristics, and they may emphasize cost and fuel economy for people attracted
       to frugality.  By separating products into different market segments, producers both
       provide consumers with goods targeted for their tastes, and may reduce competition
       among vehicle models, creating the possibility of greater profits.  From the producer
       perspective, fuel economy is not necessarily closely related to the cost-effectiveness of
                                              6-7

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                          GHG Emissions
       the technologies to consumers, but rather is one of many factors that manufacturers use to
       market their models to different consumer groups. As Fischer (2005) points out, this
       strategy can lead to inefficiencies in the market: an under-supply of fuel economy relative
       to what is cost-effective to consumers in some segments, and an over-supply of fuel
       economy in other sectors.18 The structure of the automobile industry may inefficiently
       allocate car attributes-fuel economy among them—and help to explain the existence of an
       energy efficiency gap.
    •   Chapter 4.1.3 discusses the relationship between technological innovation and the
       standards, but a shortened discussion is relevant here. In particular, in the absence of
       standards, automakers are likely to invest in small improvements upon existing
       technologies ("incremental" technologies) that can be used to improve fuel economy or
       other vehicle attributes. On the other hand, they may be more hesitant to invest in
       "major" innovations in the absence of standards, for several reasons.
          o  There may be first-mover disadvantages to investing in new technologies. Many
              manufacturers prefer to observe the market and follow other manufacturers rather
              than be the first to market with a specific technology. The "first-mover
              disadvantage" has been recognized in other research where the "first-mover" pays
              a higher proportion of the costs of developing technology, but loses the long-term
              advantage when other businesses follow quickly.19
          o  There could be "dynamic increasing returns" to adopting new technologies,
              wherein the value of a new technology may depend on how many other
              companies have adopted the technology — for instance, creating multiple
              suppliers for a technology should increase competition, improve quality, and
              reduce price.  This could be due to network effects or learning-by-doing.  In a
              network effects situation, the usefulness of the technology depends on others'
              adoption of the technology: e.g., a telephone is only useful if other people also
              have telephones.  Learning by doing is the concept that the costs (benefits) of
              using a particular  technology decrease (increase) with use.  Both of these
              incentivize firms to pursue a "wait and see" strategy when it comes to adopting
              new technologies.20
          o  There can be synergies when companies work on the same technologies at the
              same time.21 Research among multiple parties can be a synergistic process: ideas
              by one researcher may stimulate new ideas by others, and more and better results
              occur than if the one researcher operated in isolation.22>E Collaboration between
              automotive companies or automotive suppliers does occur.  For example, in 2013,
              Daimler, Ford, and Nissan teamed up to work on fuel cell vehicles,23 and Toyota
E Powell, Walter W., and Eric Giannella (2010). "Collective Invention and Inventor Networks," Chapter 13 in
  Handbook of the Economics of Innovation. Volume 1, edited by B. Hall and N. Rosenberg (Elsevier) discuss how
  a "collective momentum" has led uncoordinated research efforts among a diverse set of players to develop
  advances in a number of technologies (such as electricity and telephones). They contrast this view of
  technological innovation with that of proprietary research in corporate laboratories, where the research is part of a
  corporate strategy. Such momentum may result in part from alignment of economic, social, political, and other
  goals.
                                               6-8

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                            GHG Emissions

              and BMW teamed up to work on battery technology.24 In 2015 Toyota and
              Mazda "agreed to form a 'long-term partnership'" to collaborate on numerous
              advanced technologies, including plug-in hybrid and fuel cell systems.25
              Standards can promote research into low-CO2 technologies that would not take
              place in the absence of the standards. Because all companies (both auto firms and
              auto suppliers) have incentives to find better, less expensive ways of meeting the
              standards, the possibilities for synergistic interactions may increase.  Thus, the
              standards, by focusing all companies on finding more efficient ways of achieving
              the standards, may lead to better outcomes than if any one company operated on
              its own.

   These potential explanations are relevant, of course, if the  efficiency gap exists for vehicles.
If the gap does not exist, then there is no need to understand reasons for it. To understand the
effects of the standards, EPA has therefore been focusing on the existence of the gap.  If the gap
exists, then the standards are providing net benefits to vehicle buyers, even if it is unclear why
this is happening.26

   The existence of the gap depends on whether fuel-saving technologies that would not have
been used in the absence of the standards provide net benefits to new vehicle buyers even when
the externalities associated with the standards are not included.  The net benefits calculation
involves three components: the technology's effectiveness (which, along with fuel prices and the
amount driven, determines the fuel savings);F the technology's costs; and whether there are any
adverse unintended consequences of the technologies (hidden costs), such as interference with
the vehicle's handling or braking.0 Chapter 5 discusses the technology costs and  effectiveness of
the technologies that may be used to achieve the standards. The next section describes research
that EPA has conducted  to assess the existence of potential hidden costs associated with these
technologies.

6.4    Consumer Response to Vehicles Subject to  the Standards

6.4.1   Recent New Vehicles

6.4.1.1 Sales

   One measure of consumer response to the vehicles subject to the standards is the effects of the
standards on vehicle sales. As discussed in Chapter 3 and in Chapter 6.1, it is difficult, if not
impossible, to separately identify the effects of the standards on vehicles sales from the effects of
F Fuel-saving technologies provide different cost savings across consumers, because they drive different amounts
  under different conditions (which affect miles per gallon). As noted above, if each consumers gets individually
  optimal fuel economy in a vehicle that meets his/her other needs, then the efficiency gap does not exist even if an
  analysis done based on an average driver shows potential for increased efficiency.
G Note that the agencies' modeling work on technological effectiveness builds in the need to maintain all aspects of
  vehicle performance.  That is, the methodology includes all costs of implementing the technologies to achieve
  GHG reductions while maintaining all aspects of performance and utility. The agencies thus concluded that
  adding fuel-saving technologies results in no loss of vehicle utility, and that adding fuel-saving technologies will
  not preclude future improvements in performance, safety, or other attributes. See generally Chapter 3.2 of 2017-
  2025 MY TSD, and 77 FR 62714/2.  Chapter 4.1.3 and the next sub-chapter further discuss the relationship
  between the standards and other vehicle attributes.
                                                6-9

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                            GHG Emissions

recovery from recession.  It appears that the standards did not prevent recovery of auto sales
from the recession, but it is not possible to say whether the standards helped or hindered that
recovery.

6.4.1.2 Evaluations from Professional Auto Reviewers

   Another way that EPA is examining the effects of the standards on new vehicles is through
analysis of the evaluations that professional auto reviewers give to fuel-saving technologies.27
Auto reviews are a readily available and public source of information about the advantages and
disadvantages of new vehicle models.  We focused on professional automobile reviews because
professional reviewers have experience evaluating vehicle technologies and are expected to
identify any potential drawbacks to consumers (i.e., hidden costs) if they exist. Although
reviewers may not respond to vehicle technologies  in the same way that vehicle owners will, it
seems reasonable to expect that, if there are significant problems for particular technologies,
reviewers will comment on them.

   EPA commissioned RTI International to conduct a content analysis of auto  reviews for
MY2014 vehicles from six major websites that conduct professional auto reviews:  Automobile
Magazine, Auto Trader, Car and Driver, Consumer Reports, Edmunds, and Motor Trend.28
Content analysis is a research technique that breaks text into pre-defined sub-units  that can be
categorized and analyzed into specified definitional codes.H Staff at RTI read  each auto review
from a professional reviewer (reader reviews or comments were not included in the study) and
coded each mention of specific fuel-saving technologies for whether the reviewer evaluated it as
positive, negative, or neutral.  In addition, they coded mentions of a number of operational
characteristics, such as handling, acceleration, and noise. The initial dataset included 1023
reviews. After further review of the data, the final set includes 1,003 separate reviews,
containing 3,535 separate evaluations of various fuel-saving technologies.1

   Table 6.1  shows the results aggregated to the review level/ For each technology,  positive
evaluations exceed negative evaluations. Indeed, in the aggregate, negative evaluations are less
than 20 percent of the totals. Even the most negatively reviewed technologies  - continuously
variable transmissions (51 percent positive) and stop-start (59 percent positive) - have majority
positive evaluations. These results suggest that it is possible to implement these technologies
without significant hidden costs. The NAS report suggests a similar conclusion: '"It  is not
technology per se that  generates new problems, but rather its integration and execution,' Neal
H There are many descriptions of content analysis and its evolution as a research methodology; see Helfand et al.
  (2015), footnote 22, for background and citations.
1 The initial dataset inadvertently contained reviews of 15 vehicles not subject to the standards, primarily medium-
  duty trucks that had not previously been eliminated. In addition, due to issuance of a notice of violation about the
  compliance of some Volkswagen diesel engines with emissions standards, we dropped 5 reviews of those
  vehicles.
1 Each review could contain mentions of more than one technology, or even multiple mentions of the same
  technology. The review-level results aggregate all like mentions of a technology in one review. For instance, if a
  review contains 3 positive mentions of turbocharging, the review-level results count them as 1 positive mention.
  If the review contains 3 positive mentions and 1 negative mention, at the review level these are counted as 1
  positive and 1 negative mention. The data were analyzed both at the level of individual codes, and aggregated to
  review. With the results very similar, we here focus on the review-level results. See Helfand et al. (2015) for
  more detail, including code-level results.
                                                6-10

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                         GHG Emissions
Oddes, Director of Product Research and Analysis at J.D. Power, noted (Janes 2013), an
observation that could be made for some of the fuel-saving technologies being launched today'
(p. 9-21).

      Table 6.1 Efficiency Technology's Positive, Negative, or Neutral Evaluations by Auto Reviews
Efficiency Technology Categories
Active Air Dam
Active Grill Shutters
Active Ride Height
Electric Assist or Low Drag Brakes
Lighting - LED
Low Rolling Resistance Tires
Mass Reduction
Passive Aerodynamics
Powertrain

Engine
General
Powertrain
Transmission

Coding level
Active air dam
Active grill
shutters
Active ride height
Electric assist or
low drag brakes
Lighting-LED
Low rolling
resistance tires
Mass reduction
Passive
aerodynamics
Cylinder
deactivation
Diesel
Electronic power
steering
Full electric
GDI
General Engine
Hybrid
Plug-in hybrid
electric
Stop-start
Turbo-charged
General
Powertrain
CVT
DCT
General
Transmission
High speed
automatic
Total
Negative
-
-
-
1
1
4
-
4
1
7
45
2
6
104
16
4
14
20
8
35
16
30
60
378
-
-
-
14%
5%
24%
-
10%
3%
12%
22%
9%
9%
16%
23%
14%
27%
9%
8%
31%
24%
18%
14%
16%
Neutral
-
-
1
3
2
5
9
7
4
9
42
6
6
95
10
6
7
23
19
20
10
26
81
391
-
-
33%
43%
10%
29%
12%
18%
11%
15%
20%
27%
9%
15%
14%
21%
14%
10%
18%
18%
15%
16%
20%
16%
Positive
6
1
2
3
17
8
65
29
30
44
121
14
54
443
45
18
30
180
78
57
42
108
273
1,668
100%
100%
67%
43%
85%
47%
88%
73%
86%
73%
58%
64%
82%
69%
63%
64%
59%
81%
74%
51%
62%
66%
66%
68%
Total
6
1
3
7
20
17
74
40
35
60
208
22
66
642
71
28
51
223
105
112
68
164
414
2,437
   Further evaluation of the data involves looking at correlations between evaluations of each
technology and a range of operational characteristics (handling, acceleration, noise, etc.).  In
particular, this evaluation assesses how the technologies are related to negative evaluations of
these characteristics. If the technologies have hidden costs, the research premise is that the
                                              6-11

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                         GHG Emissions

technologies should be positively correlated with negative evaluations of operational
characteristics.  The results do not reveal much evidence of such correlation. When correlations
exist, often they are not statistically robust; their statistical significances change depending on
what covariates are considered.  For instance, seven technologies have at least one statistically
significant correlation with the characteristic of acceleration capability in six versions of the
model, but only one (continuously variable transmissions) has a statistically significant
correlation across all six model versions (its existence is correlated with negative effects on
acceleration capability). At the same time, in five of six models, the existence of stop-start
technology is significantly associated with reduced probability of negative evaluations of
acceleration capability. Indeed, across all characteristics, there are more instances of fuel-saving
technologies associated with lower probabilities of negative  evaluations of characteristics than
with increased negative evaluations. In addition, negative evaluations of characteristics are more
likely if the technology itself has a negative evaluation — in other words, it seems that a bad
implementation of the technology is associated with bad characteristics, rather than there being
some inherent problem in the technology. If it is possible to implement a technology to avoid
hidden costs,  as these data suggest, then automakers should be able to improve implementation
over time; in such a circumstance, any problems with hidden costs may be temporary.

   These findings on the relationship of technologies to hidden costs or hidden benefits  have
some limitations.  They appear sensitive to how the analysis is done, and the magnitudes are
often small. Perhaps more importantly, it is not possible to determine whether the technologies
themselves cause these effects, or whether these associations are due to the vehicles in which the
technologies are installed.  For instance, perhaps stop-start was put in vehicles that would have
had better acceleration even without it.  As a result, this research is not able to disprove  the
possibility of hidden costs (or benefits). In addition, this research cannot determine what, if any,
additional costs may have been incurred to mitigate problems with the technologies. It
nevertheless fails to find evidence of systematic hidden costs associated with fuel-saving
technologies.  The agencies seek comment providing additional evidence related to concerns
over hidden costs.

   Helfand et al. (2015)29 provides further detail about the methods and results of this work,
including additional limitations. Note that this research examines how professional auto
reviewers respond to these technologies, rather than how vehicle buyers respond. If the public
tends to be harsher critics than the reviewers, then these results may understate negative
consumer response. In addition, reviewers spend much less  time with any  one vehicle than a
vehicle owner; something that a reviewer may not notice in a few hours of test driving may
become significant to an owner over time.  On the other hand, we expect professional auto
reviewers, as  experts, to be aware of vehicle characteristics and technologies more than  the
general public.  Thus, consumer response to these technologies may be either more or less
critical than reviewer response.

6.4.1.3 Consumer Responses to New Vehicles

   Another potential source of information on consumer response to vehicles subject to  the GHG
and fuel economy standards can come from market research firms that conduct surveys  of new
vehicle buyers.  These surveys, typically conducted a few months after purchase of a new
vehicle, ask the buyer's views on a wide range of vehicle attributes. EPA has been pursuing
access to one of these survey data sets.  Our goal would be to look for associations between the
                                              6-12

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                           GHG Emissions

existence of fuel-saving technologies and consumer responses to vehicle attributes: for instance,
do consumers rate satisfaction with their vehicles differently for vehicles with stop-start systems
relative to those without such systems, controlling for other vehicle characteristics? This
research would provide direct insights into consumer attitudes.

   EPA is still pursuing access to such a database; results from it are not available for this Draft
TAR. If we are successful in gaining access, we intend to use the information to inform the
midterm evaluation.

6.4.2  MY2022-25 Vehicles

   To date, it seems difficult to find evidence that the standards have posed significant obstacles
to consumer acceptance: vehicle sales are very strong, and we have not found evidence of
inherent "hidden costs" of the technologies, at the same time that the auto industry as a whole has
over-complied with the standards (see Chapter 3.3).K As the standards continue to become more
stringent, though, there will be both more application of existing technologies to new vehicles,
and new or improved technologies are likely to be developed.  As discussed in Chapter 4.1.3,
these standards themselves may be  contributing to innovation that would not have happened in
their absence. As a result, it is difficult to extrapolate to future technologies from findings
related to existing ones.

   There is, of course, uncertainty about which technologies will be necessary to achieve the
MY2022-25 standards. In the MY2017-25 rulemaking analysis, EPA projected that the
standards could be achieved primarily with gasoline  vehicles; it estimated only  about 2 percent
penetration of plug-in electric vehicles (PEVs), either plug-in hybrid electric vehicles  (PHEVs)
or all-battery EVs (BEVs).30 The NAS also expects  the spark-ignition gasoline engine to
dominate the auto market through, and beyond, 2025.31 For these vehicles, the  effects of the
standards on consumer acceptance depend on the costs, effectiveness, and potential tradeoffs or
synergies of those technologies with other attributes; there is already an established infrastructure
for fuel availability. If the standards can be achieved primarily with greater penetration of
existing technologies, we  do not have evidence of significant problems for consumer acceptance.
On the other hand, if the standards can be achieved only with increased utilization of new
technologies, these new technologies could raise the  possibility of new challenges.

   The role of electrified vehicles in particular in achieving the standards has led to questions
about consumer acceptance of those vehicles.L Some states,M led by California, are requiring
greater use of plug-in electric vehicles (PEVs) and fuel cell electric vehicles (FCEVs) for
meeting state air quality and greenhouse gas targets,  and these vehicles are also included in
automaker fleets that are subject to  the National Program. If EVs become a more important part
of the compliance strategy for the 2022-25 standards, then their unique features — in particular,
K Design elements of program, such as targeting emissions rather than specific technologies, averaging and banking
  credits, and allowing credit trades, are expected to have facilitated compliance by providing manufacturers with
  great flexibility in meeting the standards.
L We do not include conventional hybrid-electric vehicles (HEVs) in this discussion. Because they are fueled solely
  by gasoline and rely on the same infrastructure as other gasoline vehicles, they are part of the gasoline-vehicle
  market.
M Connecticut, Maine, Maryland, Massachusetts, New Jersey, New York, Oregon, Rhode Island, and Vermont have
  adopted the California Zero Emission Vehicle program.
                                               6-13

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                         GHG Emissions

the need for infrastructure and the associated concerns over vehicle range, as well as differences
(many positive) in other attributes — are likely to have an effect on consumer acceptance.

   As noted in the 2017-25 Preamble,32 the National Program standards are performance-based;
there is no mandate under the National Program for any manufacturer to use any particular kind
of technology, or for any consumer to choose,  any particular kind of vehicle. If the variety of
vehicles in the conventional fleet does not shrink, the availability of PEVs should not reduce
consumer welfare compared to a fleet with no PEVs: increasing options should not reduce
consumer well-being, because other existing options still are available.  An individual consumer
will buy a PEV only if the price and characteristics of the vehicle make it more attractive to her
than other vehicles. Already, many current PEV options are  versions of gasoline-only vehicles,
for example, the  Chevrolet Spark EV, the FIAT 500e, all of Ford's PEV products, and the
Volkswagen e-Golf  The forthcoming Hyundai loniq will be offered as a conventional hybrid,
plug-in hybrid, and all-battery electric vehicle, allowing consumers to choose the degree of
electrification best suited to their needs.  Similarly, both Volvo and BMW have announced plans
to offer plug-in hybrid variants over a wide range of existing and new models.

   On the other hand, if the only compliance path available to automakers involves more use of
PEVs than markets would normally support (in the absence of government incentives), then
achieving the standards may lead automakers and dealers to encourage the market for PEVs by
providing incentives for PEV purchase sufficient to meet the standards. This encouragement can
come in various forms — for instance, through marketing and advertising, through sales
incentives, or through increased education about PEVs to potential buyers to increase buyer
familiarity with the technology. Automakers may also cross-subsidize sales as they have long
been able to do to meet fleet average standards; in this case using higher prices on conventional
vehicles to support lower prices on PEVs, to increase sales of PEVs relative to gasoline vehicles
beyond levels that markets would support in the absence of the standards.  Cross-subsidization
would be expected to reduce auto industry profits.

   If consumers are willing to purchase PEVs (and other low-GHG-emitting vehicles) at prices
that provide adequate profits to manufacturers, then consumer acceptance is sufficient to
maintain a functioning auto market.  As discussed in Section 3.1.5,  PEVs are currently estimated
to be about 1.1 percent of MY2015  sales. Section 5.2.4 discusses these technologies and the
technological advances being made. As that section presents, this market is evolving rapidly,
with expected increases in model diversity, vehicle range, decreased costs, and expansion of
infrastructure (see Chapter 9). Although PEV range is often  cited as a concern for consumer
acceptance, it should be noted that PEVs have some desirable characteristics relative to gasoline
vehicles, including higher low end torque, potentially higher acceleration,  lower operating costs,
and the convenience of refueling by plugging in at home.N>33  Consumer acceptance of these
vehicles will depend on the degree of all  these factors, plus the differences in attributes, both
positive and negative, of PEVs relative to gasoline vehicles.  Additionally, many automakers
have announced moderately priced BEVs with longer ranges, and various public and/or private
initiatives continue to increase investments in public and workplace infrastructure that will
further alleviate concerns about range.
N The Tesla Model S, an all-electric vehicle, for instance, has regularly been achieving top ratings from standard
  auto reviewers for its handling and power.
                                              6-14

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                       GHG Emissions

   While concerns over range and cost are often cited as primary obstacles to PEV adoption, lack
of awareness and understanding of PEVs, perhaps including misunderstanding, itself creates
another barrier to adoption.34 A 2015 survey by the National Renewable Energy Laboratory
(NREL) of over 1,000 U.S. households found that less than half of the respondents could name a
specific PEV model, despite being available on the market for over four years.35 Using this same
measure, awareness levels were even lower in a 2015 University of California, Davis survey of
5,600 households that purchased a new vehicle after 2008.36

   The National Academy of Sciences Committee on Overcoming Barriers to Electric-Vehicle
Deployment37 notes that many people consider PEVs, as new technologies, to involve
uncertainty and risk compared to gasoline vehicles, and thus are hesitant to consider them. It
cites as barriers "the limited variety and availability of PEVs; misunderstandings concerning
range of PEVs; difficulties in understanding electricity consumption, calculating fuel costs, and
determining charging infrastructure needs; complexities of installing home charging; difficulties
in determining the 'greenness' of the vehicle; lack of information on incentives; and lack of
knowledge of unique  PEV benefits" (p. 47).

   Some  studies suggest that experience with the technology increases acceptance.38 Indeed, a
survey of PEV drivers in California shows that the vehicle test drive and other PEV drivers to be
the two information sources most influential in a consumer's purchase decision. Yet, if people
view PEVs as risky and are thus reluctant to try them, then it will be difficult for them to gain
experience that would make them more comfortable with the technology.

   The NAS Committee discusses the role of auto dealers in helping consumers to understand
PEVs. It notes PEV buyers' dissatisfaction with the dealer experience, greater than that of buyers
of conventional vehicles.39 It cites evidence that salespeople are not very knowledgeable about
PEVs, and may not get adequate financial incentives for the extra time that PEV buyers may
require. Many dealers have no or few PEVs in their stock. At most dealerships the explanation
for not having PEVs in stock is "high demand" for the vehicles; the second-most common
explanation, in contrast, is a "lack of consumer interest" (p. 52).  These problems with
consumers' experiences with PEV dealers may contribute to the slow adoption of PEVs in the
market.

   For a small segment of the public, PEVs already are suitable for their purposes. As the
technology of PEVs evolves, especially as range and fueling infrastructure expand, it is likely
that a larger segment  could find PEVs suitable. As the NAS Committee notes, these issues arise
with adoption and diffusion of many new technologies, and are not unique to PEVs.
Overcoming these barriers, it argues, will require both public policy incentives and methods to
promote consumer experience with them. As noted, some research suggests that some perceived
barriers, such as concerns over charging, may become smaller with experience, while some
perceived advantages may be strengthened.40  Thus, consumer acceptance of PEVs may depend,
not only on technological advances, but also on the feedback loop associated with other
consumers purchasing PEVs.
                                             6-15

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                        GHG Emissions

6.5    Impacts of the Standards on Vehicle Affordability

   Because the standards are expected to increase the up-front costs of new vehicles, with the
fuel savings that recover those costs coming over time, questions arose in comments on the
2017-25 LD GHG rule about the effects of the standards on affordability. We analyze this
question by considering the effects of the standards on lower-income households, on the used
vehicle market, on whether access to credit  may limit consumers' ability to purchase new
vehicles, and on the availability of low-priced vehicles.  Further detail may be found in Cassidy
etal.41

6.5.1   Effects on Lower-Income Households

   We begin here by examining the effects of the standards separately for lower- and higher-
income households. We consider lower-income households to be those that had after-tax
incomes below the weighted median0 income in a given year, and higher-income households to
be those that had after-tax incomes above that threshold. For example, the weighted median in
2013 is $33,371. For this analysis, we use the 2007-2013 Consumer Expenditure Survey (CES),
which is conducted annually by the Bureau  of Labor Statistics of the U.S. Department of Labor
and provides information on the expenditures, income and characteristics of U.S. households, as
well as federal poverty levels.42'1"

   The effects of this rule on lower income households depend on its impacts, not only in the
new vehicle market, but also in the used vehicle market. Using CES data from 2007-2013, on
average, 29 percent of new car buyers were lower income according to our definition.*2 The
2013 Consumer Expenditure Survey  data indicate that lower income households on average
spent  more in 2013 on gasoline ($2,154) than on vehicles ($670); in addition, they spent more on
used vehicles ($362) than on new vehicles ($308). These results are analogous to those that
Consumer Federation of America (CFA) provided in comments on the 2017-25 standards.  CFA
found that households with income less than $20,000 per year in 2010 accounted for 22 percent
of households but only 2 percent of money spent on new vehicles; those households spent 7.3
times  as much on gasoline as on new car payments.43 These data suggest that lower income
households are more affected by the impact of the rule on the used vehicle market than on the
new vehicle market, and that they are more  vulnerable to changes in fuel prices than they are to
changes in vehicle prices.

6.5.2   Effects on the Used Vehicle  Market

   The effect of this rule on the used vehicle market will be related to its effects on new vehicle
prices, the fuel efficiency of new vehicle models, the fuel efficiency of used vehicles, and the
total sales of new vehicles.  If the consumer value of fuel savings resulting from improved fuel
efficiency outweighs the average increase in new models' prices to potential  buyers of new
vehicles, sales of new vehicles could rise, and the used vehicle market may increase in volume as
0 The weighting, from the Bureau of Labor Statistics, corrects for under- or over-representation of certain
  households in each sample. The weighted median thus reflects the U.S. median rather than the sample median.
p The Federal Poverty Level is calculated annually by the Department of Health and Human Services. It varies with
  household size and for households in Alaska and Hawaii.
Q The CES data have many missing data. We present these results on the assumption that omitted information on
  vehicle purchases is not affected by household income.
                                              6-16

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                         GHG Emissions

new vehicle buyers sell their older vehicles. In this case, used vehicle buyers, including lower-
income households, are likely to benefit from the increased inventory of used vehicles.
However, if potential buyers value future fuel savings resulting from the increased fuel
efficiency of new models at less than the increase in their selling prices,  sales of new vehicles
may decline, and the used vehicle market may see price increases as people hold onto their
vehicles longer.

   Jacobsen and van Bentham (2015) look at the effect of fuel prices and fuel standards on the
used vehicle market.44 They argue that the increased price of new vehicles subject to the
standards will decrease new vehicle sales, and increase sales and prices in the used vehicle
market. As people switch to used vehicles, the greenhouse gas benefits of more efficient new
vehicles will be reduced.  Their results depend on the standards depressing new vehicle sales.R
As discussed in Chapter 6.2, we have not identified ways to estimate the effects of the standards
on new vehicle sales.

   Figure 6.2 presents data from the Consumer Price Index for used45 and new vehicle.46 Each
series has been adjusted to a year 2013 reference base with underlying prices in 2013$ (using
price deflators for GDP47) so that numbers on the_y-axis represent the percentage difference from
price levels in 2013 (in 2013$).  Used vehicle prices have decreased since 1995, and have varied
in a small range between 2008 and 2015.  The used car price index closely follows the new car
price index, although used car prices have more volatility across all years.  Mannheim
Consulting indicates that volumes at auto auctions have increased steadily from 2011-2015, with
relatively small fluctuations in its value index during that time.48 These  suggest that the increase
in new vehicle sales since the recession ended (see Chapter 6.1) has had  the expected positive
effect on used vehicle volumes; price reflects "strong new vehicle pricing, exceptional credit
conditions, higher employment levels, record job stability, and the often  overlooked factor of
increased dealership operating efficiencies" (Mannheim Consulting, p. 15). The average loan
payment for used vehicles, in nominal terms, increased by $6/month between 2014 and 2015;49
in constant 2013$, the payment is approximately constant, at $350/month.  This observation
again  does not suggest great movement in overall used vehicle prices.  Additionally, trends in the
new vehicle market, supply of used vehicles, and changing consumer preferences may even
result in used prices falling for certain market segments; January 2016 used vehicle prices for
compact and luxury cars fell relative to the prior year, while prices  for used pickups increased.50
As with the effects of the standards on new vehicle  sales, it is possible that the GHG/fuel
economy standards have had some influence on these trends, but their effect is likely swamped
by the effects of the economic recovery.
R The applicability of their empirical analysis is limited due to their use of pre-2009 data (including cost data from
  2002) and a flat (not footprint-based) standard, among other assumptions.
                                              6-17

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                            GHG Emissions

   165
    85


    75
• Used Car CPI (Urban Consumers)
• New Vehicles CPI (Urban Consumers)
                Figure 6.2 Used and New Car Consumer Price Index, 2013=100 (2013$).

   A recent Heritage Foundation analysis51 by Furth and Kreutzer (2016) cites a similar set of
price trends to argue that prices of new vehicles are higher by larger amounts (up to $7100) than
they would be if they had followed trends before 2009, trends in furnishings and durable
household equipment, or trends in vehicle prices in the United Kingdom or in Australia.  It
implies that the  standards created this divergence between the previous trend and current prices.
This change in the price trend is unlikely to be due only, or even primarily, to the standards,
though. These price trends are based on the vehicles that people are buying, not on a constant
vehicle model; that is, if people are switching from less expensive to more expensive vehicles,
then price trends would increase, even if the prices of individual vehicles had stayed constant.
As discussed in  Chapter 3.1.4, fleet mix has been changing during this time, with sales of SUVs
and pickup trucks higher than the estimates in the 2012 final rule. For instance, the share of the
fleet that is car (sedan) and not car SUV, truck SUV, pickup, or minivan went from 61 percent in
MY 2009 to 49 percent in MY 2014.52  To the extent that the latter vehicles are more expensive
than car sedans, the  change in sales mix will have affected the trend. Note  as well that the price
trend changes in 2008, at the start of the Great Recession, before the standards went into effect
for MY 2012.s Without a good way to separate effects  on prices due to the  standards from other
s Further evidence that these price trends are not due to the standards is found in comparing the trend in the United
  Kingdom (UK) with the trends in France, Germany, and Italy reported by Furth and Kreutzer (2016). The UK has
  a fairly steady, steep decrease in prices from 1999 to 2015, while France, Italy, and Germany have much flatter
  price trends; France and Italy show small decreases followed by a small upturn, while Germany has a steady but
  small decrease. All these countries are in the European Union, which provides a common set of standards for all
  countries. If standards alone were driving price trends, then these countries should all see similar trends. Instead,
  even if the France, Italy, and Germany patterns are similar, the UK pattern is very different. Thus, vehicle
  standards alone do not seem to be driving price trends.
                                                6-18

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                        GHG Emissions

factors affecting prices, the Furth and Kreutzer (2016) assessment does not provide a sound basis
for estimating the effects of the standards on vehicle prices.

   The benefits of the standards for buyers of used vehicles will depend on two countervailing
effects from the improvement in fuel economy: the increased cost of the used vehicles attributed
to fuel-saving technologies, and the savings in fuel costs over time.  Depreciation of new vehicle
prices reduces the cost of the additional fuel economy for used vehicle buyers.  On the other
hand, because older vehicles are used less on average than new vehicles, the fuel savings will
accrue more slowly. On net, in this current Draft TAR, reduced up-front costs exceed the
reduction in fuel savings so that the payback period is shorter for used cars than for new cars; see
Chapter 12 for more details.

6.5.3  Effects on Access to Credit

   Even though projected fuel savings are expected to outweigh increased vehicle costs, some
concerns have been raised about whether higher vehicle prices may exclude prospective
consumers from the new vehicle market through effects on consumers' ability to finance
vehicles. If lenders focus on the amount of the vehicle loan, the person's current debt, and the
person's income when issuing loans, and do not consider the reduced operating costs associated
with fuel savings, then the higher up-front costs of the new vehicles  subject to the standards
could reduce buyers' ability to get loans (holding down payments constant). Thus, if lenders do
not take fuel savings into account in providing some loans, households that are borrowing near
the limit of their abilities to borrow may either have to change what vehicles they buy (including
possibly switching from new to used vehicles), or defer buying vehicles.

   The financing market appears to be evolving, apparently in response to consumers buying
more expensive vehicles, among other factors.  One way that the loan market appears to be
evolving is that the available term length of auto loans has increased. The average new car loan
in mid-2015 has a record repayment period of 67 months, and 29 percent of loans were for 73-84
months.53  While interest rates have been low by historic standards since the recession, longer
loans typically reduce (or keep constant) the monthly payments that  consumers make, though
with more payments required and perhaps higher interest rates. Though these longer terms may
ease consumers' abilities to buy more expensive vehicles than they otherwise would, they
increase the chances that a vehicle owner may end up "under water"  - that is, with a vehicle
worth less than the amount that the buyer still owes.  In addition, the number of new vehicles
being leased has increased, from 19 percent in 2010 to 27 percent in 2015.54 These changes
show an evolving financing market, though why the market is evolving is not clear: it may be
that vehicles have become more expensive, or it may be that consumers are choosing more
expensive vehicles, or that consumer preferences toward ownership are changing. Any link
between these changes and the standards is speculative.

   Another market innovation suggests that parts of the loan market  take fuel savings into
account in the lending decision.  Some lenders currently give discounts for loans to purchase
more fuel-efficient vehicles.55  An internet search on the term "green auto loan" produced more
than 50 lending institutions that provide reduced loan rates for more  fuel-efficient vehicles.56 A
third of credit unions responding to a recent survey offered some type of green auto loan.57 It
seems that some auto loan makers incentivize the financing of more  fuel-efficient vehicles.
                                             6-19

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                       GHG Emissions

   Comments from the National Automobile Dealers Association (NADA) on the 2017-25 LDV
standard58 argue that an increase in the purchase price of new vehicles would increase the debt-
to-income ratio (DTI) of potential buyers beyond a critical threshold, which may prevent these
buyers from being eligible for a loan. As discussed in the 2012 FRM,59 their assessment looked
at the number of drivers living in households who would be eligible for a loan of $11,750, but
not $14,750. It did not examine households likely to be in the market for new vehicles and was
based on inaccurate assumptions about the impacts of the standards on new vehicle prices.
Among other assumptions, it implies the disappearance of low-priced new vehicles, a topic
discussed below.

   Another assumption of the NADA analysis was that the DTI is an impassible obstacle for
lending. To determine whether this DTI threshold is rigid, we used CES to identify households
with over 36 percent DTI in order to gauge whether exceeding this threshold precludes
households from being able to finance a vehicle purchase. We chose this threshold based on
guidance from online sources stating that lenders prefer to give loans to consumers who  have a
DTI under 36 percent.60 In 2013, the CES data indicated that over 66 percent of households that
purchased either a new or used vehicle with a DTI of over 36 percent financed their car
purchases.  This suggests that it is possible to obtain a loan for a new vehicle even with a DTI
over the assumed threshold.  Thus, if increases in vehicle prices push some households over the
36 percent DTI, it nevertheless appears  possible for them to get loans.

6.5.4   Effects on Low-Priced Cars

   Low-priced vehicles may be considered an entry point for people into buying new vehicles
instead of used ones; automakers may seek to entice people to buy new vehicles through a low
price point, perhaps to build brand loyalty for future, more profitable sales.61 In comments on
the MY2017-25 LD GHG rule, concerns were raised that the standards would increase the cost
of low-priced vehicles sufficiently to eliminate this segment.  To examine this question,  we used
Ward's Automotive datasets62 to explore low-priced new car models over time. Low-priced new
models - in particular, those with manufacturer's suggested retail price (MSRP) of less than
$15,000 (2013$) for the base version — continue to exist in the automobile market. As  shown in
Figure 6.3, the number of new car models offered with an MSRP of under $15,000 (2013$) is not
large, but automakers to date have been able to preserve the number of offerings in this segment.
                                             6-20

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                      GHG Emissions
          o _
          CM
        
-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                        GHG Emissions
          o
          o
          o -
          CO
          o
          o
          in
          CN
        CO
        58
         .
        Of.
        CO
          o
          o
          in
          o
          o
          o -
             2000
2005                 2010
        Model Year
2015
        Figure 6.4 Minimum MSRP of All Car Models Available, from Ward's Automotive Data
   Note, however, that the lowest prices were observed in the years surrounding the recession;
recent higher prices may be driven, in part, by the strength of the U.S. economy. In the past, not
only was the low-priced vehicle segment a way to encourage first-time new vehicle purchasers,
but it also tended to include more fuel-efficient vehicles that assisted automakers in achieving
CAFE standards.63 The footprint-based standards, by encouraging improvements in GHG
emissions and fuel economy across the vehicle fleet, reduce the need for low-priced vehicles to
be a primary means of compliance with the standards. This change in incentives for the
marketing of this segment may contribute to the increases in the prices of vehicles previously in
this category. In addition, these vehicles may be gaining more content, such as improved
entertainment systems and electric windows, if they develop an identity as a desirable market
segment without regard to their previous purpose in enabling the sales of less efficient vehicles
and compliance with CAFE standards.64 For instance, the Nissan Versa, the lowest-priced
vehicle since MY2011, added Bluetooth, audio controls on the steering wheel, and speed-
sensitive volume control in MY2015.  It may be that the small, fuel-efficient vehicles previously
sold with low prices are evolving to fit consumer demand that prefers content to low prices.

   In sum, the low-priced vehicle segment still exists. Whether it continues to exist, and in what
form, may depend on the marketing plans of manufacturers: whether benefits are greater from
offering basic new vehicles to first-time new-vehicle buyers, or from making small vehicles
more attractive by adding more desirable features to them.

6.5.5   Conclusion
                                             6-22

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                         GHG Emissions

   It is difficult to assess the effects of the LDV GHG standards on vehicle affordability, due to
both challenges in defining affordability, and difficulties in separating the effects of the standards
from other market changes. Because lower-income households are likely to buy used vehicles,
the effects of the standards on lower-income households depend on its effects in both the new
and used vehicles.  In the used vehicle market, used vehicle prices do not appear to be increasing.
The effects of the standards on access to sufficient financing to purchase a new vehicle may not
be large: there continue to be loan discounts for fuel-efficient vehicles, and people with high
debt-to-income ratios appear able to get loans. The low-priced vehicle segment still exists,
though perhaps in changing form.  In sum, if the standards have affected vehicle affordability,
those effects do not appear to have been large enough to be obvious in our considerations of the
data.

   This assessment has focused on the effects of the standards on purchase affordability of
vehicles - that is, whether they become more difficult to purchase because of the increase in up-
front costs. The vehicles will also become less expensive to operate. The reduced operating
costs from fuel savings over time are still expected to exceed the increase in up-front vehicle
costs, as a further mitigation of any effects on vehicle affordability.
                                              6-23

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                                   GHG Emissions

   References
1 77 Federal Register 62784.
2 Bureau of Economic Analysis. "Real gross domestic product per capita." BEA Account Code A939RXO,
downloaded 2/10/2016; U.S. Environmental Protection Agency 2015. Light-Duty Automotive Technology, Carbon
Dioxide Emissions, and Fuel Economy Trends: 1975 through 2015. U.S. EPA-420-R-15-001, Office of
Transportation and Air Quality, December 2015.
3 American Automotive Policy Council (2015). "State of the U.S. Automotive Industry: Investment, Innovation,
Jobs, Exports, and America's Economic Competitiveness." http://americanautocouncil.org/sites/default/files/2015-
AAPC-Economic-Contribution-Report%28FINAL%29.pdf.
4 Environmental Protection Agency 2015.  Light-Duty Automotive Technology, Carbon Dioxide Emissions, and
Fuel Economy Trends: 1975  Through 2015. U.S. EPA-420-R-15-001, Office of Transportation and Air Quality,
December 2015.
5 Whitefoot, Kate S.,  and Steven J. Skerlos  (2012). "Design incentives to increase vehicle size created from the U.S.
footprint-based fuel economy standards." Energy Policy 41: 402-411.
6 See also Helfand, Gloria, and Ann Wolverton (2011). "Evaluating the Consumer Response to Fuel Economy: A
Review of the Literature." International Review of Environmental and Resource  Economics 5: 103-146.
7 Greene, David L. (2010). "How Consumers Value Fuel Economy: A Literature Review." EPA Report EPA-420-R-
10-008, http://www.epa.gov/otaq/climate/regulations/420rl0008.pdf.
8 Haaf, C.G., J.J. Michalek, W.R. Morrow,  and Y. Liu (2014). "Sensitivity of Vehicle Market Share Predictions to
Discrete Choice Model Specification." Journal of Mechanical Design 136: 121402-121402-9.
9 Raynaert, Mathias (2014). "Abatement Strategies and the Cost of Environmental Regulation: Emission Standards
on the European Car Market." KU Leuven Center for Economic Studies Discussion Paper Series DPS14.31.
10 Oak Ridge National Laboratory (2012). "Consumer Vehicle Choice Model Documentation." EPA-420-B-12-052,
http://www.epa.gov/otaq/climate/documents/420bl2052.pdf; Systems Research and Applications International, Inc.
(2012). "Peer Review for the  Consumer Vehicle Choice Model and Documentation."  EPA-420-R-12-013,
http://www.epa.gov/otaq/climate/documents/420rl2013.pdf.
11 Helfand, Gloria, Changzheng Liu, Marie  Donahue, Jacqueline Doremus, Ari Kahan, and Michael Shelby (2015).
"Testing a Model of Consumer Vehicle Purchases."  EPA-420-D-15-011,
http://www3.epa.gov/otaq/climate/documents/mte/420dl501 l.pdf.
12 Jaffe, A.B., and Stavins, R.N. (1994). "The Energy Paradox and the Diffusion  of Conservation Technology."
Resource and Energy Economics 16(2): 91-122.
13 Helfand, G., & Wolverton,  A. (2011). "Evaluating the consumer response to fuel economy: A review of the
literature." International Review of Environmental and Resource Economics 5(2), 103-146; Allcott, H., &
Greenstone, M. (2012). "Is there an energy  efficiency gap?" Journal of Economic Perspectives 26(1). 3-28;
Gillingham, K., and K. Palmer (2014). "Bridging the Energy Efficiency Gap: Policy Insights from Economic Theory
and Empirical Evidence." Review of Environmental Economics and Policy 8(1): 18-38.
14 Greene,  David L. (2010). "How Consumers Value Fuel Economy:  A Literature Review." EPA-420-R-10-008.
15 Allcott, Hunt, and Nathan Wozny (2014). "Gasoline Prices, Fuel Economy, and the Energy Paradox." Review of
Economics and Statistics 96: 779-795; Busse, Meghan R., Christopher R. Knittel, and Florian Zettelmeyer (2013).
"Are Consumers Myopic? Evidence from New and Used Car Purchases." American Economic Review 103: 220-
256; Sallee, James, Sarah West, and Wei Fan (2016). "Do Consumers Recognize the Value of Fuel Economy?
Evidence from Used Car Prices and Gasoline Price Fluctuations." Journal of Public Economics, forthcoming.
16 National Research  Council (2015). Cost.  Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles. Washington, D.C.: The National Academies Press.
17 Fischer,  Carolyn (2005). "On the Importance of the Supply Side in Demand-Side Management." Energy
Economics 27: 165-180; Blumstein, Carl, and Margaret Taylor (2013). "Rethinking the Energy-Efficiency Gap:
Producers, Intermediaries, and Innovation." Energy Institute at Haas  Working Paper WP 243; Houde, Sebastien, and
C. Anna Spurlock (2015). "Do Energy Efficiency Standards Improve Quality? Evidence from a Revealed Preference
Approach." Ernest Orlando Lawrence Berkeley National Laboratory  Working Paper LBNL-182701.
18 Fischer,  Carolyn (2005). "On the Importance of the Supply Side in Demand-Side Management." Energy
Economics 27:165-180.
                                                    6-24

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                                   GHG Emissions
19 Blumstein, Carl and Margaret Taylor (2013). "Rethinking the Energy-Efficiency Gap: Producers, Intermediaries,
and Innovation," Energy Institute at Haas Working Paper 243, University of California at Berkeley; Tirole, Jean
(1998). The Theory of Industrial Organization Cambridge, MA: MIT Press, pp. 400, 402.
20 Popp, D., Newell, R.G., and Jaffe, A.B. (2010). "Energy, the environment and technological change." In
Handbook of the Economics of Innovation 2nd ed. B.H. Hall, and N. Rosenberg, Elsevier; Vollebergh, Herman R.I,
and Edwin van der Werf (2014). "The Role of Standards in Eco-Innovation: Lessons for Policymakers." Review of
Environmental Economics and Policy 8(2): 230-248.
21 Powell, Walter W., and Eric Giannella (2010). "Collective Invention and Inventor Networks," Chapter 13 in
Handbook of the Economics of Innovation. Volume 1, ed. B. Hall and N. Rosenberg, Elsevier.
22 Powell, Walter W., and Eric Giannella (2010). "Collective Invention and Inventor Networks," Chapter 13 in
Handbook of the Economics of Innovation. Volume 1, edited by B. Hall and N. Rosenberg (Elsevier).
23 Hetzner, Christiaan. "UPDATE 3-Daimler, Ford and Nissan team up on fuel cell cars." Reuters. January 28, 2013.
http://www.reuters.com/article/2013/01/28/daimer-ford-nissan-idUSL5NOAX5QU20130128
24 Kubota, Yoko.  "Toyota, BMW to research lithium-air battery." Reuters. January 24, 2013.
http://www.reuters.com/article/2013/0 l/24/us-tovota-bmw-fuelcell-idUSBRE90NOL020130124.
25 Greimel, Hans. "Toyota, Mazda form Partnership to Share Technologies, Confront Cost Challenges." Automotive
News. May 13, 2015. http://www.autonews.com/article/20150513/OEM01/150519954/tovota-mazda-form-
partnership-to-share-technologies-confront-cost.
26 Helfand, Gloria, and Reid Dorsey-Palmateer (2015). "The Energy Efficiency Gap in EPA's Benefit-Cost Analysis
of Vehicle Greenhouse Gas Regulations: A Case Study," Journal of Benefit-Cost Analysis 6.
27 Helfand, Gloria, Jean-Marie Revelt, Lawrence Reichle, Kevin Bolon, Michael Me Williams, Mandy Sha, Amanda
Smith, and Robert Beach (2015). "Searching for Hidden Costs: A Technology-Based Approach to the Energy
Efficiency Gap in Light-Duty Vehicles." EPA-420-D-15-010.
28 Sha, Mandy, and Robert Beach (2015). "Content Analysis of Professional Automotive Reviews." Final Report,
Work Assignment 3-01, EPA Contract Number EP-C-11-045; Helfand, Gloria, Jean-Marie Revelt, Lawrence
Reichle, Kevin Bolon, Michael McWilliams, Mandy Sha, Amanda Smith, and Robert Beach (2015). "Searching for
Hidden Costs: A Technology-Based Approach to the Energy Efficiency Gap in Light-Duty Vehicles." EPA-420-D-
15-010.
29 Helfand, Gloria, Jean-Marie Revelt, Lawrence Reichle, Kevin Bolon, Michael McWilliams, Mandy Sha, Amanda
Smith, and Robert Beach (2015). "Searching for Hidden Costs: A Technology-Based Approach to the Energy
Efficiency Gap in Light-Duty Vehicles."  EPA-420-D-15-010.
30 U.S. Environmental Protection Agency (2012). "Regulatory Impact Analysis: Final Rulemaking for 2017-2025
Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards." EPA-
420-R-12-016, pp. 3-52 - 3-53.
31 National Research Council (2015). Cost. Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles. Washington, D.C.: The National Academies Press, p. S-4.
32 77 Federal Register 62917
33 DeMorro, Christopher (2015). "How Many Awards Has Tesla Won? This Infographic Tells Us. Clean Technical"
http://cleantechnica.com/2015/02/18/many-awards-tesla-won-infographic-tells-us/.
34 Consumer Federation of America (2015). "Knowledge Affects Consumer Interest in EVs, New EVs Guide to
Address Info Gap." http://consumerfed.org/press release/knowledge-affects-consumer-interest-in-evs-new-evs-
guide-to-address-info-gap/, accessed 3/15/16.
35 Singer, Mark (2016). "Consumer Views on Plug-In Electric Vehicles ~ National Benchmark Report." U.S.
Department of Energy, National Renewable Energy Laboratory Technical Report NREL/TP-5400-65279,
http://www.afdc.energv.gov/uploads/publication/consumer views_pev benchmark.pdf.
36 K. Kurani, N. Caperello, J. TyreeHageman; New Car Buyers' Valuation of Zero-Emission Vehicles: California,
March 2016, http://www.arb.ca.gov/research/apr/past/12-332.pdf.
37 National Research Council (2015). Overcoming Barriers to Deployment of Plug-In Electric Vehicles.
Washington, D.C.: National Academies Press.
38 Buhler, Franziska, Peter Cocron, Isabel Neumann, Thomas Franke, and Josef F. Krems (2014). "Is EV Experience
Related to EV Acceptance? Results from a German Field Study." Transportation Research Part F 25: 34-49;
Dudenhoffer (2013). "Why Electric Vehicles Failed." Journal of Management Control 24:95-124; Singer, Mark
(2016). Consumer Views on Plug-in Electric Vehicles — National Benchmark Report. National Renewable Energy
Laboratory Technical Report NREL/TP-5400-65279.


                                                    6-25

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                                   GHG Emissions
39 Cahill, E., J. Davies-Shawhyde, and T. Turrentine. 2014. "Zero-emission Vehicles and Retail Innovation in the
U.S. Automotive Sector: An Exploration of the Consumer Purchase Experience for Plug-in Electric Vehicles."
University of California, Davis Institute of Transportation Studies Working Paper, August 2014, cited in National
Research Council (2015). Overcoming Barriers to Deployment of Plug-In Electric Vehicles. Washington, D.C.:
National Academies Press.
40 Buhler, Franziska, Peter Cocron, Isabel Neumann, Thomas Franke, and Josef F. Krems (2014). "Is EV Experience
Related to EV Acceptance? Results from a German Field Study." Transportation Research Part F 25: 34-49.
41 Cassidy, Alecia, Geoffrey Burmeister, and Gloria Helfand. "Impacts of the Model Year 2017-25 Light-Duty
Vehicle Greenhouse Gas Emission Standards on Vehicle Affordability." Working paper, docket EPA-HQ-OAR-
2015-0827.
42 U.S. Bureau of Labor Statistics. "Consumer Expenditure Survey." http://www.bls.gov/cex/;  see Cassidy et al.
ibid., for more detail.
43 Consumer Federation of America, et al. "Comments of Consumer Groups." Docket EPA-HQ-OAR-2010-0799-
9419, http://www.regulations.gov/#!documentDetail:D=EPA-HQ-OAR-2010-0799-9419
44 Jacobsen, Mark, and Arthur van Bentham (2015). "Vehicle Scrappage and Gasoline Policy." American Economic
Review 105: 1312-1338.
45 U.S. Bureau of Labor Statistics. "Consumer Price Index for All Urban Consumers: Used cars and
trucks [CUSROOOOSETA02]," retrieved from FRED, Federal Reserve Bank of St. Louis,
https://research.stlouisfed.org/fred2/series/CUSROOOOSETA02, accessed 3/23/2016; see
"Annual_used_and_new_CIP_price_index_with_GDP_deflator.xlsx," Docket EPA-HQ-OAR-2015-0827.
46 U.S. Bureau of Labor Statistics. "Consumer Price Index for All Urban Consumers: New
vehicles [CUSROOOOSETA01]," retrieved from FRED, Federal Reserve Bank of St. Louis,
https://research.stlouisfed.org/fred2/series/CUSROOOOSETA01, accessed 3/23/2016; see
"Annual_used_and_new_CIP_price_index_with_GDP_deflator.xlsx," Docket EPA-HQ-OAR-2015-0827.
47 U.S. Department of Commerce, Bureau of Economic Analysis. "Table 1.1.9 Implicit Price Deflators for Gross
Domestic Product," http://www.bea.gov/iTable/iTable.cfm?ReqID=9&step=l#reqid=9&step=3&isuri=l&903=13.
accessed 3/23/2016; see "Annual_used_and_new_CIP_price_index_with_GDP_deflator.xlsx," Docket EPA-HQ-
OAR-2015-0827.
48 Mannheim (2016). "Used Car Market Report." http://www.manheim.com/contentjdfs/products/UCMR-
2016.pdf?WT.svl=m_prod consulting latestupdates button 2016 , accessed 2/11/2016.
49 Zabritski, Melinda (2015). "State of the Automotive Finance Market Second Quarter 2015."  Experian
Automotive, http://www.experian.com/assets/automotive/white-papers/experian-auto-2015-
q2.pdf?WT.srch=Auto Q22015FinanceTrends PDF , accessed 9/25/2015.
50 Mannheim (2016). "Mannheim Used Vehicle Value Index."
https://www.manheim.com/content images/content/ManheimUsedVehicleValueIndex-BarGraph0116.jpg  , accessed
3/15/16.
51 Furth, Salim, and David W. Kreutzer (2016). "Fuel Economy Standards are a Costly Mistake." The Heritafe
Foundation Backgrounder, http://www.heritage.org/research/reports/2016/03/fuel-economy-standards-are-a-costly-
mistake, downloaded 5/20/2016.
52 U.S. EPA (2015). "Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends:
1975-2015," EPA-420-S-15-001, Appendix D, https://www3.epa.gov/fueleconomy/fetrends/l975-2015/420rl5016-
appendix-d.xlsx, accessed 5/23/16.
53 Zabritski, Melinda (2015). "State of the Automotive Finance Market Second Quarter 2015."  Experian
Automotive, http://www.experian.com/assets/automotive/white-papers/experian-auto-2015-
q2.pdf?WT.srch=Auto Q22015FinanceTrends PDF , accessed 9/25/2015; Gardner, Greg (2015). "New-car loans
keep getting longer." USAToday  June 1. 2015. http://www.usatodav.com/storv/monev/cars/2015/06/01/new-car-
loans-term-length/28303991/. downloaded 9/25/2015.
54 Zabritski, Melinda (2015). "State of the Automotive Finance Market Second Quarter 2015."  Experian
Automotive, http://www.experian.com/assets/automotive/white-papers/experian-auto-2015-
q2.pdf?WT.srch=Auto Q22015FinanceTrends PDF , accessed 9/25/2015.
55 See, for instance, Ladika, Susan (2009). '"Green auto loans offer lower rates," Bankrate.com,
http://www.bankrate.com/finance/auto/green-auto-loans-offer-lower-rates-l.aspx, accessed 7/29/15, Docket EPA-
HQ-OAR-2010-0799-11829.
                                                    6-26

-------
  Assessment of Consumer Acceptance of Technologies that Reduce Fuel Consumption and
                                                                                   GHG Emissions
56Cassidy, Alecia, Geoffrey Burmeister, and Gloria Helfand. "Impacts of the Model Year 2017-25 Light-Duty
Vehicle Greenhouse Gas Emission Standards on Vehicle Affordability." Working paper, docket EPA-HQ-OAR-
2015-0827.
57 Baumhefner, Max (2013). "Why Can't Your Loan be as Green and Efficient as Your Vehicle?" Natural Resources
Defense Council, http://switchboard.nrdc.org/blogs/mbaumhefner/whv cant your  loan  be  as green.html, accessed
7/29/2015.
58 Comment submitted by Douglas I. Greenhaus, Director, Environment, Health and Safety,  National Automobile
Dealers Association (NADA). Docket EPA-HQ-OAR-2010-0799-9575.
59 77 Federal Register 62950-51; U. S. EPA (2012), "2017 and Later Model Year Light-Duty Vehicle Greenhouse
Gas Emissions and Corporate Average Fuel Economy Standards: EPA Response to Comments," EPA-420-R-12-
017, https://www3.epa.gov/otaq/climate/documents/420rl2017.pdf. Chapter 18.7.1, pp. 18-202 to 18-213.
60 See Bankrate (2015). "Debt-to-income ratio calculator," Bankrate.com,
http://www.bankrate.com/calculators/mortgages/ratio-debt-calculator.aspx: Keythman, Bryan (2015). "What Is the
28/36 Rule of Debt Ratio?" http:^udgeting.thenest.com/28-36-rule-debt-ratio-22412.html: Zillow (2015). "Debt-to-
income calculator," Zillow.com, http://www.zillow.com/mortgage-calculator/debt-to-income-calculator/
61 Deep, Said (1999). "Small in Stature, Big in the Market-Why automakers maintain their small-car focus." Wards
Auto, http://wardsauto.com/news-analvsis/small-stature-big-market-whv-automakers-maintain-their-small-car-
focus, accessed 6/16/2016.
62 Ward's Automotive. '07 [and subsequent, to 2015] Model Year U.S. Car and Light Truck Specifications and
Prices. Accessed 6/16/2015: http://wardsauto.com/data-center see Cassidy, Alecia, Geoffrey Burmeister, and Gloria
Helfand. "Impacts of the Model Year 2017-25 Light-Duty Vehicle Greenhouse Gas Emission Standards on Vehicle
Affordability." Working paper, docket EPA-HQ-OAR-2015-0827.
63 See, for example, Austin, David, and Terry Dinan (2005). "Clearing the Air: The Costs and Consequences of
Higher CAFE Standards and Increased Gasoline." Journal of Environmental Economics  and Management 50(3):
562—82; and Kleit, Andrew N. (2004). "Impacts of Long-Range Increases in the Fuel Economy (CAFE) Standard."
Economic Inquiry 42(2): 279—294.
64 Cassidy, Alecia, Geoffrey Burmeister, and Gloria Helfand (2015). "Impacts of the Model Year 2017-25 Light-
Duty Vehicle Greenhouse Gas Emission Standards on Vehicle Affordability."  Working paper, docket EPA-HQ-
OAR-2015-0827.
                                                    6-27

-------
                                                                    Employment Impacts
Table of Contents

Chapter 7:   Employment Impacts	7-1
  7.1    Introduction	7-1
  7.2    Employment in the Auto Sector in Recent Years	7-1
  7.3    Current State of Knowledge of Employment in the Automotive Sector Based on the
  Peer-Reviewed Literature	7-4
     7.3.1   Regulatory Effects at the Firm Level	7-4
     7.3.2   Regulatory Effects at the Industry Level	7-5
     7.3.3   Peer-Reviewed Literature	7-6
  7.4    Employment Impacts in the Motor Vehicle and Parts Manufacturing Sector	7-7
     7.4.1   The Output Effect	7-7
     7.4.2   The Substitution Effect	7-7
     7.4.3   Summary of Employment Effects in the Motor Vehicle Sector	7-12
     7.4.4   Motor Vehicle Parts Manufacturing Sector	7-12
  7.5    Employment Impacts in Other Affected Sectors	7-12
     7.5.1   Effects on Employment for Auto Dealers	7-12
     7.5.2   Effects on Employment for Fuel Suppliers	7-13
     7.5.3   Effects on Employment due to Impacts on Consumer Expenditures	7-13
  7.6    Summary	7-13

Table of Figures
Figure 7.1 Auto Sector Employment and Production3	7-3
Figure 7.2 Indexed Auto Sector Employment and Production, and Gross Domestic Product (GDP) per Capita,3 2005
           = 100 for all data series	7-3

Table of Tables
Table 7.1 Employment per $1 Million Expenditures (2013$) in the Motor Vehicle Manufacturing Sector3	7-10
Table 7.2 Partial Employment Impact due to Substitution Effect of Increased Costs of Vehicles and Parts, in Job-
           years3	7-11

-------
                                                                   Employment Impacts
Chapter 7: Employment Impacts

7.1    Introduction

   The Presidential Memorandum that requested the agencies to develop the National Program
sought a program that would "strengthen the [auto] industry and enhance job creation in the
United States."1  Executive Order 13563, "Improving Regulation and Regulatory Review"
(January 18, 2011), states, "Our regulatory system must protect public health, welfare, safety,
and our environment while promoting economic growth, innovation, competitiveness, and job
creation."2  In addition, the 2017-25 final rule lists "Impacts on employment, including the auto
sector" as one of the factors to be considered in this Draft TAR.3  Although analysis of
employment impacts is not part of a cost-benefit analysis (except to the extent that labor costs
contribute to costs), EPA is accordingly providing this discussion of the potential employment
effects of the standards. This section begins with an overview of employment in the auto
industry in recent years, and then discusses estimating the employment effects of the standards.
While the 2022-2025 standards may have some effect on employment in the auto sector, this
effect is likely to be small enough that it cannot be distinguished from other factors affecting
auto sector employment.

7.2    Employment in the Auto Sector in Recent Years

   Figure 7.1 shows employment in three segments of the U.S. auto industry from 2005 through
2014: Motor Vehicles; Motor Vehicle Parts; and Automobile Dealers. The Motor Vehicle sector
itself, which includes the major manufacturers, employs the fewest people of these three sectors;
Motor Vehicle Parts, suppliers to the auto industry, employs roughly two to three times as many
people, and the Automobile Dealers sector employs more than the sum of the manufacturing and
parts sectors.

   As this chart shows, in all three segments, employment was decreasing before the recession
began in 2009, and has been increasing in recent years with recovery from the recession. Auto
dealers had a smaller percentage decrease than Motor Vehicles or Motor Parts, though all  have
recovered back to employment levels of 2007-2008 by 2014.

   Figure 7.1 includes  vehicle salesA during this period (see also Chapters 3 and 6.1); it shows a
similar overall pattern  of decrease followed by increase, though sales have increased more
rapidly on a percentage basis than employment since 2009 (see Figure 7.2).  The similarities in
the patterns for sales and employment suggest, unsurprisingly, that one of the key drivers of
employment in auto-related sectors is vehicle production. Indeed, the American Automotive
Policy Council cites a prediction from the Center for Automotive Research that auto employment
will increase by more than a third from 2011 to 2016, as production of vehicles in the U.S.
increases from 5.8 million in 2009 to at least 11.5 million vehicles in 2016,4 and total sales
reached a record high of 17.5 million in 2015.5  The differences in changes in magnitude for
employment compared to sales may be due to a number of factors; one of those factors may be
A Vehicle production data represent production volumes delivered for sale in the U.S. market, rather than actual
  sales data. They include vehicles built overseas imported for sale in the U.S., and exclude vehicles built in the
  U.S. for export.
                                              7-1

-------
                                                                   Employment Impacts
changes in the production process and in productivity; another factor might be the GHG/fuel
economy standards.

   The effects of the standards on employment are difficult to identify.  As Chapter 6.1
discusses, it is difficult, if not impossible, to disentangle the effects of the standards on vehicle
production (or employment) from changes in other factors, especially the state of the
macroeconomy. Figure 7.2 shows the same employment sectors and production as in Figure 7.1,
now indexed to show each value as a percent of its value in 2005; it also includes Gross
Domestic Product (GDP) per capita.6 This figure suggests that auto sector production and
employment declined earlier and more deeply than the economy  as a whole, and rebounded more
vigorously.

   EPA's Regulatory Impact Analysis for the MY2017-25 light-duty vehicle standards included a
discussion of the effects of the standards on employment in the automotive and directly related
sectors (e.g., the parts sector) (see Chapter 8.2).6  It did not quantify the overall net effects of the
standards on U.S employment. Nor did it quantify the effects of the standards on vehicle sales,
and thus did not quantify the effects of employment changes in these sectors due to changes in
vehicle sales. It did provide partial estimates of the effects of increased expenditures on
employment in these sectors: some of those increased expenditures would be on labor. Those
estimates were provided to suggest the magnitude of employment impacts, even though they
were only one pathway through which employment in these sectors would be affected. It
estimated increases on the order of 700 to 3,200 jobs in 2017 (p.  8-28) due to those expenditures,
with the range dependent on whether the increased expenditures  occurred in the light duty
vehicle manufacturing sector or the parts sector. Given levels of employment in the auto sector in
2015, this increase would be less than 1  percent of employment in the auto sector, and it does not
account for any effects of the standards on vehicle sales.  As Figure  7.1 and Figure 7.2 suggest,
employment is likely to vary much more than that proportion due to macroeconomic factors.
Thus, while the MY2012-16 standards are likely to have had some effect on employment in the
auto sector, this effect is likely to have been small enough that it cannot be distinguished from
other factors affecting auto sector employment. In addition, the standards are not expected to
have had any notable inflationary or recessionary effect.
 ' Graphing in this way facilitates comparison of percentage changes in the data series compared to 2005.
                                              7-2

-------
                                                                             Employment Impacts
                               •Motor vehicles empl

                               •Automobile dealers empl
             •Motor vehicle parts empl

              LDV production
                0                                                                      0.00
                    2005   2006  2007  2008  2009  2010  2011  2012  2013  2014  2015
                          Figure 7.1 Auto Sector Employment and Production"

Note:aEmployment data are from http://www.bls.gov/iag/tgs/iagauto.htm. Production data are for model years,
from U.S. EPA 2015.7 Note that 2015 production data are projected, not actual, values.
             no
                   2005   2006   2007   2008   2009   2010   2011   2012    2013    2014    2015

                  -Motor vehicles empl    	Motor vehicle parts empl —  —Automobile dealers empl
                    LDV production
-Real GDP/Capita
   Figure 7.2 Indexed Auto Sector Employment and Production, and Gross Domestic Product (GDP) per
                                 Capita,3 2005 = 100 for all data series.

Note:aEmployment data are from http://www.bls.gov/iag/tgs/iagauto.htm. Production data are for model years,
from U.S. EPA 2015.8 Note that 2015 production data are projected, not actual, values. GDP per capita data are
found at https://research.stlouisfed.org/fred2/series/A939RXOQ048SBEA/downloaddata.
                                                     7-3

-------
                                                                   Employment Impacts
7.3    Current State of Knowledge of Employment in the Automotive Sector
Based on the Peer-Reviewed Literature

   As suggested in the previous section, the employment effects of environmental regulation are
difficult to disentangle from other economic changes and business decisions that affect
employment, over time and across regions and industries. In light of these difficulties, we look
to economic theory to provide a constructive framework for approaching these assessments and
for better understanding the inherent complexities in such assessments.

   If the U.S. economy is at full employment, even a large-scale environmental regulation is
unlikely to have a noticeable impact on aggregate net employment.0 Instead, labor would
primarily be reallocated from one productive use to  another, and net national employment effects
from environmental regulation would be small and transitory (e.g., as workers move from one
job to another).9

   Affected sectors may experience transitory effects as workers change jobs.  Some workers
may retrain or relocate in anticipation of new requirements or require time to search for new
jobs, while shortages in some sectors or regions could bid up wages to attract workers.  These
adjustment costs can lead to local labor disruptions.  Although the net change in the national
workforce is expected to be small, localized reductions in employment may adversely impact
individuals and communities just as localized increases may have positive impacts.

   If the economy is operating  at less than full employment, economic theory does not clearly
indicate the direction or magnitude of the net impact of environmental regulation on
employment; it could cause either a short-run net increase or short-run net decrease.10 An
important research question is how to accommodate unemployment as a structural feature in
economic models.  This may be important in assessing large-scale regulatory impacts on
employment.11

   Environmental regulation may also affect labor supply. In particular, pollution and other
environmental risks may impact labor productivity or employees' ability to work.12 While the
theoretical framework for analyzing labor supply effects is analogous to that for labor demand, it
is more difficult to study empirically. There is a small emerging literature described in the next
section that uses detailed labor and environmental data to assess these impacts.

7.3.1  Regulatory Effects at the Firm Level

   Neoclassical microeconomic theory provides insights into how profit-maximizing firms adjust
their use of productive inputs in response to changes in their economic conditions.13 Berman and
Bui (2001, pp. 274-75) model two components that drive changes in firm-level labor demand:
output effects and substitution effects.14,0  Regulation can affect the profit-maximizing quantity
of output by changing the marginal cost of production.  If regulation causes marginal cost to
c Full employment is a conceptual target for the economy where everyone who wants to work and is available to do
  so at prevailing wages is actively employed. The unemployment rate at full employment is not zero.
D Berman and Bui also discuss a third component, the impact of regulation on factor prices, but conclude that this
  effect is unlikely to be important for large competitive factor markets, such as labor and capital. Morgenstern,
  Pizer and Shih (2002) use a very similar model, but they break the employment effect into three parts: 1) a
  demand effect; 2) a cost effect; and 3) a factor-shift effect.
                                              7-4

-------
                                                                    Employment Impacts
increase, it will place upward pressure on output prices, leading to a decrease in the quantity
demanded, and resulting in a decrease in production. The output effect describes how, holding
labor intensity constant, a decrease in production causes a decrease in labor demand.  As noted
by Berman and Bui, although many assume that regulation increases marginal cost, it need not
be the case.  A regulation could induce a firm to upgrade to less polluting and more efficient
equipment that lowers marginal production costs, or it may induce use of technologies that may
prove popular with buyers or provide positive network externalities (see Chapter 6.3 for
discussion of this effect).  In such a case, output could increase.
   The substitution effect describes how, holding output constant, regulation affects labor-
intensity of production. Although increased environmental regulation may increase use of
pollution control equipment and energy to operate that equipment, the impact on labor demand is
ambiguous.  For example, equipment inspection requirements,  specialized waste handling, or
pollution technologies that alter the production process may affect the number of workers
necessary to produce a unit of output. Berman and Bui (2001)  model the substitution effect as
the effect of regulation on pollution control equipment and expenditures required by the
regulation and the corresponding change in labor-intensity of production.

   In summary, as output and substitution effects may be positive or negative, theory alone
cannot predict the direction of the net effect of regulation on  labor demand at the level of the
regulated firm.  Operating within the bounds of standard economic theory, however, empirical
estimation of net employment effects on regulated firms is possible when data and methods of
sufficient detail and quality are available.  The literature, however, illustrates difficulties with
empirical estimation. For example, studies sometimes rely on confidential plant-level
employment data from the U.S. Census Bureau, possibly combined with pollution abatement
expenditure data that are too dated to be reliably informative. In addition, the most commonly
used empirical methods do not permit estimation of net effects.

7.3.2  Regulatory Effects at the Industry Level

   The conceptual framework described thus far focused on regulatory effects on plant-level
decisions within a regulated industry. Employment impacts at an individual plant do not
necessarily represent impacts for the sector as a whole. The approach must be modified when
applied at the industry level.

   At the industry level, labor demand is more responsive if:  (1) the price elasticity of demand
for the product is high, (2) other factors of production can be easily substituted for labor, (3) the
supply of other factors is highly elastic, or (4) labor costs are a  large share of total production
costs.15  For example, if all firms in an industry are faced with the same regulatory compliance
costs and product demand is inelastic, then industry output may not change much, and output of
individual firms may change slightly.16  In this case, the output effect may be small, while the
substitution effect depends on input substitutability. Suppose, for example, that new  equipment
for GHG emissions reductions requires labor to install and operate. In this case, the substitution
effect may be positive, and with a small output effect, the total  effect may be positive. As with
potential effects for an individual firm, theory cannot determine the sign or magnitude of
industry-level regulatory effects on labor demand.  Determining these signs and magnitudes
requires additional sector-specific empirical study.  For environmental rules, much of the data
needed for these empirical studies is not publicly available.
                                              7-5

-------
                                                                   Employment Impacts
   In addition to changes to labor demand in the regulated industry, net employment impacts
encompass changes in other related sectors. For example, the standards are expected to increase
demand for fuel-saving technologies. This increased demand may increase revenue and
employment in the firms supporting this technology. At the same time, the regulated industry is
purchasing the equipment, and these costs may impact labor demand at regulated firms.
Therefore, it is important to consider the net effect of compliance actions on employment across
multiple sectors or industries.

   Affected sectors may  experience transitory effects as workers change jobs.  Some workers
may retrain or relocate in anticipation of new requirements or require time to search for new
jobs, while shortages in some sectors or regions could bid up wages to attract workers.  These
adjustment costs can lead to local labor disruptions. Although the net change in the national
workforce is expected to be small, localized reductions in employment may adversely impact
individuals and communities just as localized increases may have positive impacts.

   To summarize, economic theory provides a framework for analyzing the impacts of
environmental regulation on employment. The net employment effect incorporates expected
employment changes (both positive and negative) in the regulated sector and elsewhere. Labor
demand impacts for regulated firms, and also for the regulated industry, can be decomposed into
output and  substitution effects which may be either negative or positive. Estimation of net
employment effects for regulated sectors is  possible when data of sufficient detail and quality are
available.  Finally, economic theory suggests that labor supply effects are also possible. In the
next section, we discuss the empirical literature.

7.3.3  Peer-Reviewed Literature

   In the labor economics literature there is an extensive body of peer-reviewed empirical work
analyzing various aspects of labor demand,  relying on the above theoretical framework.17 This
work focuses primarily on the effects of employment policies, e.g. labor taxes, minimum wage,
etc.18 In contrast, the peer-reviewed empirical literature specifically estimating employment
effects of environmental regulations is very limited. Several empirical studies, including Berman
and Bui (2001),19  Morgenstern, Pizer and Shih (2002),20  Gray et al (2014),21 and Ferris,
Shadbegian and Wolverton (2014)22 suggest that net employment impacts may be zero or
slightly positive but small even in the regulated sector. Other research suggests that more highly
regulated counties may generate fewer jobs than less regulated ones.23  However, since these
latter studies compare more regulated to less regulated counties, they overstate the net national
impact of regulation to the extent that regulation causes plants to locate in one area of the
country rather than another. List  et al. (2003)24 find some evidence that this type of geographic
relocation may be occurring.  Overall, the peer-reviewed  literature does not contain evidence that
environmental regulation has a large impact on net employment (either negative or positive) in
the long run across the whole economy.

   Analytic challenges make it very difficult to accurately produce net employment estimates for
the whole economy that would appropriately capture the way  in which costs, compliance
spending, and environmental benefits propagate through the macro-economy. Quantitative
estimates are further complicated by the fact that macroeconomic models often have very little
sectoral detail and usually assume that the economy is at  full employment.  EPA is currently in
the process of seeking input from an independent expert panel on modeling economy-wide
impacts, including employment effects. For more information, see:
                                              7-6

-------
                                                                   Employment Impacts
ocument.

7.4    Employment Impacts in the Motor Vehicle and Parts Manufacturing
Sector

   This chapter describes estimated changes in employment in the motor vehicle, trailer, and
parts (hence, motor vehicle) manufacturing sectors associated with the MY2022-25 standards.
We focus on the motor vehicle manufacturing sector because it is directly regulated by the
GHG/fuel economy  standards, and because it is likely to bear most of any employment changes
due to the standards. We include discussion of effects on the parts manufacturing sector,
because the motor vehicle manufacturing sector can either produce parts internally or buy them
from an external supplier, and we do not have estimates of the likely breakdown of effort
between the two sectors.

   We follow the theoretical structure of Berman and Bui 25 of the impacts of regulation in
employment in the regulated sectors. In Berman and Bui's (2001, p. 274-75) theoretical model,
as described above, the change in a firm's labor demand arising from a change in regulation is
decomposed into two main components: output and substitution effects.  As the output and
substitution effects may be both positive, both negative, or some combination, standard
neoclassical theory alone does not point to a definitive net effect of regulation on labor demand
at regulated firms.

   Following the Berman and Bui framework for the impacts of regulation on employment in the
regulated sector, we consider two effects for the motor vehicle sector: the output effect and the
substitution effect.

7.4.1   The Output Effect

   The output effect measures the effect due to new vehicle sales only.  If vehicle sales increase,
then more people will be required to assemble vehicles and their components. If vehicle sales
decrease, employment associated with these activities will decrease. The effects of the MY2022-
25 standards on vehicle sales thus depend on the perceived desirability of the new vehicles
relative to other transportation options. On one hand, these  standards will increase vehicle  costs;
by itself, this effect would reduce vehicle sales. In addition,  while adverse effects on other
vehicle characteristics would also decrease sales, there is currently no evidence of systematic
adverse effects of fuel-saving technologies (see Chapter 6.3). On the other hand, these standards
will reduce the fuel costs of operating the vehicles; by itself, this effect would increase vehicle
sales, especially if potential buyers have an expectation of increasing fuel prices.  EPA has  not
made an estimate of the effects of the standards on vehicles  sales (see Chapter 6.1).

7.4.2   The Substitution Effect

   The substitution effect includes the impacts due to the changes in technologies needed for
vehicles to meet the  standards, separate from the effect due to vehicle sales (that is, as though
holding output constant). This effect includes both changes in employment due to incorporation
of abatement technologies and overall changes in the labor intensity of manufacturing. We here
capture these effects using estimates of the historic share of labor as a part of the cost of
production, which we then extrapolate to provide future estimates of the share of labor as a cost
                                              7-7

-------
                                                                    Employment Impacts
of production. When these shares are multiplied by the change in the cost of production, they
approximate the change in labor associated with the cost increases associated with the standards.
We present estimates for this effect to provide a sense of the order of magnitude of expected
impacts on employment, which we expect to be small in the automotive sector, and to repeat that
regulations may have positive as well as negative effects on employment.

   One way to estimate this effect, given the cost estimates for complying with the rule, is to use
the ratio of workers to each $1  million of expenditures in that sector.  The use of these ratios has
both advantages and limitations.  It is often possible to estimate these ratios for quite specific
sectors of the economy: for instance, it is possible to estimate the average number of workers in
the light-duty vehicle manufacturing sector per $1 million  spent in the sector, rather than use the
ratio from another, more aggregated sector, such as motor vehicle manufacturing. As a result, it
is not necessary to extrapolate employment ratios from possibly unrelated sectors.  On the other
hand, these estimates are averages for the sectors, covering all the activities in those sectors; they
may not be representative of the labor required when expenditures are required on specific
activities, or when manufacturing processes change sufficiently that labor intensity changes. For
instance, the ratio for the motor vehicle manufacturing sector represents the ratio for all vehicle
manufacturing, not just for emissions reductions associated with compliance activities. In
addition, these estimates do not include changes in sectors that supply these sectors, such as steel
or electronics producers. They thus may best be viewed as the effects on employment in the auto
sector due to the changes in expenditures in that sector, rather than as an assessment of all
employment changes due to these changes in expenditures. In addition, this approach estimates
the effects of increased expenditures while holding constant the labor intensity of manufacturing;
it does not take into account changes in labor intensity due to changes in the nature of
production. This latter effect could either increase or decrease the employment impacts
estimated here.E

   Some of the costs of this rule will be spent directly in the motor vehicle manufacturing sector,
but it is  also likely that some of the costs will be spent in the motor vehicle parts manufacturing
sector. The analysis here draws on estimates of workers per $1 million of expenditures for both
of these sectors.

   There are several public sources for estimates of employment per $1 million expenditures.
The U.S. Bureau of Labor  Statistics (BLS) provides its Employment Requirements Matrix
(ERM),26 which provides direct estimates of the employment per $1 million in sales of goods in
202 sectors.  The values considered here are for Motor Vehicle Manufacturing (NAICS 3361)
and Motor Vehicle Parts Manufacturing (NAICS  3363) for 2014. These values are updated from
the 2012 FRM, which used the 2010 ERM data.

   The U.S. Census Bureau provides both the Annual Survey of Manufacturers27 (ASM) and the
Economic Census (EC). The ASM is a subset of the Economic Census, based on a sample of
establishments; though the Census itself is more complete, it is conducted only every 5 years,
while the ASM is annual.  Both include more sectoral detail than the BLS ERM: for instance,
while the ERM includes the Motor Vehicle Manufacturing sector, the ASM and EC have detail
E As noted above, Morgenstern et al. (2002) separate the effect of holding output constant into two effects: the cost
  effect, which holds labor intensity constant, and the factor shift effect, which estimates those changes in labor
  intensity.
                                               7-8

-------
                                                                     Employment Impacts
at the 6-digit NAICS code level (e.g., light truck and utility vehicle manufacturing). While the
ERM provides direct estimates of employees/Si million in expenditures, the ASM and EC
separately provide number of employees and value of shipments; the direct employment
estimates here are the ratio of those values.  The values reported are for Motor Vehicle
Manufacturing (NAICS 3361), Automobile and Light Duty Motor Vehicle Manufacturing
(NAICS 33611), and Motor Vehicle Parts Manufacturing (NAICS 3363), for 2014 for the ASM
and 2012 for the EC.  These values are updated from the 2012FRM, which used 2010 values for
the ASM, and 2007 values from the EC.

   The values used here are adjusted to remove the employment effects of imports through use of
a ratio of domestic production to domestic sales of 0.663.F

   Table 7.1 provides the values, either given (BLS) or calculated (ASM and EC) for
employment per $1 million of expenditures in 2014 (2012 for EC),  all adjusted to 2013 dollars
using the Bureau of Economic Analysis's Implicit GDP Price Deflators.0 Although the ASM
appears to provide slightly higher values than the ERM, the different data sources provide
similar patterns for the estimates for the sectors. These updated values differ slightly (under 10
percent) from the values used in the 2012 FRM in 2013$.
F To estimate the proportion of domestic production affected by the change in sales, we use data from Ward's
  Automotive Group for total car and truck production in the U.S. compared to total car and truck sales in the U.S.
  Over the period 2006-2015, the proportion averages 66.3 percent. From 2012-2015, the proportion average is
  slightly higher, at 69.2 percent.
G At the time of access, the EC data was only available by 2-, 3-, or 6-digit NAICS industry code. To construct the
  4- and 5-digit numbers, we separately summed total employees and total expenditure for each 6-digit
  subcategory.
                                               7-9

-------
                                                                   Employment Impacts
  Table 7.1 Employment per $1 Million Expenditures (2013$) in the Motor Vehicle Manufacturing Sector3
Source
BLS ERM
BLS ERM
ASM
ASM
ASM
ASM
EC
EC
EC
EC
Sector
Motor vehicle mfg (3361)
Motor vehicle parts mfg (3363)
Motor vehicle mfg (3361)
Automobile and light duty motor vehicle
mfg (33611)
Automobile mfg (336111)
Motor vehicle [arts mfg (3363)
Motor vehicle mfg (3361)
Automobile and light duty motor vehicle
mfg (33611)
Automobile mfg (336111)
Motor vehicle parts mfg (3363)
Ratio of
workers per $1
million
expenditures
0.39
1.71
0.58
0.54
0.63
2.08
0.59
0.55
0.63
2.13
Ratio of workers per $1
million expenditures,
adjusted for domestic vs.
foreign production
0.26
1.13
0.39
0.36
0.42
1.38
0.39
0.36
0.42
1.41
   Note:
   a BLS ERM refers to the U.S. Bureau of Labor Statistics' Employment Requirement Matrix, 2014 values. ASM
   refers to the U.S. Census Bureau's Annual Survey of Manufactures, 2014 values. EC refers to the U.S. Census
   Bureau's Economic Census, 2012 values.

   Over time, the amount of labor needed in the motor vehicle industry has changed: automation
and improved methods have led to significant productivity increases. The BLS ERM, for
instance,  provided estimates that, in 1997, 1.09 workers in the Motor Vehicle Manufacturing
sector were needed per $1 million, but only 0.39 workers by 2014 (in 2013$).28 Because the
ERM is available annually for 1997-2014, we used these data to estimate productivity
improvements over time.  We regressed logged ERM values on a year trend for the Motor
Vehicle Manufacturing and Motor Vehicle Parts Manufacturing sectors. We used this approach
because the coefficient describing the relationship between time and productivity is a direct
measure of the average percent change in productivity per year. The results suggest a 6.6 percent
per year productivity improvement in the Motor Vehicle Manufacturing Sector, and a 4.9 percent
per year improvement in the Motor Vehicle Parts Manufacturing Sector.

   We then used the regression results to project the number of workers per $1 million through
2025.  We calculated separate sets of projections (adjusted to 2013$) for both the BLS ERM data
as well as the EC and ASM for all sectors discussed above. The BLS ERM projections were
calculated directly from the fitted regression equations since the regressions themselves used
ERM data.  For the ASM and EC projections, we used the ERM's ratio of the projected value in
each future year to the projected value in 2014 for the ASM and 2012 for the EC (the base years
in our data) to determine how many workers will be needed per $1 million of 2013$. In other
words, we apply the projected productivity growth estimated using the ERM data to the ASM
and EC numbers.

   Finally, to simplify the presentation and give a range of estimates, we compared the projected
employment among the sectors for the ERM, EC, and ASM, and we provide here only the
maximum and minimum effects in each year across all sectors.  We provide the range rather than
a point estimate because of the inherent difficulties in estimating employment impacts; the range
                                              7-10

-------
                                                                    Employment Impacts
gives an estimate of the expected magnitude. The details of the calculations may be found in the
docket.  The Motor Vehicle Parts Manufacturing Sector value from the ASM provides the
maximum employment estimates per $1 million; the Motor Vehicle Manufacturing Sector value
from the ERM provides the minimum estimates.

   Chapter 12 of this Draft TAR discusses the vehicle cost estimates developed for this rule.  The
final step in estimating employment impacts is to multiply costs (in $ millions) by workers per
$1 million in costs, to estimate employment impacts in the regulated and parts manufacturing
sectors.  Table 7.2  presents the projected reference case costs and the corresponding minimum
and maximum estimated employment impacts. For each year, additional ranges in parentheses
are included that reflect estimates from projections using high and low fuel price scenarios.11
Increased costs of vehicles and parts, by itself, and holding labor intensity constant, would be
expected to increase employment between 2021 and 2025 by several hundred to 12,000 jobs
each year.  These values are lower than those estimated in the 2012 FRM, primarily because the
cost estimates are lower, for reasons explained in Chapter  12.

   While we estimate employment impacts, measured in job-years, beginning with program
implementation, some of these employment gains may occur earlier as vehicle manufacturers
and parts suppliers hire staff in anticipation of compliance with the standards. A job-year is a
way to calculate the amount of work needed to complete a specific task.  For example, a job-year
is one year of full-time work for one person.
Table 7.2 Partial Employment Impact due to Substitution Effect of Increased Costs of Vehicles and Parts, in
                                        Job-years"
Year
2021
2022
2023
2024
2025
Costs (Millions of
2013$)
$3,045
($2,872 - 2,876)
$5,877
($5,766 - $5,769)
$8,736
($8,620 - $8,709)
$11,649
($11,483 - $11,727)
$14,678
($14,433 -$14,871)
Minimum Employment Due to
Substitution Effect (ERM
estimates, expenditures in the
Motor Vehicle Mfg Sector)
300
(300 - 300)
600
(600 - 600)
800
(800 - 800)
1,000
(1,000 - 1,100)
1,200
(1,200 - 1,300)
Maximum Employment Due to
Substitution Effect (ASM estimates,
expenditures in the Parts Sector)
3,000
(2,800 - 2,800)
5,500
(5,400 - 5,400)
7,800
(7,700 - 7,700)
9,800
(9,700 - 9,900)
11,800
(11,600 - 12,000)
Note:
a Numbers in parentheses reflect the estimates derived from scenarios with high and low fuel prices.
H As discussed in Chapter 12, the costs for the reference fuel price scenario do not necessarily fall between those of
  the high and low fuel price scenarios, because fuel prices are not the only difference in the scenarios; they differ
  in assumptions about the vehicle fleet as well.
                                              7-11

-------
                                                                    Employment Impacts
7.4.3   Summary of Employment Effects in the Motor Vehicle Sector

   The overall effect of the rule on motor vehicle sector employment depends on the relative
magnitude of the output effect and the substitution effect. Because we do not have quantitative
estimates of the output effect, and only a partial estimate of the substitution effect, we cannot
reach a quantitative estimate of the overall employment effects of the standards on auto sector
employment or even whether the total effect will be positive or negative.

   The standards are not expected to provide incentives for manufacturers to shift employment
between domestic and foreign production. This is because the standards will apply to vehicles
sold in the U.S. regardless of where they are produced.  Ward's automotive data suggest that the
current share  of domestic production for cars and trucks is very similar to the  share in 2006: 66
percent in 2006, and 68 percent in 2015. If production  overseas already involved increased
expertise in satisfying the requirements  of the standards, there may be some initial incentive for
foreign production, but meeting the standards may lead to increased opportunities for domestic
production to sell in other markets. To the extent that the requirements of these standards might
lead to installation and use of technologies that other countries may seek now or in the future,
developing this capacity for domestic production now may provide some additional ability to
serve those markets.

7.4.4   Motor Vehicle Parts Manufacturing  Sector

   Some vehicle parts are made in-house and would be included directly in the regulated sector.
Others are made by independent suppliers and are not directly regulated, but they will be affected
by the rules as well. The parts manufacturing sector will be involved primarily in providing
"add-on" parts, or components  for replacement parts built internally. If demand for these parts
increases due to the increased use of these parts, employment effects in this sector are expected
to be positive. If the output effect in the regulated sectors is significantly negative enough, it is
possible that demand for other parts may decrease. As  noted, the agencies do not predict a
magnitude or direction for the output effect.

7.5    Employment Impacts in Other Affected Sectors

7.5.1   Effects on Employment for Auto Dealers

   The effects of the standards  on employment for auto dealers depend principally on the effects
of the standards on light duty vehicle sales: increases in sales are likely to contribute to
employment at dealerships, while reductions in sales  are likely to have the opposite effect.  As
discussed in Chapter 6, it is difficult to separate the effects of the standards on vehicle sales from
effects due to macroeconomic conditions; however, the standards have not prevented sales from
returning to (and exceeding)  pre-recession levels.  In addition, auto dealers may be affected by
any changes in maintenance and service costs.  Increases in those costs are likely to increase
labor demand in dealerships, and reductions are likely to decrease labor demand.

   Concerns have been raised about consumer acceptance of technologies used to meet the
standards, though these effects  do not seem significant to date (see Chapter 6). Auto dealers may
play a major role in explaining the merits and disadvantages of these new technologies to vehicle
buyers.  This  additional role may also affect employment levels at dealers.
                                              7-12

-------
                                                                   Employment Impacts
7.5.2   Effects on Employment for Fuel Suppliers

   In addition to the effects on the auto manufacturing and parts sectors, the standards result in
changes in fuel use that lower GHG emissions.

   Expected petroleum fuel consumption reductions can be found in Chapter 12.  While this
reduced consumption represents fuel savings for purchasers of fuel, it represents a loss in value
of output for the petroleum refinery industry, fuel distributors, and gasoline stations. The loss of
expenditures to petroleum fuel suppliers throughout the petroleum fuel supply chain, from the
petroleum refiners to the gasoline stations, is likely to result in reduced employment in these
sectors. Because the fuel production sector is material-intensive, the employment effect is not
expected to be large.1 Although gasoline stations will sell less fuel, the fact that many provide
other goods, such as food and car washes, moderates losses in this sector. In addition, it may be
difficult to distinguish these effects from other trends, such as increases in petroleum sector labor
productivity that may also lower labor demand.

   Auto manufacturers may choose to meet the standards through alternatively-fueled vehicles,
such as those that use electricity, hydrogen, or compressed natural gas (CNG), though the
agencies do not project large use of these vehicles.  Such fuels may require additional
infrastructure, such as electricity charging locations or hydrogen fueling stations.  See Chapter 9.
Providing this infrastructure will require some increased employment. In addition, the production
of these fuels is likely to require some additional labor. We have  insufficient information at this
time to predict whether the increases in labor associated with increased infrastructure provision
and generation for electricity and hydrogen production will be greater or less than the
employment reductions associated with reduced demand for petroleum fuels.

7.5.3   Effects on Employment due to Impacts on Consumer Expenditures

   As a result of these standards, consumers will likely pay higher up-front costs for the vehicles,
but they are expected to recover those costs in a fairly short payback period (see Chapters 6 and
12). As a result, consumers are expected to have additional money to spend on other goods and
services, though the timing for access to that additional money depends on the payback period
and whether the consumer borrows money to buy the vehicle. These increased expenditures
could support employment in those sectors where consumers spend their savings.

   These increased expenditures will occur in the years in which the fuel savings exceed
expenditures on the up-front costs. If, on the one hand, the economy is at full employment
during that time, any change in consumer expenditures would primarily represent a shift in
employment among sectors.  If, on the other hand, the economy has substantial unemployment,
these expenditures would contribute to employment through increased consumer demand.

7.6     Summary

   The primary employment effects of these standards are expected to be found in several key
sectors: auto manufacturers, auto parts manufacturing, auto dealers, fuel production and supply,
and consumers. In an economy with full employment, the primary employment effect of a
1 In the 2014 BLS ERM cited above, the Petroleum and Coal Products Manufacturing sector has a ratio of workers
  per$l million of 0.215, lower than all but two of the 181 sectors with non-zero employment per $1 million.
                                              7-13

-------
                                                                   Employment Impacts
rulemaking is likely to be to shift employment from one sector to another, rather than to increase
or decrease employment.  For that reason, we focus our partial quantitative analysis on
employment in the regulated sector, to examine the impacts on that sector directly. We discuss
the likely direction of other impacts in the regulated sector as well  as in other directly related
sectors, but we do not quantify those impacts, because they are more difficult to quantify with
reasonable accuracy, particularly so far into the future.

   For the regulated sector, the partial employment impact due to the substitution effect of
increased costs of autos is expected to be positive.  The total effect of the standards on motor
vehicle employment depends in addition on changes in vehicle sales, which are not quantified;
thus, we do not estimate the total effects of the standards in the regulated industry.
   Effects in other sectors that are affected by vehicle sales are also ambiguous.  Reduced
petroleum fuel production implies less employment in the petroleum sectors, although there
could be increases in employment related to providing infrastructure for alternative fuels if
manufacturers choose to comply with the standard through increased production of vehicles that
use those fuels. Finally, consumer spending is expected to affect employment through changes
in expenditures in general retail sectors; net fuel savings by consumers are expected to increase
demand (and therefore employment) in other sectors.  Thus, while the standards are likely to
have some effect on employment, this effect is likely to be small enough that it cannot be
distinguished from other factors affecting employment, especially macroeconomic conditions.
As has been noted, under  conditions of full employment, any changes in employment levels in
the regulated sector due to this program are mostly expected to be offset by  changes in
employment in other sectors.
                                              7-14

-------
                                                                            Employment Impacts
   References
1 President Barack Obama. "Presidential Memorandum Regarding Fuel Efficiency Standards. The White House,
Office of the Press Secretary, May 21, 2010. http://www.whitehouse.gov/the-press-office/presidential-
memorandum-regarding-fuel-efficiency-standards.
2 President Barack Obama. "Executive Order 13563 of January 18, 2011: Improving Regulation and Regulatory
Review." Federal Register 76(14) (January 21, 2011): 3821-3823.
3 77 Federal Register 62784.
4 American Automotive Policy Council  (2015). "State of the U.S. Automotive Industry: Investment, Innovation,
Jobs, Exports, and America's Economic Competitiveness." http://americanautocouncil.org/sites/default/files/2015-
AAPC-Economic-Contribution-Report%28FINAL%29.pdf.
5 Woodall, Bernie (2015). "U.S. Auto Sales in 2015 Set Record after Strong December." Reuters,
Mtei/Mwwjiuteisxoi^^                                        , accessed 2/12/2016.
6 Environmental Protection Agency (EPA) (2012). Regulatory Impact Analysis: Final Rulemaking for 2017-25
Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards,"
Chapter 8.2. EPA-420-R-12-016, htte://www3Jepa1gQy/Qtag/c!ir^^                        .
7 Environmental Protection Agency (EPA) 2015. Light-Duty Automotive Technology, Carbon Dioxide Emissions,
and Fuel Economy Trends: 1975 through 2015. U.S. EPA-420-R-15-001, Office of Transportation and Air Quality,
December 2015.
8 Environmental Protection Agency (EPA) 2015. Light-Duty Automotive Technology, Carbon Dioxide Emissions,
and Fuel Economy Trends: 1975 through 2015. U.S. EPA-420-R-15-001, Office of Transportation and Air Quality,
December 2015.
9 Arrow et al. (1996). "Benefit-Cost Analysis in Environmental, Health, and Safety Regulation: A Statement of
Principles." American Enterprise Institute, The Annapolis Center, and Resources for the Future. Docket EPA-HQ-
OAR-2014-0827-0073.  See discussion on bottom of p. 6. In practice, distributional impacts on individual workers
can be important, as discussed later in this section.
10 Schmalensee, Richard, and Robert N.  Stavins. "A Guide to Economic and Policy Analysis of EPA's Transport
Rule." White paper commissioned by Excelon Corporation,  March 2011. Docket EPA-HQ-OAR-2014-0827-0071.
11 Klaiber, H. Allen, and V. Kerry Smith (2012). "Developing General Equilibrium Benefit Analyses for Social
Programs: An Introduction and Example." Journal of Benefit-Cost Analysis 3(2). Docket EPA-HQ-OAR-2014-
0827-0085.
12 Graff Zivin, J., and M. Neidell (2012). "The Impact of Pollution on Worker Productivity." American Economic
Review 102: 3652-3673. Docket EPA-HQ-OAR-2014-0827-0092.
13 Layard, P.R.G., and A. A. Walters (1978). Microeconomic Theory (McGraw-Hill, Inc.), Chapter 9, "The Derived
Demand for Factors."
14 Berman, E. and L. T. M. Bui (2001). "Environmental Regulation and Labor Demand: Evidence from the South
Coast Air Basin." Journal of Public Economics 79(2): 265-295. Docket EPA-HQ-OAR-2014-0827-0086.
15 Ehrenberg, Ronald G., and  Robert S. Smith (2000). Modern Labor Economics: Theory and Public Policy
(Addison Wesley Longman, Inc.), p. 108. Docket EPA-HQ-OAR-2014-0827-0077.
16 This  discussion draws from Berman, E. and L. T. M. Bui (2001). "Environmental Regulation and Labor Demand:
Evidence from the South Coast Air Basin." Journal of Public Economics 79(2): 265-295, p. 293. Docket EPA-HQ-
OAR-2014-0827-0074.
17 Hamermesh (1993). Labor Demand (Princeton, NJ: Princeton University Press), Chapter 2. Docket EPA-HQ-
OAR-2014-0827-0082.
18 Ehrenberg, Ronald G., and  Robert S. Smith (2000). Modern Labor Economics: Theory and Public Policy.
Addison Wesley Longman, Inc., Chapter 4. Docket EPA-HQ-OAR-2014-0827-0077.
19 Berman, E. and L. T. M. Bui (2001). "Environmental Regulation and Labor Demand: Evidence from the South
Coast Air Basin." Journal of Public Economics 79(2): 265-295. Docket EPA-HQ-OAR-2014-0827-0086.
20 Morgenstern, Richard D., William A.  Pizer, and Jhih-Shyang Shih (2002). "Jobs Versus the Environment: An
Industry-Level Perspective." Journal of Environmental Economics and Management 43: 412-436. Docket EPA-HQ-
OAR-2014-0827-0088.
21 Gray, Wayne B., Ronald J.  Shadbegian, Chunbei Wang, and Merve Meral (2014). "Do EPA Regulations Affect
Labor Demand? Evidence from the Pulp and Paper Industry." Journal of Environmental Economics and
Management 68:  188-202. Docket EPA-HQ-OAR-2014-0827-0080.
                                                    7-15

-------
                                                                           Employment Impacts
22 Ferris, Ann, Ronald J. Shadbegian and Ann Wolverton (2014. "The Effect of Environmental Regulation on Power
Sector Employment: Phase I of the Title IV SO2 Trading Program." Journal of the Association of Environmental
and Resource Economists 1(4): 521-553. Docket EPA-HQ-OAR-2014-0827-0078.
23 Greenstone, M. (2002). "The Impacts of Environmental Regulations on Industrial Activity: Evidence from the
1970 and 1977 Clean Air Act Amendments and the Census of Manufactures." Journal of Political Economy 110(6):
1175-1219, Docket EPA-HQ-OAR-2014-0827-0081; Walker, Reed. (2011). "Environmental Regulation and Labor
Reallocation." American Economic Review: Papers and Proceedings 101(3): 442-447, Docket EPA-HQ-OAR-2014-
0827-0091.
24 List, J. A., D. L. Millimet, P. G. Fredriksson, and W. W. McHone (2003). "Effects of Environmental Regulations
on Manufacturing Plant Births: Evidence from a Propensity Score Matching Estimator." The Review of Economics
and Statistics 85(4): 944-952. Docket EPA-HQ-OAR-2014-0827-0087.
25 Herman, E. and L. T. M. Bui (2001). "Environmental Regulation and Labor Demand: Evidence from the South
Coast Air Basin." Journal of Public Economics 79(2): 265-295. Docket EPA-HQ-OAR-2014-0827-0086.
26 httEZ/ww^M^^                                  see "Substitution Effect Employment Impacts
calculation," Docket EPA-HQ-OAR-2015-0827.
27 httE/Ml^CCTS^                              see "Substitution Effect Employment Impacts
calculation," Docket EPA-HQ-OAR-2015-0827.
28 http://www.bls.gov/emp/ep_data_emp_requirements.htm; this analysis used data for sectors 80 (Motor Vehicle
Manufacturing) and 82 (Motor Vehicle Parts Manufacturing) from "Chain-weighted (2009 dollars) real domestic
employment requirements tables;" see "Substitution Effect Employment Impacts calculation," Docket EPA-HQ-
OAR-2015-0827.
                                                   7-16

-------
                                                    Assessment of Vehicle Safety Effects
Table of Contents

Chapter 8:  Assessment of Vehicle Safety Effects	8-1
  8.1    Safety Considerations in Establishing CAFE/GHG Standards	8-1
     8.1.1   Why Do the Agencies Consider Safety?	8-1
     8.1.2   How Do the Agencies Consider Safety?	8-3
  8.2    What is the Current State of the Research on Statistical Analysis of Historical Crash
  Data?  8-5
     8.2.1   Background	8-5
     8.2.2   Historical Activities Informing the 2017-2025 Final Rule	8-8
       8.2.2.1   2011 NHTSA Workshop on Vehicle Mass, Size and Safety	8-8
       8.2.2.2   Report by Green et. al., UMTRI - "Independent Review: Statistical Analyses of
       Relationship between Vehicle Curb Weight, Track Width, Wheelbase and Fatality
       Rates," April 2011	8-9
       8.2.2.3   2012NHTSA, LBNL, and DRI Reports	8-10
     8.2.3   Final Rule for Model Years 2017-2025	8-11
     8.2.4   Activities and Development since 2017-2025 Final Rule	8-11
       8.2.4.1   2013 Workshop on Vehicle Mass, Size and Safety	8-11
       8.2.4.2   Subsequent Analyses by LBNL	8-14
       8.2.4.3   2013 Presentations to NAS Subcommittee	8-15
       8.2.4.4   2015 National Academy of Sciences' Report	8-15
       8.2.4.5   2016 NHTSA/Volpe Study Reported in "Relationships between Fatality Risk,
       Mass, and Footprint in Model Year 2003-2010 Passenger Cars and LTVs: Preliminary
       Report," June 2016	8-16
       8.2.4.6   Report by Tom Wenzel, LBNL, "An Assessment of NHTSA's Report
       'Relationships between Fatality Risk, Mass, and Footprint in Model Year 2003-2010
       Passenger Cars and LTVs,'"2016	8-29
       8.2.4.7   Fleet Simulation Model	8-37
     8.2.5   Based on this Information, What do the Agencies Consider to be the Current State of
     Statistical Research on Vehicle Mass and Safety?	8-42
  8.3    How do the Agencies Think Technological Solutions Might Affect the Safety
  Estimates Indicated by the Statistical Analysis?	8-44
     8.3.1   Workshops on Technological Opportunities and Constraints to Improving Safety
     under Mass Reduction	8-45
       8.3.1.1   2011 Workshop on Vehicle Mass, Size and Safety	8-45
       8.3.1.2   2013 Workshop on Vehicle Mass, Size and Safety	8-47
     8.3.2   Technical Engineering Projects	8-49
       8.3.2.1   Honda Accord Study	8-49
       8.3.2.2   Second Honda Accord Study	8-50
       8.3.2.3   NHTSA Silverado Study and Light-Duty Fleet Analysis	8-51
       8.3.2.4   EPA Midsize CUV "Low Development" Study	8-53
       8.3.2.5   CARB Phase 2 Midsize CUV "High Development" Study	8-55
       8.3.2.6   EPA Light Duty Truck Study	8-56
  8.4    How have the Agencies Estimated Safety Effects for the Draft TAR?	8-57
     8.4.1   What was the Agencies' Methodology for Estimating Safety Effects?	8-57
     8.4.2   Why Might the Real-World Safety Effects be Less Than or Greater Than What the
     Agencies Have Calculated?	8-61

-------
                                                             Assessment of Vehicle Safety Effects
     8.4.3   What Are the Agencies' Plans Going Forward?	8-62

Table of Figures

Figure 8.1 Actual (Unadjusted) U.S. Societal Fatality Risk per VMT vs. Curb Weight, By Vehicle Type and Model
             	8-35
Figure 8.2 Adjusted U.S. Societal Fatality Risk per VMT vs. Curb Weight, by Vehicle Type and Model, After
             Accounting for All Driver, Crash, and Vehicle Variables except Mass and Footprint	8-36
Figure 8.3 Vehicle Crash Simulations	8-39
Figure 8.4 Diagram of Computation for Overall Change in Societal Risk	8-40


Table of Tables

Table 8.1 Mass Reductions Foreseen by NHTSA/EPA and by the Committee	8-16
Table 8.2 Passenger Car and LTV Classes in the 2012 and 2016 Analyses	8-18
Table 8.3 Results of 2012 NHTSA Final Report: Fatality Increase (%) per 100-Pound Mass Reduction While
             Holding Footprint Constant	8-21
Table 8.4 Results of 2016 NHTSA Preliminary Report: Fatality Increase (%) per 100-Pound Mass Reduction While
             Holding Footprint Constant	8-21
Table 8.5 Societal Fatality Increase (%) Per 100-Pound Mass Reduction While Holding Footprint* Constant.... 8-27
Table 8.6 Societal Fatality Increase (%) Per 100-Pound Mass Reduction While Holding Footprint* Constant from
             Wenzel Study	8-31
Table 8.7 Estimated Annual Change in Fatalities from Six Different Fleetwide Mass Reduction Scenarios, Using
             Coefficients Estimated By 2012 and 2016 NHTSA Baseline Models and 2016 DRI  Measures	8-37
Table 8.8 Base Vehicle Models Used in the Fleet Simulation Study	8-39
Table 8.9 Overall Societal Risk Calculation Results for Model Runs, with Base Vehicle Restraint and Airbag
             Settings Being the same for All Vehicles, in Frontal Crash Only	8-41
Table 8.10 Additional Safety Requirements Post 2010 (FMVSS, IIHS>	8-43
Table 8.11 Mass Reduction Levels to Achieve Safety Neutral Results in the Draft TAR Analysis	8-58
Table 8.12 Examples of Mass Reduction (in Pounds) for Different Vehicle Subclasses Using the Percentage
             Information as Defined for the CAFE Draft TAR Analysis	8-59
Table 8.13 Mapping between Safety Classes and Technology Classes in the CAFE Analysis	8-60
Table 8.14 NHTSA Calculated Mass-Safety-Related Fatality Impacts of the Draft TAR Analysis  over the Lifetime
             of the Vehicles Produced in each Model Year  Using 2015 Baseline	8-60
Table 8.15 Metrics  Used in the OMEGA Safety Analysis	8-61
Table 8.16 EPA's Net Fatality Impacts over the Lifetimes of MY2021-2025 Vehicles	8-61

-------
                                                       Assessment of Vehicle Safety Effects
Chapter 8:  Assessment of Vehicle Safety Effects

8.1    Safety Considerations in Establishing CAFE/GHG Standards

8.1.1   Why Do the Agencies Consider Safety?

   The primary goals of CAFE and GHG standards are to reduce fuel consumption and GHG
emissions from the on-road light-duty vehicle fleet, but in addition to these intended effects, the
agencies also consider the potential of the standards to affect vehicle safety.A As a safety
agency, NHTSA has long considered the potential for adverse safety consequences when
establishing CAFE standards,6 and under the CAA, EPA considers factors related to public
health and human welfare, including safety, in regulating emissions of air pollutants from mobile
sources.0 Safety trade-offs associated with fuel economy increases have occurred in the past,
particularly before NHTSA CAFE standards were attribute- based,1 and the agencies must be
mindful of the possibility of future ones.  These past safety  trade-offs may have occurred because
manufacturers chose at the time, partly in response to CAFE standards, to build smaller and
lighter vehicles, rather than adding more expensive fuel-saving technologies while maintaining
vehicle size  and safety, and the smaller and lighter vehicles did not fare as well in crashes as
larger and heavier vehicles.  Historically, as shown in FARS data analyzed by NHTSA (e.g.,
Kahane,  20122), the safest cars generally have been heavy and large, while the cars with the
highest fatal-crash rates have been light and small.

   The question, then, is whether past is necessarily prologue when it comes to potential changes
in vehicle size (both footprint and "overhang") and mass in response to the more stringent future
CAFE and GHG standards.  Manufacturers have stated that they will reduce  vehicle mass as one
of the cost-effective means of increasing fuel economy and reducing CO2  emissions in order to
meet the standards, and the agencies have incorporated this expectation into  our modeling
analysis  supporting the standards. Because the agencies discern a historical relationship between
vehicle mass, size, and safety, one potential means of assessing the impact of future standards on
vehicle safety is to assume that these relationships will  continue in the future. In formulating the
MY2017-2025 final rule, the agencies were encouraged by  comments to the  NPRM from the
Alliance of Automotive Manufacturers reflecting a commitment to safety stating that, while
improving the fuel efficiency of the vehicles, the vehicle manufacturers are "mindful that such
improvements must be implemented in a manner that does not compromise the rate of safety
improvement that has been achieved to date."  The question of whether vehicle design can
mitigate  the adverse effects of mass reduction is discussed below.
A In this document, "vehicle safety" is defined as societal fatality rates per vehicle miles traveled (VMT), which
  include fatalities to occupants of all the vehicles involved in the collisions, plus any pedestrians.
B This practice is recognized approvingly in case law.  As the United States Court of Appeals for the D.C. Circuit
  stated in upholding NHTSA's exercise of judgment in setting the 1987-1989 passenger car standards, "NHTSA
  has always examined the safety consequences of the CAFE standards in its overall consideration of relevant
  factors since its earliest rulemaking under the CAFE program." Competitive Enterprise Institute v. NHTSA ("CEI
  I"), 901 F.2d 107, 120 atn. 11 (D.C. Cir. 1990).
c As noted in Section ID above, EPA has considered the safety of vehicular pollution control technologies from the
  inception of its Title II regulatory programs.  See also NRDC v. EPA. 655 F. 2d 318, 332 n. 31 (D.C. Cir. 1981).
  (EPA may consider safety in developing standards under section 202 (a) and did so appropriately in the given
  instance).
                                                8-1

-------
                                                        Assessment of Vehicle Safety Effects
   Due to the structure of the standards put in place by the MY2017-2025 rulemaking,
manufacturers are less likely than they were in the past to reduce vehicle footprint in order to
reduce mass for increased fuel economy. This factor is important because, as the agencies have
noted, historic studies have shown a positive relationship between overall vehicle size and safety,
although the relationship should continuously be re-tested as materials change in the future.  This
will be described in greater detail below.

   The primary mechanism in the MY2017-2025 rulemaking for mitigating the potential
negative effects on safety was the application of footprint-based standards, which create a
disincentive for manufacturers to produce smaller-footprint vehicles (Section II.G. 1, MY 2017-
2025  Final Rule).  This is because, as footprint decreases, the corresponding fuel economy/GHG
emission target becomes more stringent.  We also believe that the shape of the footprint curves
themselves is approximately "footprint-neutral," that is, that it should neither encourage
manufacturers to increase the footprint of their fleets, nor to decrease it. Upsizing footprint is
also discouraged through the curve "cut-off at larger footprints.0 However, the footprint-based
standards do not discourage downsizing the portions of a vehicle in front of the front axle and to
the rear of the rear axle,  or of other areas of the vehicle outside the wheels.  The crush space
provided by those portions of a vehicle can make important contributions to managing crash
energy. Additionally, simply because footprint-based standards minimize the incentive to
downsize vehicles does not mean that some manufacturers will not downsize if doing so makes it
easier for them to  meet the overall CAFE/GHG standard in a cost-efficient manner,  as for
example, if the smaller vehicles are so much lighter (or de-contented) that they exceed their
targets by much greater amounts.  On balance, however,  we believe the target curves and the
incentives they provide generally  will not encourage down-sizing (or up-sizing) in terms of
footprint reductions (or increases).E

   Given that we expect manufacturers to reduce vehicle mass in response to the standards, and
do not expect manufacturers to reduce vehicle footprint in response to the standards, the agencies
must  attempt to predict the safety effects, if any, of the final rule based on the best information
currently available. This section explained why the agencies consider safety; the following
section discusses how the agencies consider safety.
D The agencies recognize that at the other end of the curve, manufacturers who make small cars and trucks below 41
  square feet (the small footprint cut-off point) have some incentive to downsize their vehicles to make it easier to
  meet the constant target.  That cut-off may also create some incentive for manufacturers who do not currently
  offer models that size to do so in the future.  However, at the same time, the agencies believe that there is a limit
  to the market for cars and trucks smaller than 41 square feet: most consumers likely have some minimum
  expectation about interior volume, for example, among other things. Additionally, vehicles in this segment are
  the lowest price point for the light-duty automotive market, with several models in the $10,000-$15,000 range.
  Manufacturers who find themselves incentivized by the cut-off will also find themselves adding technology to the
  lowest price segment vehicles, which could make it challenging to retain the price advantage. Because of these
  two reasons, the agencies believe that the incentive to increase the sales of vehicles smaller than 41 square feet
  due to the final rule, if any, is small. See Chapter 1 of the Joint TSD for more information on the agencies' choice
  of "cut-off points for the footprint-based target curves.
E This statement makes no prediction of how consumer choices of vehicle size will change in the future, independent
  of the standards.

-------
                                                      Assessment of Vehicle Safety Effects
8.1.2   How Do the Agencies Consider Safety?

   Assessing the effects of vehicle mass reduction and size on societal safety is a complex issue.
One part of estimating potential safety effects involves trying to understand better the
relationship between mass and vehicle design.  The extent of mass reduction that manufacturers
may be considering to meet more stringent fuel economy and GHG standards may raise different
safety concerns from what the industry has previously faced. Heavier vehicles, especially truck-
based LTVs and lighter vehicles, perform differently in collisions with each other than in
collisions with another car or LTV.  When two vehicles of unequal mass collide, the change in
velocity (delta V) is higher in the lighter vehicle, similar to the mass ratio proportion.  As a result
of the higher change in velocity in lighter vehicles, the fatality risk may  also increase. Removing
more mass from the heavier vehicle than in the lighter vehicle by amounts that bring the mass
ratio closer to 1.0 reduces the delta V in the lighter vehicle and thereby reducing fatality risk and
possibly resulting in a net societal benefit.

   Another complexity is that if a vehicle is made lighter, adjustments must be made  to the
vehicle's structure such that it will be able to manage the energy in a crash while limiting
intrusion into the occupant compartment.  To maintain an acceptable occupant compartment
deceleration, the effective front-end stiffness has to be managed such that the crash pulse does
not increase as lighter yet stiffer materials are utilized. If the energy is not well managed, the
occupants may have to "ride down" a more severe crash pulse, putting more burdens  on the
restraint systems to protect the occupants3. There may be technological  and physical limitations
to how much the restraint system may mitigate these effects.

   The agencies must attempt to estimate now, based on the best information currently available
to us for analyzing these CAFE  and GHG standards, how the assumed levels of mass reduction
without additional changes (i.e.  footprint, performance, functionality) might affect the safety of
vehicles, and how lighter vehicles might affect the safety of drivers and  passengers in the entire
on-road fleet. The agencies seek to ensure that the standards are designed to encourage
manufacturers to pursue a path toward compliance that is both cost-effective and safe.

   To estimate the possible safety effects  of the MY2022-2025 standards, then, the agencies have
undertaken research that approaches this question from  several angles. First, we are using a
statistical approach to study the  effect of vehicle mass reduction on safety  historically, as
discussed in greater detail in section 8.2 below.  Statistical analysis is performed using the most
recent historical crash data available (calendar year 2005-2011 data for MY2003-2010 vehicles),
and is considered as the agencies' best estimate of potential mass-safety effects. The agencies
recognize that negative safety effects estimated based on the historical relationships could
potentially be tempered with safety technology advances in the future, and may not represent the
current or future fleet. Second,  we are using an engineering approach to investigate what amount
of mass reduction is affordable and feasible while maintaining vehicle safety and functionality
such as durability, drivability, NVH, and acceleration performance. Third, we are also studying
the new challenges these lighter vehicles might bring to vehicle safety and potential
countermeasures available to manage those challenges effectively. Comments received to the
proposed 2012 Final Rule are summarized in the 2012 Final Rule preamble.

   The agencies have looked closely at these issues, and we believe that our approach of using
both statistical analyses of historical data to assess societal safety effects, and design  studies to
                                               3-3

-------
                                                      Assessment of Vehicle Safety Effects
assess the ability of individual designs to comply with the FMVSS and perform well on NCAP
and IIHS tests responds to these concerns.

   A large body of traffic safety literature exists that examines the relationship between vehicle
mass and traffic fatality rates.  Most of the literature estimates aggregate State-level time series
correlations (Khazzoom, 19944; Noland, 20045; Ahmad and Greene, 20056; Evans, 20017) from
various angles or on a specific crash type. In general,  these studies come to varying conclusions
regarding the sign of the relationship between average vehicle mass and overall fatality rates, but
all conclude that the magnitude of this relationship is  relatively modest.

   In recent years economists have studied the "arms  race" nature of vehicle choice, and the
effect of disparity in the mass and/or size in the vehicle fleet on fleetwide safety. In particular,
they focus on the internal and external safety effect posed by larger vehicles -pickup trucks and
sport utility vehicles (SUVs)--relative to passenger cars.  Anderson and Auffhammer 2014,8
White 2004,9 Gayer 2004,10 Anderson 2008,11 Li 2012,12 and Jacobsen 201313 all conclude that
light trucks (pickups  and SUVs) impose significant societal risks relative to passenger cars.
Overall, light trucks pose a significant hazard to other users of the highway system but on
average provide no additional protection to their own occupants. Anderson (2008) estimates the
implied Pigovian tax is approximately $3850 per light truck sold, using standard value of
statistical  life figures. Anderson and Auffhammer (2014) recommend two policy options for
internalizing the external safety cost, a weight-varying mileage tax and a gas tax, and find  that
they are similar for most vehicles.

   Some of these papers use State-level data on fatalities and VMT, instead of data at the
individual vehicle level. Some estimate fatality risk once a crash has occurred, but do not
account for the effect of crash frequency on risk.  Some account for vehicle type, but not for
vehicle mass, footprint, and other characteristics by vehicle model, or for driver characteristics or
crash  circumstances.  None of the listed literature includes all of these elements in its analysis or
serves the purpose of estimating the change in societal fatality risk from reducing vehicle mass,
while holding  size (footprint) unchanged.

   It should be noted that those safety articles on the "arms race" focus on the potential role of
policy in changing the size mix, or the type mix, of the vehicle fleet. As discussed in the TSD for
the MY 2017-25 final rulemaking, Chapter 2, in developing the footprint-based standards the
agencies sought to preserve rather than change the distribution of vehicle sizes; and by
continuing to set a standard for light trucks distinct from that for cars, the agencies sought  to
preserve consumer choice  for different types of vehicles that fit their transportation needs.

   The safety analysis presented in this chapter is a statistical analysis that, unlike these cited
papers, takes all the factors listed above into account. To consider what technologies are
available for improving fuel  economy, including mass reduction, the agencies have to consider
the potential effect that those technologies may have on safety.  The purpose of our analysis is to
find a statistical relationship between mass, footprint, and safety. Specifically, the analysis is to
estimate the fatality risk effect per 100 pounds mass reduction while holding the vehicle footprint
constant.  The results of the analysis are applied in estimating fatality risk in the NHTSA Volpe
model or EPA OMEGA model. The relationships among a vehicle's mass, size, and fatality risk
are complex, and they vary in different types of crashes and by different vehicle categories.  The
performed analysis is built on the weighted logistic regression model at each fatality case level

-------
                                                      Assessment of Vehicle Safety Effects
by using updated micro data from historic annual NHTSA fatality data and State police-reported
crash data.

   The safety analysis presented in this chapter says that reducing the mass of the heavier
vehicles enhances societal safety, while reducing the mass of the lighter vehicles diminishes
societal safety .These findings agree with the disparity research discussed above that less mass
disparity is a good thing.  The agencies believe that the safety analysis in this chapter is the most
comprehensive analysis available at this time of the relationship between vehicle weight,
footprint, and societal fatality risk, and is the most appropriate to estimate what effect reduction
in vehicle mass, while holding footprint constant, of current vehicles will have on societal
fatality risk per VMT.

   The sections below discuss more specifically the state of the research on the mass-safety
relationship, and how the agencies have integrated that research into our assessment of the safety
effects of the MY2017-2025 CAFE and GHG standards.

8.2    What is the Current State of the Research on Statistical Analysis of
Historical Crash Data?

8.2.1   Background

   Researchers have been using statistical analysis to examine the relationship of vehicle mass
and safety in historical crash data for many years, and continue to refine their techniques over
time.  In the MY2012-2016 final rule, the agencies conducted further study and research into the
interaction of mass, size and safety to assist future rulemakings, and started to work
collaboratively by developing an interagency working group between NHTSA, EPA, DOE, and
CARB to evaluate all aspects  of mass, size and safety.  The team coordinated government
supported studies and independent research, to the greatest extent possible, to help ensure the
work is complementary to previous and ongoing research and to guide further research in this
area.

   The agencies also identified three specific areas to direct research in preparation for future
CAFE/GHG rulemaking in regards to statistical analysis of historical data.

   First, NHTSA would contract with an independent institution to review the statistical methods
that NHTSA and DRI have used to analyze historical data related to mass, size and safety, and to
provide recommendations on whether the existing methods or other methods should be used for
future statistical analysis of historical data. This study would include a consideration of potential
near multi-collinearity in the historical data and how best to address it in a regression analysis.
The 2010 NHTSA report was also peer reviewed by two other experts in the safety field -
Charles Farmer (Insurance Institute for Highway Safety) and Anders Lie (Swedish Transport
Admini strati on).F

   Second, NHTSA and EPA, in consultation with DOE, would update the MY 1991-1999
database on which the safety analyses in the NPRM and final rule are based with newer vehicle
F All three of the peer reviews are available in Docket No. NHTSA-2010-0152. You can access the docket at
  http://www.regulations.gov/#!home by typing 'NHTSA-2010-0152' where it says "enter keyword or ID" and then
  clicking on "Search."
                                               8-5

-------
                                                     Assessment of Vehicle Safety Effects
data, and create a common database that could be made publicly available to help address
concerns that differences in data were leading to different results in statistical analyses by
different researchers.

   And third, in order to assess if the design of recent model year vehicles that incorporate
various mass reduction methods affect the relationships among vehicle mass, size and safety, the
agencies sought to identify vehicles that are using material substitution and smart design, and to
try to assess if there is sufficient crash data involving those vehicles for statistical analysis.  If
sufficient data exists, statistical analysis would be conducted to compare the relationship among
mass, size and safety of these smart design vehicles to vehicles of similar size and mass  with
more traditional designs.

   By the time of the MY2017-2025 final rule, significant progress had been made on these tasks
since the MY2012-2016 final rule:  The independent review of recent and updated statistical
analyses of the relationship between vehicle mass, size, and crash fatality rates had been
completed. NHTSA contracted with the University of Michigan Transportation Research
Institute (UMTRI) to conduct this review, and the UMTRI team led by Paul Green evaluated
over 20 papers, including studies done by NHTSA's Charles Kahane, Tom Wenzel of the U.S.
Department of Energy's Lawrence Berkeley National Laboratory, Dynamic Research, Inc., and
others. UMTRI's basic findings will be discussed below.

   Some commenters in recent CAFE rulemakings, including some vehicle manufacturers,
suggested that the designs and materials of more recent model year vehicles may have weakened
the historical statistical relationships between mass, size, and  safety. The agencies agreed that
the statistical analysis would be improved by using an updated database that reflects more recent
safety technologies, vehicle designs and materials, and reflects changes in the overall vehicle
fleet, and an updated database was created and employed for assessing safety effects in the final
rule. The agencies also believed, as UMTRI also found, that different statistical analyses may
have produced different results because they each used slightly different datasets for their
analyses.

   In order to try to mitigate this issue and to support 2012 rulemaking, NHTSA created a
common, updated database for statistical analysis that consisted of crash data of model years
2000-2007 vehicles in calendar years 2002-2008, as compared to the database used in prior
NHTSA analyses based on model years 1991-1999 vehicles in calendar years 1995-2000.  The
2012 database was the most up-to-date possible at that time, given the processing lead time for
crash data and the need for enough  crash cases to permit statistically meaningful analyses.
NHTSA made the preliminary version  of the new database, which was the basis for NHTSA's
2011 report, available to the public  in May 2011, and an updated version in April 2012,°
enabling other researchers to analyze the same data and hopefully minimizing discrepancies in
the results that would have been due to inconsistencies across databases.14

   The agencies were aware that several studies had been conducted using the 2011 version or
the 2012 version of NHTSA's safety database. In addition to three NHTSA studies, which are
discussed in Section 8.2.5, other studies included two by Wenzel at Lawrence Berkeley National
Laboratory (LBNL) under contract  with the U.S. DOE, and one by Dynamic Research, Inc.
} These databases are available at ftp://ftp.nhtsa.dot.gov/CAFE/.

-------
                                                         Assessment of Vehicle Safety Effects
(DRI) contracted by the International Council on Clean Transportation (ICCT).  These studies
took somewhat different approaches to examine the statistical relationship between fatality risk,
vehicle mass and size.  In addition to a detailed assessment of the NHTSA 2011 report, Wenzel
considered the effect of mass and footprint reduction on casualty risk per crash, using data from
thirteen states, where casualty risk included both fatalities and serious or incapacitating injuries.
Both LBNL studies were peer reviewed and subsequently revised and updated. DRI used
models that separate the effect of mass reduction on two components of fatality risk, crash
avoidance and crashworthiness.  DRI studies were also peer reviewed and revised in response to
peer reviewer's questions.  The LBNL and DRI studies were made available in the docket for the
2012 final rule.H The database was made available for download to the public from NHTSA's
website.

   Finally, EPA and NHTSA with DOT's Volpe Center, part of DOT's Office of the Assistant
Secretary for Research and Technology, attempted to investigate the implications of "Smart
Design," by identifying and describing the types of "Smart Design" and methods for using
"Smart Design" to result in vehicle mass reduction,  selecting analytical pairs of vehicles, and
using the appropriate crash database to analyze vehicle crash data. The analysis identified
several one-vehicle and two-vehicle crash datasets with the potential to shed light on the issue,
but the  available data for specific crash scenarios was insufficient to produce consistent results
that could be used to support conclusions regarding  historical performance of "Smart Designs."
This study was also available in the docket for the final rule.15
H Wenzel, T. (201 la). Assessment of NHTSA's Report "Relationships Between Fatality Risk, Mass, and Footprint
  in Model Year 2000-2007 Passenger Cars and LTVs - Draft Final Report." (Docket No. NHTSA-2010-0152-
  0026). Berkeley, CA: Lawrence Berkeley National Laboratory; Wenzel, T. (201 Ib). An Analysis of the
  Relationship between Casualty Risk Per Crash and Vehicle Mass and Footprint for Model Year 2000-2007 Light-
  Duty Vehicles - Draft Final Report." (Docket No. NHTSA-2010-0152-0028). Berkeley, CA: Lawrence Berkeley
  National Laboratory; Wenzel, T. (2012a). Assessment of NHTSA's Report "Relationships Between Fatality
  Risk, Mass, and Footprint in Model Year 2000-2007 Passenger Cars and LTVs - Final Report." (To appear in
  Docket No. NHTSA-2010-0152). Berkeley, CA: Lawrence Berkeley National Laboratory; Wenzel, T. (2012b).
  An Analysis of the Relationship between Casualty Risk Per Crash and Vehicle Mass and Footprint for Model
  Year 2000-2007 Light-Duty Vehicles - Final Report." (To appear in Docket No. NHTSA-2010-0152). Berkeley,
  CA: Lawrence Berkeley National Laboratory; Van Auken, R.M., and Zellner, J. W. (2012a).  Updated Analysis
  of the Effects of Passenger Vehicle Size and Weight on Safety, Phase I. Report No. DRI-TR-11-01. (Docket No.
  NHTSA-2010-0152-0030). Torrance, CA: Dynamic Research, Inc.; Van Auken, R.M., and Zellner, J. W.
  (2012b). Updated Analysis of the Effects of Passenger Vehicle Size and Weight on Safety, Phase II; Preliminary
  Analysis Based on 2002 to 2008 Calendar Year Data for 2000 to 2007 Model Year Light Passenger Vehicles to
  Induced-Exposure and Vehicle Size Variables.  Report No. DRI-TR-12-01, Vols. 1-3. (Docket No. NHTSA-
  2010-0152-0032). Torrance, CA: Dynamic Research, Inc.; Van Auken, R.M., and Zellner, J. W. (2012c).
  Updated Analysis of the Effects of Passenger Vehicle Size and Weight on Safety, Phase II; Preliminary Analysis
  Based on 2002 to 2008  Calendar Year Data for 2000 to 2007 Model Year Light Passenger Vehicles to Induced-
  Exposure and Vehicle Size Variables. Report No. DRI-TR-12-01, Vols. 4-5. (Docket No. NHTSA-2010-0152-
  0033). Torrance, CA: Dynamic Research, Inc.; Van Auken, R.M., and Zellner, J. W. (2012d). Updated Analysis
  of the Effects of Passenger Vehicle Size and Weight on Safety; Sensitivity of the Estimates for 2002 to 2008
  Calendar Year Data for 2000 to 2007 Model Year Light Passenger Vehicles to Induced-Exposure and Vehicle
  Size Variables. Report No. DRI-TR-12-03. (Docket No. NHTSA-2010-0152-0034). Torrance, CA: Dynamic
  Research, Inc.
                                                 8-7

-------
                                                      Assessment of Vehicle Safety Effects
   Since the publication of the MY2017-2025 final rule, NHTSA has sponsored new studies and
research to inform the midterm evaluation and the MY2022-2025 rulemaking.  A newly updated
NHTSA study, presented in Section 8.2.5, represents the latest iteration of the database and
analysis applied in the 2011 and 2012 NHTSA reports. The updated database created for the
study consists of crash data of MY2003-2010 vehicles in calendar years 2005-2011, and follows
the identical analytical structure as the peer-reviewed method applied in the 2011 and 2012
reports. NHTSA published a separate preliminary report in 2016, applying this newly updated
database.1 The agencies recognize, however, that the updated database may not represent the
future fleet, because vehicles have continued and will continue to change.

   Wenzel at Lawrence Berkeley National Laboratory (LBNL) also conducted a statistical
analysis using the new database.  Wenzel's new findings are summarized in Section 8.2.6.

   In addition, the National Academy of Sciences published a new report in this area in 2015,
discussed in Section 8.2.4.16

   Throughout the midterm evaluation process, NHTSA's goal is to publish as much of our
research as possible.  Thus, while some of these reports have already been published, all are
summarized below.  In establishing standards, the agencies will consider all available data,
studies and information objectively without regard to whether they were sponsored by the
agencies.

   Technical assessment and review of previous studies and current findings helps the agencies
come closer to resolving some of the ongoing debates in statistical  analysis research of historical
crash data that are detailed later in this chapter. We intend to apply these conclusions going
forward in Draft TAR future rulemakings, and we believe that the public discussion of the issues
will be facilitated by the research conducted.

   The following sections chronologically discuss the findings from these studies and others in
greater detail. Section 8.2.2 summarize historical activities leading up to the 2017-2025 final
rule published in 2012, and sections 8.2.4 cover developments since 2012 conducted for the
midterm evaluation and anticipation of rulemaking  for model years 2022-2025, including
updated analyses.

8.2.2  Historical Activities Informing the 2017-2025 Final Rule

8.2.2.1 2011 NHTSA  Workshop  on Vehicle Mass,  Size and Safety

   On February 25, 2011, NHTSA hosted a workshop on mass reduction, vehicle size, and fleet
safety at the Headquarters of the U.S. Department of Transportation in Washington, DC.J  The
purpose of the workshop was to provide the agencies with a broad understanding of current
research in the field and provide stakeholders and the public with an opportunity to weigh in on
this issue, by bringing together experts in the field to discuss some of the overarching questions
1 The preliminary report can be found in Docket No. NHTSA-2010-0131.
1A video recording, transcript, and the presentations from the NHTSA workshop on mass reduction, vehicle size and
  fleet safety is available at http://www.nhtsa.gov/fuel-economy (look for "NHTSA Workshop on Vehicle Mass-
  Size-Safety on Feb. 25").

-------
                                                      Assessment of Vehicle Safety Effects
to be examined in NHTSA's impending CAFE rulemaking. NHTSA also created a public docket
to receive comments from interested parties that were unable to attend.

   The speakers included Charles Kahane of NHTSA, Tom Wenzel of Lawrence Berkeley
National Laboratory, R. Michael Van Auken of Dynamic Research Inc. (DRI), Jeya Padmanaban
of JP Research, Inc., Adrian Lund of the Insurance Institute for Highway Safety, Paul Green of
the University of Michigan Transportation Research Institute (UMTRI), Stephen Summers of
NHTSA, Gregg Peterson of Lotus Engineering, Koichi Kamiji of Honda, John German of the
International Council on Clean Transportation (ICCT), Scott Schmidt of the Alliance of
Automobile Manufacturers, Guy Nusholtz of Chrysler, and Frank Field of the Massachusetts
Institute of Technology.

   The wide participation in the workshop allowed the agencies to hear from a broad range of
experts  and stakeholders. The contributions were particularly relevant to the agencies' analysis
of the effects of mass reduction for the MY2017-2025 final rule. The presentations were divided
into two sessions that addressed the two expansive sets of issues: statistical evidence of the roles
of mass and size on safety, and engineering realities regarding structural crashworthiness,
occupant injury and advanced vehicle design.

   Some main points from the workshop were:

   •  Statistical studies of crash data that attempt to identify the relative recent historical
       effects of vehicle mass and size on fleet safety shows complicated relationships with
       many confounding influences in the data.
   •  Analyses must also control  for individual technologies with significant safety effects
       (e.g., Electronic Stability Control,  airbags).
   •  The physics of a two-vehicle crash require that the lighter vehicle experience a greater
       change in velocity, which, all else  being equal, often leads to disproportionately more
       injury risk.
   •  The separation of key parameters is a challenge to the analyses,  as vehicle size has
       historically been highly correlated with vehicle mass.
   •  There was no consensus on whether smaller, lighter vehicles maneuver better, and thus
       avoid more crashes, than larger, heavier vehicles.
   •  Kahane's results from his 2010 report found that a scenario which took some mass out of
       heavier vehicles but little or no mass out of the lightest vehicles did not impact safety in
       absolute terms,  and noted that if the analyses were able to consider the mass of both
       vehicles in a two-vehicle crash, the results may be more indicative of future crashes.
8.2.2.2 Report by Green et. al, UMTRI- "Independent Review: Statistical Analyses of
Relationship between Vehicle Curb Weight, Track Width, Wheelbase  and Fatality Rates,"
April 2011

   As explained above, NHTSA contracted with the University of Michigan Transportation
Research Institute (UMTRI) to conduct an independent reviewK of a set of statistical analyses of
relationships between vehicle curb weight, the footprint variables (track width, wheelbase) and
fatality rates from vehicle crashes.  The purpose of this review was to examine analysis methods,
K The review is independent in the sense that it was conducted by an outside third party without any interest in the
  reported outcome.

-------
                                                       Assessment of Vehicle Safety Effects
data sources, and assumptions of the statistical studies, with the objective of identifying the
reasons for any differences in results. Another objective was to examine the suitability of the
various methods for estimating the fatality risks of future vehicles.

   UMTRI reviewed a set of papers, reports, and manuscripts provided by NHTSA (listed in
Appendix A of UMTRI's report, which is available in the docket to the MY2017-2025
rulemaking) that examined the statistical relationships between fatality or casualty rates and
vehicle properties such as curb weight, track width, wheelbase and other variables.

   Fundamentally, the UMTRI team concluded that the database created by Kahane appeared to
be an impressive  collection of files from appropriate sources and the best ones available for
answering the research questions considered in this study; and that the disaggregate logistic
regression model used by NHTSA in the 2003 report17 seemed to be the most appropriate model,
and valid for the analysis in the context that it was used: finding general associations between
fatality risk and mass - and the general directions of the reported associations were correct.

8.2.2.3 2012 NHTSA, LBNL, andDRIReports

   NHTSA published a study in 2012 (Kahane, 2012) that estimated the effect of mass reduction
on US societal fatality risk per VMT, using light vehicles from model years 2000 to 2007 in
calendar years 2002 to 2008.  NHTSA's methodology in part responded to comments Paul  Green
made in his 2011  review.  For the first time NHTSA included the correlated variables vehicle
curb weight and footprint in its baseline regression model, for two reasons:  an analysis indicated
that the model variance inflation factors were not  high enough to preclude including the two
correlated variables in the same regression model, and the fuel economy/greenhouse gas
emission standards adopted for model years 2012  to 2016 were based on a vehicle's footprint, so
the regression model needed to estimate the effect mass reduction would have on safety while
holding footprint constant. The model used came to be known as the "baseline" model, and the
study found that mass reduction in only lighter-than-average cars was associated with a
statistically-significant increase in fatality risk; for the other vehicle types, mass reduction was
associated with increases or decreases in fatality risk that were not statistically significant.  This
study is cited in more detail in Section 8.2.6, detailing the current follow-up. NHTSA published
a preliminary report in 2011 that was subject to external review; the final  report was published  in
2012.

   In its 2012 "Phase 1" report18, LBNL replicated the 2012 NHTSA baseline results, and
conducted 19 alternative regression models to test the sensitivity of the NHTSA baseline model
to changes in the  measure of risk, the variables included, and the data used.  In its report LBNL
pointed out that other vehicle attributes, driver characteristics, and crash circumstances were
associated with much larger changes in risk than mass reduction.L  LBNL also demonstrated  that
L As stated at p. iv, Executive Summary of LBNL 2012 Phase 1 report, "many of the control variables NHTSA
  includes in its logistic regressions are statistically significant, and have a much larger estimated effect on fatality
  risk than vehicle mass. For example, installing torso side airbags, electronic stability control, or an automated
  braking system in a car is estimated to reduce fatality risk by about 10%; cars driven by men are estimated to
  have a 40% higher fatality risk than cars driven by women; and cars driven at night, on rural roads, or on roads
  with a speed limit higher than 55 mph are estimated to have a fatality risk over 100 times higher than cars driven
  during the daytime on low-speed non-rural roads.

-------
                                                     Assessment of Vehicle Safety Effects
there was little correlation between mass and fatality risk by vehicle model, even after
accounting for all other vehicle attributes, driver characteristics, and crash circumstances.

   In its 2012 "Phase 2" report19, LBNL used data from police reported crashes in the 13 states to
study casualty (fatality plus severe injury) risk per VMT, and to divide risk per VMT into its two
components, crash frequency (crashes per VMT) and crashworthiness/crash compatibility (risk
per crash). LBNL found that mass reduction was associated with increases in crash frequency,
and decreases in risk per crash. Preliminary versions LBNL's Phase 1 and Phase 2 reports were
reviewed by external reviewers20, and comments incorporated into the final versions published in
2012.

   DRI published three preliminary reports in 2012.  DRI's preliminary Phase I report updated
its analysis of data from 1995 to 2000, and was able to replicate the results from NHTSA's 2003
report.  DRI's preliminary Phase II report replicated the 2012 NHTSA baseline results,  and used
a simultaneous two-stage model to estimate the separate effects of mass reduction on crash
frequency and fatality risk per crash.  The results from DRI's two-stage model were comparable
to LBNL's Phase 2 analysis: that mass reduction was associated with increases in crash
frequency, and decreases in risk per crash. DRI's preliminary Summary report showed the effect
of two alternative regression models: using stopped rather than non-culpable vehicles as the basis
for the induced exposure database, and replacing vehicle footprint with its components
wheelbase and track width. Under these two alternatives, mass  reduction was associated with
more beneficial changes in fatality risk. The three preliminary DRI reports were peer-reviewed,
with comments incorporated into the final versions published in 2013.

   The results from LBNL's Phase 2 and DRI's Phase II reports implied that the increase in
fatality risk per VMT from mass reduction in lighter cars estimated by the NHTSA baseline
model was due to increasing crash frequency, and not increasing fatality risk once a crash had
occurred, as mass is reduced. In the final version of its 2012 report NHTSA argued that the
effects of crash frequency could not be separated from risk per crash because of reporting bias in
state crash data, such as lack of a crash severity measure, and possible bias due to under-
reporting of less severe crashes in certain States.

8.2.3   Final Rule for Model Years 2017-2025

   In August 2012, EPA and NHTSA jointly published the Joint Technical Support Document:
Final Rulemaking for (Model Years) 2017-2025, Light-Duty Vehicle Greenhouse Gas Emission
Standards (EPA) and Corporate Average Fuel Economy Standards (NHTSA); EPA-420-R-12-
901. Since NHTSA rules are always in lengths of five years, the standards for model years 2022-
2025 for Corporate Average Fuel Economy (CAFE)  are considered "augural" and must be
revisited for a permanent rule. Analyses described in the following sections will inform not only
the midterm evaluation of the 2017-2025 rule but the final CAFE rule for MY's 2022-2025.

8.2.4   Activities and Development since 2017-2025  Final Rule

8.2.4.1 2013 Workshop on Vehicle Mass, Size and Safety

   On May 13-14, 2013, NHTSA hosted a follow-on symposium to  continue to explore the
relevant issues and concerns with mass, size, and potential safety tradeoffs, bringing together
experts in the field to discuss questions to address CAFE standards for model years 2022-2025.
The first day of the two-day symposium focused on engineering, while the second day
                                             8-11

-------
                                                      Assessment of Vehicle Safety Effects
investigated various methodologies for assessing statistical evidence of the roles of vehicle mass
and size on occupant safety.  All presentations may be seen on NHTSA's web site at:
   The speakers for the second day, focusing on the subject matter of this chapter, included
Charles Kahane of NHTSA, Joe Nolan of the Insurance Institute for Highway, Guy Nusholtz of
Chrysler, Mike van Auken of Dynamic Research Incorporated, and Tom Wenzel of Lawrence
Berkeley National Laboratory. Summaries of the topics follow:

   •  Kahane gave an overview of statistical studies designed to determine the incremental
      change in societal risk as vehicle mass of a particular vehicle is modified while keeping its
      footprint (the product of wheel base and track width) is kept constant. The physics of
      crashes, in particular conservation of momentum and equal and opposite forces, imply
      that mass reduction in the heaviest vehicles and/or mass increase in the lightest vehicles
      can reduce societal risk in two-vehicle crashes. It is therefore reasonable that reducing
      disparities in mass ratio in the vehicle fleet (such as by reducing the mass of heavy
      vehicles by a larger percentage than that of light vehicles) should reduce societal harm.
      This trend was noticed in the data for model year 2000-2007 vehicles, but only
      statistically significant for the lightest group of vehicles.  This is  similar to the results
      found for model year 1991-1991 vehicles in a 2003 study. Kahane acknowledged
      numerous confounding factors such as maneuverability of different vehicle classes
      (although data indicated smaller cars were more likely to be involved in crashes), driver
      attributes and vulnerabilities, advances in restraint safety systems and vehicle structures,
      and, and electronic stability control.

   •  Wenzel replicated Kahane' s results using the same data and methods, but came to slightly
      different conclusions. He demonstrated that the effect of mass or footprint reduction that
      Kahane estimated on societal risk is much smaller than the effect Kahane estimated for
      other vehicle attributes, driver characteristics, or crash circumstances. Wenzel plotted
      actual fatality risk vs. weight by vehicle make and model, and estimated predicted risk by
      make and model after accounting for all control variables used in NHTSA's baseline
      model  except for mass and footprint.  The remaining,  or residual  risk, not explained by the
      control variables has no correlation with vehicle weight.  He presented results of the 19
      alternative regression models he conducted to test the sensitivity  of the results from
      NHTSA's baseline model. He also presented results from LBNL's Phase 2 analysis,
      which  examined the effect of mass or footprint reduction on the two components of risk
      per VMT: crashes per VMT (crash frequency), and risk per crash (crashworthiness). Both
      his analysis of casualty risk using crash data from 13 states, and his replication of the DRI
      two-state simultaneous regression model, indicate that mass reduction is associated with
      an increase in crash frequency, but a decrease in risk per crash.

   •  Van Auken also replicated Kahane' s results from the NHTSA baseline model, and
      presented results from three sensitivity regression models. Replacing footprint with its
      components wheelbase and track width reduces the estimated increase in risk from mass
      reduction in cars, and suggests that mass reduction in light trucks decreases societal risk.
      Using stopped rather than non-culpable vehicles to derive the induced exposure  dataset
      also reduces the estimated increase in risk from mass reduction in lighter-than-average

-------
                                                   Assessment of Vehicle Safety Effects
   cars and light trucks, and estimates that mass reduction in heavier cars and trucks
   decreases societal risk. Including both of these changes to the NHTSA baseline model
   greatly reduces the estimated increase in risk from mass reduction in the lightest cars, and
   is associated with decreases in risk for all other vehicle types. Van Auken described in
   more detail his two-stage simultaneous regression model, that allows risk per vehicle mile
   of travel to be decomposed into crashes per VMT (crash frequency) and risk per crash
   (crashworthiness/crash compatibility). As with Wenzel's analysis, Van Auken found that
   mass reduction is associated with an increases in crash frequency, but with a decrease in
   risk per crash. Once again, the resulting trends were similar to those from Kahane and
   Wenzel. Van Auken explored the issue of inducing the exposure of vehicles via crash
   statistics in which relative exposure was measured by non-culpable vehicles in the crash
   database versus by its subset of stopped vehicles in the data, and also investigated the
   impact in substituting footprint for track width and wheelbase as size variables in the
   regression.

•  Nusholtz of Chrysler presented an analysis of the sensitivity of the fleet-wide fatality risk
   to changes in vehicle mass and size. He noted  the difficulty in finding a definitive metric
   for "size." He dismissed some assertions of mass having negligible (or purely negative)
   effect on safety as leading to absurd conclusions in the extreme.  He extended the methods
   of Joksch (1993) and Evans (1992) to estimate risk as a function of readily measurable
   vehicle attributes and reported crash characteristics. He used crash physics (closing
   speed, estimates of inelastic stiffness and energy absorption) to estimate changes in fleet
   risk as a function of changes in these parameters. He observed that mass is a dominant
   factor but believes crush space could begin to dominate if vehicles could be made larger.
   He concurred that removing more mass from larger vehicles can reduce overall risk but is
   not convinced that such a strategy will be sufficient to meet fuel economy goals. He
   regards the safety implications of mass reduction to be transition issues, of greater
   importance so long as legacy heavier vehicles are used in significant numbers.

•  Nolan analyzed historical trends in the fleet. While median vehicle mass has increased,
   safety technologies have enhanced the safety of current small cars to the level only
   achieved by larger cars in the past. In particular, electronic stability control has reduced
   the relative importance of some severe crash modes. While acknowledging that smaller
   vehicles will always be at a disadvantage, there is hope that further technological
   advances such as crash avoidance systems hold promise in advancing safety.  Fleet safety
   would be enhanced if these technologies could quickly penetrate across the fleet to small
   cars as well as large ones.

•  An attempt was made to separate the effect of mass on crash outcome as distinct from the
   likelihood of the crash itself.  It was acknowledged that mass can affect both. Nusholtz
   emphasized that crash parameters (e.g., closing speed) necessarily dominate.  Kahane
   suggested that reporting rates might be sufficiently different to affect results.  Nusholtz
   cautioned that physics and statistics must be considered but in a way that connects them to
   reality rather than abstractions. Nolan hopes that crash avoidance effects could be very
   significant. Nusholtz noted that  assessments of that effect are difficult in that determining
   when and why a crash didn't occur is problematic against the backdrop of confounding
   information.

-------
                                                     Assessment of Vehicle Safety Effects
8.2.4.2 Subsequent Analyses by LBNL

   As part of its review of the 2012 DRI studies,21 LBNL recreated DRI's two-stage
simultaneous regression model, which estimated the effect of mass or footprint reduction on the
two components of fatality risk per VMT: the number of crashes per VMT and the risk of fatality
per crash (Wenzel 2013). LBNL first replicated DRI's methodology of taking a random
"decimated" sample of the crash data from 10 states for the induced exposure records.  Although
LBNL was not able to exactly recreate DRI's results, its results were comparable to DRI's, and
LBNL's Phase 2 analysis: mass reduction is associated with increases in crash frequency for all
vehicle types, and with decreases in fatalities per crash for all vehicle types except heavier cars.
LBNL then re-ran the two-stage regression model using all crash data from the 13  states NHTSA
used in their baseline model, and obtained similar results.

   The LBNL Phase 2 study and DRI Phase II study had two unexpected results: that mass
reduction is associated with increased crash frequency, but decreased risk per crash; and the
signs on some of the control variables are in the unexpected direction.  For example, side airbags
in light trucks and CUVs/minivans were estimated to reduce crash frequency; the crash
avoidance technologies electronic stability control (ESC) and antilock braking systems (ABS)
were estimated to reduce risk once a crash had occurred; and all-wheel-drive and brand new
vehicles were estimated to increase risk once a crash had occurred. In addition, male drivers
were estimated to have essentially no effect on crash frequency, but were associated with a
statistically significant increase in fatality risk once a crash had occurred. And driving at night,
on high-speed or rural roads, were associated with higher increases in risk per crash than on
crash frequency. A possible explanation for these unexpected results is that important control
variables were not being included in the regression models. For example, crashes involving male
drivers, in vehicles equipped with AWD, or that occur at night on rural or high-speed roads, may
not be more frequent but rather more severe than other crashes, and thus lead to greater fatality
or casualty risk.  And drivers who select vehicles with certain safety features may tend to drive
more carefully, resulting in vehicle safety features designed to improve crashworthiness or
compatibility, such as side airbags, being also associated with lower crash frequency.

   LBNL made several attempts to create a regression model that "corrected" these unexpected
results. 22 LBNL first examined the results of three vehicle braking and handling tests conducted
by Consumer Reports: the maximum speed achieved during the avoidance maneuver test,
acceleration time from 45 to 60 mph, and dry braking  distance.  When these three test results
were added to the LBNL baseline regression model of the number of crashes per mile of vehicle
travel in cars, none of the three handling/braking variables had the expected effect  on crash
frequency.  In other words, an increase in maximum maneuver speed, the time to reach 60 miles
per hour, or braking distance on dry pavement in cars, either separately or combined, was
associated with a decrease in the likelihood of a crash, of any type or with a stationary object.
Adding one or all of the three handling/braking variables had relatively little effect on the
estimated relationship between mass or footprint reduction in cars and crash frequency, either in
all types of crashes or only in crashes with stationary.

   LBNL next tested the sensitivity of the relationship between mass or footprint reduction and
crash frequency by adding five additional variables to the regression models: initial vehicle price,
average household income, bad driver rating, alcohol/drug use, and seat belt use. An increase in
vehicle price, household income, or belt use was associated with a decrease in crash frequency,

-------
                                                     Assessment of Vehicle Safety Effects
while an increase in alcohol/drug use was associated with an increase in crash frequency, for all
three vehicle types; a poor bad driver rating increases crash frequency in cars, but unexpectedly
decreases crash frequency in light trucks and CUVs/minivans. Including these five variables,
either individually or including all in the same regression model, did not change the general
results of the baseline LBNL regression model: that mass reduction is associated with an increase
in crash frequency in all three types of vehicles, while footprint reduction is associated with an
increase in crash frequency in cars and light trucks, but with a decrease in crash frequency in
CUVs/minivans. The variable with the biggest effect was initial vehicle purchase price, which
dramatically reduced the estimated increase in crash frequency in heavier-than-average cars (and
in heavier-than-average light trucks, and all CUVs/minivans).  These results suggest that other,
more subtle, differences in vehicles and their drivers account for the unexpected finding that
lighter vehicles have higher crash frequencies than heavier vehicles, for all three types of
vehicles.

   In its 2012 report NHTSA suggested two possible explanations for the unexpected results in
the LBNL Phase 2 analysis and the DRI and LBNL two-stage regression models: that the
analyses did not account for the severity of the crash, and possible bias in the crashes reported to
police in different states, with less severe crashes being under-reported for certain vehicle types.
LBNL analyzed the first of Kahane's explanations for the unexpected result of mass reduction
being associated with decreased risk per crash, by re-running the baseline Phase 2 regressions
after excluding the least-severe crashes from the state crash databases objects.23 Only vehicles
that were described as "disabled" or as having "severe" damage were included, while vehicles
which were driven away from the crash site or had functional, none, or unknown damage were
excluded.  Excluding non-severe  crashes had little effect on the relationship between mass
reduction and crash frequency, in either LBNL's Phase 2 baseline model or the two-stage
simultaneous model: mass reduction was associated with an increase in crash frequency, and a
decrease in risk per crash. Excluding the non-severe crashes also did not change the unexpected
results for the other control variables: most of the  side airbag variables, and the crash
compatibility variables  in light trucks, continued to be associated with an increase in crash
frequency, while antilock braking systems, electronic stability control, all-wheel drive, male
drivers, young drivers, and driving at night, in rural counties, and on high speed roads all
continued to be associated with an increase in risk per crash.

8.2.4.3 2013 Presentations to NAS Subcommittee

   Chuck Kahane, Tom Wenzel,  Stephen Ridella, (and Chuck Thomas, Honda, and Chuck
Nolan, IIHS) were invited to the June 2013 NAS subcommittee on light-duty fuel economy to
present the results from their 2012 analyses.   At the meeting committee members raised several
questions about the studies; the presenters responded to these questions at the meeting, as well as
in two emails in August 2013 and December 2014.

8.2.4.4 2015 National Academy of Sciences' Report

   In 2015, the National Academy of Sciences published the report "Cost, Effectiveness and
Deployment of Fuel Economy Technologies for Light-Duty Vehicles."  The report is the result
of the work of the Committee on  Assessment of Technologies for Improving the Fuel Economy
of Light-Duty Vehicles, Phase 2,  established upon the request of NHTSA to help inform the
midterm review. The committee  was asked to assess the CAFE standard program and the
analysis leading to the setting of the  standards, as  well as to provide its opinion on costs and fuel

-------
                                                     Assessment of Vehicle Safety Effects
consumption improvements of a variety of technologies likely to be implemented in the light-
duty fleet between now and 2030 (see further discussion in Chapter 2.2.1).

   In the particular area of mass and safety, as shown below, the Committee found the agencies'
estimates of mass reductions to be conservative, particularly for mid-size and small vehicles.
             Table 8.1  Mass Reductions Foreseen by NHTSA/EPA and by the Committee
Mass Reductions Foreseen by NHTSA/EPA and by the Committee (percent)24
Vehicle
Small Car
Midsize Car
Large Car
Minivan
Light duty truck
NHTSA/EPA TSD Estimate
0
3.5
10
20
20
Committee Estimate
5
10
15
20
20
   The Committee acknowledged the possibility of negative safety impacts during the transition
period, due to variances in how reductions occurred. Because of this, the Committee
recommended NHTSA consider and, if necessary, take steps to mitigate this possibility.

8.2.4.5 2016 NHTSA/Volpe Study Reported in "Relationships between Fatality Risk, Mass,
and Footprint in Model Year 2003-2010 Passenger Cars andLTVs: Preliminary Report,"
June 2016

   The relationship between a vehicle's mass, size, and fatality risk is complex, and varies
depending upon the type of crash.  NHTSA, along with others, has been examining this
relationship for over a decade. The safety chapter of NHTSA's 2012 final regulatory impact
analysis (FRIA) of CAFE standards for MYs 2017-2025  passenger cars and light trucks included
a statistical analysis of relationships between fatality risk, mass, and footprint in MY2000-2007
passenger cars and LTVs (light trucks and vans), based on calendar year (CY) 2002-2008 crash
and vehicle-registration data (Kahane, Aug. 2012).

   The principal findings of NHTSA's 2012 analysis were that mass reduction, while holding
footprint constant, was estimated to result in a statistically significant increase in societal fatality
risk in lighter cars, but a statistically significant decrease in societal fatality risk in heavier LTVs
by decreasing the fatality risk of occupants in lighter vehicles which collide with the heavier
LTVs. NHTSA concluded that, as a result,  any reasonable combination of mass reductions while
holding footprint constant in MYs 2017-2025 vehicles -  concentrated, at least to some extent, in
the heavier LTVs and  limited in the lighter cars - would  likely be approximately safety-neutral;
it would not significantly increase fatalities and might well decrease them. LBNL replicated
these results in its 2012 assessment of the NHTSA study.

   NHTSA's 2012 report partially agreed and partially disagreed with analyses published during
2010-2012 by Dynamic Research, Inc. (DRI).  NHTSA, LBNL, and DRI all found a significant
protective effect for footprint, and that reducing mass and footprint together (downsizing) on
smaller vehicles was harmful. DRI's analyses estimated a statistically significant decrease in
fatalities from mass reduction in all light-duty vehicles if wheelbase and track width were
maintained, whereas NHTSA's report showed overall fatality reductions only in the heavier
LTVs, and benefits only in some types of crashes for other vehicle types. Much of the NHTSA,

-------
                                                       Assessment of Vehicle Safety Effects
LBNL, and DRI 2012 reports involved sensitivity tests on the databases and models, which
generated a range of estimates somewhere between the initial DRI and NHTSA results.M

   In May 2015, NHTSA, working closely with EPA and the Department of Energy (DOE),
commenced a new statistical analysis of the relationships between fatality rates, mass and
footprint, updating the crash and exposure databases to the latest available model years, and
utilizing the same methodology as in the 2012 NHTSA report.  The new databases use the most
up-to-date data available, given the processing lead time for crash data and the need for enough
crash cases to permit statistically meaningful analyses. NHTSA made the first version of the
new databases available to the public in 2016, concurrently with the release of its 2016
preliminary report,25 enabling other researchers to analyze the same data and hopefully
minimizing discrepancies in the results due to inconsistencies across the data used.26

   One way to estimate the effect of mass reduction on safety is the use of statistical analyses of
societal fatality risk per vehicle miles traveled (VMT) for the current on-road vehicle fleet.
Consistent with this, the analysis follows the identical approach employed in the 2012 NHTSA
report, centering on cross-sectional logistic regressions of societal fatality risk per billion vehicle
miles of travel (the dependent variable), as a function of driver- (e.g., driver age and gender),
vehicle- (e.g., safety features) and crash-specific factors (e.g., times, locations). Societal fatality
risk represents total fatalities to all vehicle occupants, pedestrians, cyclists and motorcyclists
involved in collisions per volume of VMT.

   The paramount purpose of the analysis is to develop five parameters for use in the CAFE
Compliance and Effects Modeling System (usually  referred to as the "Volpe model," developed
for NHTSA by the Volpe National Transportation Systems Center) to estimate the safety effects,
if any,  of the modeled mass  reductions in MY2022-2025 vehicles over their lifetime.  The
primary difference from the 2012 report is that the set of case vehicles and time period for
observed vehicle incidents is more recent, involving model year (MY) 2003-2010 vehicles in
calendar year (CY) 2005-2011, versus MY2000-2007 vehicles  in CY2002-2008 in the 2012
report. The most notable vehicle-specific factors for this analysis are curb weight and vehicle
size (represented as footprint in the preferred model structure).

   After controlling for driver-, crash- and other vehicle-specific factors including footprint, the
logistic regression estimates percentage changes in  societal fatalities as curb weight varies by
100 pounds.  The logistic regressions in the analysis are applied to five vehicle classes: two
passenger car classes, two LTV classes, and one class combining crossover (CUV)  vehicles and
minivans. In both the 2012  report and this analysis, the vehicle classes for passenger cars and
M Van Auken, R. M., and Zellner, J. W. (2003). A Further Assessment of the Effects of Vehicle Weight and Size
  Parameters on Fatality Risk in Model Year 1985-98 Passenger Cars and 1986-97 Light Trucks. Report No. DRI-
  TR-03-01. Torrance, CA: Dynamic Research, Inc.; Van Auken, R. M., and Zellner, J. W. (2005a). An
  Assessment of the Effects of Vehicle Weight and Size on Fatality Risk in 1985 to 1998 Model Year Passenger
  Cars and 1985 to 1997 Model Year Light Trucks and Vans.  Paper No. 2005-01-1354. Warrendale, PA: Society
  of Automotive Engineers; Van Auken, R. M., and Zellner, J. W. (2005b). Supplemental Results on the
  Independent Effects of Curb Weight, Wheelbase, and Track on Fatality Risk in 1985-1998 Model Year Passenger
  Cars and 1986-97 Model YearLTVs. Report No. DRI-TR-05-01. Torrance, CA: Dynamic Research, Inc.; Van
  Auken, R.M., and Zellner, J. W. (2011).2012a).  Updated Analysis of the Effects of Passenger Vehicle Size and
  Weight on Safety, Phase I. Report No. DRI-TR-11-01. (Docket No. NHTSA-2010-0152-0030).  Torrance, CA:
  Dynamic Research, Inc.

-------
                                                     Assessment of Vehicle Safety Effects
LTVs are defined as the subsets of vehicles above and below the median curb weight in fatal
crashes for a given group of vehicles (i.e., passenger cars or LTVs). Due to the increase in the
weight of the LTV fleet, the median curb weights used to define LTV classes are notably higher
than in the 2012 report, as detailed in Table 8.2
               Table 8.2 Passenger Car and LTV Classes in the 2012 and 2016 Analyses
Vehicle Class
Lighter Passenger Cars
Heavier Passenger Cars
Lighter LTVs
Heavier LTVs
2012 Report
< 3,106 pounds
3,106+ pounds
< 4,594 pounds
4,594+ pounds
2016 Analysis
< 3,197 pounds
3,197+ pounds
< 4,947 pounds
4,947+ pounds
Difference in Median
91 pounds
91 pounds
353 pounds
353 pounds
   The curb weight threshold defining passenger car classes in the update is only 91 pounds
higher, while the curb weight threshold defining LTV classes in the update is 353 pounds higher,
than the corresponding threshold in the 2012 report. The expected tendency of the influence of a
heavier light truck fleet is to magnify estimated beneficial effects for mass reduction in those
heavier LTVs, and to reduce estimated detrimental effects for lighter LTVs relative to the
previous analysis.

   The relatively short interval between the 2012 report and the update enables a generally direct
comparison of findings between the two studies.  However, there are at least two key empirical
outcomes associated with the updated safety dataset that limit its comparability with the 2012
analysis. Firstly, CY2009-2011  data replace CY2002-2004 data within the sample. New vehicle
registrations were below trend for CY2009-2011 (and hence, below corresponding levels in
CY2002-2004). In turn, and in conjunction with general (improving) trends in vehicle safely, the
number of fatal crashes in CY2009-2011 is about 25 percent lower than the number of crashes in
CY2002-2004.  Hence, the results of the analysis are calibrated with respect to a smaller number
of fatal crashes, resulting in larger estimated standard errors and associated confidence bounds
for the point estimates in the analysis.

   Secondly, as noted in the 2012 report, light-duty trucks (LTVs) began increasing in mass
around the year 2000; this trend  did not appear to abate for MY2008-2010 LTVs. The heavier
(relative to similar models from  previous model years on or near 2000) LTVs comprised a
relatively small share of the sample in the 2012 report, because relatively  early-model vehicles
comprise a much larger share of the observations in the database than late-model vehicles.
However, the sample in the update involves not only a large share of relatively heavy LTVs in
common with models in the 2012 report, but also MY2008-2010 vehicles that tend to be heavier
than the MY2000-2002 vehicles no longer in the sample.

   The analysis incorporates data from multiple sources required to represent fatalities, baseline
driving risk (i.e., induced exposure), and VMT across distributions of driver-, crash- and vehicle-
specific factors. The primary sources applied within the analysis are: the Fatality Analysis
Reporting System (FARS), State crash records, IHS Automotive's (formerly R.L. Polk & Co.),
National Vehicle Population Profiles (NVPP) and odometer readings, and a range of sources of
values for curb weight, footprint, track width, wheelbase and other vehicle attributes.

   FARS provides most of the information about fatal crashes needed for this study: the type of
crash and number of fatalities, the vehicle identification number (VIN) of the vehicles involved,

-------
                                                     Assessment of Vehicle Safety Effects
the age and gender of the driver(s), the time and location.  The 2005-2011 PARS files contain
85,890 records of crash-involved vehicles of model years 2003-2010 with decodable VTNs that
can be assigned a model year, curb weight, and footprint, and identified as passenger cars or
LTVs (pickup trucks, CUVs, truck-based SUVs and vans, excluding incomplete vehicles but
including "300-series" pickups and vans with GVWR sometimes over 10,000 pounds). The set
of PARS records in this analysis represents a decrease of around 24 percent relative to the 2012
analysis (113,248 records), due to both a general downward trend in fatalities and a decrease in
new vehicle registrations beginning in 2009.

   No single database has comparable detailed information on the number of total vehicles, their
drivers, and their use, which is necessary to estimate exposure in order to compute fatality risk
per VMT.  The NVPP data count the number of vehicles of a given make-model and model year
registered in any calendar year. The NVPP data specify the number of vehicles registered as of
July 1 of every calendar year, and provide estimates of vehicle registrations by MY, CY,  vehicle
group, make-model, body style/truck type and, where needed, by State. NVPP data have no
information, for example, on the age or gender of the drivers, or the annual VMT, or whether the
vehicles were driven by day or at night.  A file of odometer readings, also supplied by IHS
Automotive was used to derive estimates of annual VMT by make and model.

   Police-reported crash data from 13 states were used to develop the induced-exposure crashes;
the state crash data provide information on not only the vehicles involved but also driver  age and
gender, urban/rural and other characteristics corresponding to the PARS  data. Induced-exposure
crashes are a subset of two-vehicle collisions where one vehicle can be identified as "culpable"
and the other as "non-culpable." The distribution of such vehicles within a particular area is
believed to be an essentially random sample of driver and vehicle combinations travelling
through that area. Accurate estimates of the curb weight and footprint of vehicles, as well as
other attributes such as the presence of electronic stability control (ESC), antilock brake systems
(ABS), and side or curtain air bags are assembled from several publications.

   The State data represent a sample of 13 States that provide the VIN (all in common with the
2012 report): Alabama, Florida, Kansas, Kentucky, Maryland, Michigan, Missouri, Nebraska,
New Jersey, Pennsylvania, Washington, Wisconsin and Wyoming.  The  State data include
2,255,398 records of induced-exposure cases, a decrease of around eight percent relative  to the
2012 database (2,457,228 records), compared to a 24 percent decrease in PARS records relative
to the 2012 database. The difference in sizes of the State and PARS data between the 2012 and
2016 reports indicate the presence of a larger decrease in the fatality rate than in the crash rate
between the two samples.

   The 85,890 records in the database of PARS fatal crash involvements come from all 50 States
and the District of Columbia.  Each of the 2,255,398 records on the database of induced-
exposure crash involvements is nominally a specific crash involvement in one of 13 States, a
discrete unit.  But when each induced exposure record is weighted by its allocation of vehicle
registrations or VMT, it becomes a cohort of vehicle registrations or VMT in the United States.
The weighted induced-exposure records are a national census of model year 2003 to 2010
vehicle registrations and VMT in each calendar year. Fatal-crash records are weighted by the
number of fatalities in the crash, including fatalities in the crash partner vehicle and any cyclists
or pedestrians. After combining the PARS and induced exposure data, the sum of the fatalities in

-------
                                                      Assessment of Vehicle Safety Effects
the fatal crashes divided by the sum of the VMT in the induced exposure crashes is the national
fatality risk per mile driven, which serves as the dependent variable in the regression analyses.

   The curb weight of passenger cars is formulated, as in previous reports, as a two-piece linear
variable in order to estimate one effect of mass reduction in the lighter cars and another effect in
the heavier cars. The boundary between "lighter" and "heavier" vehicles is itemized in Table 8.2
above. Curb weight is formulated as a simple linear variable for CUVs and minivans: because
CUVs and minivans account for a relatively small share of new-vehicle sales, there are less crash
data available than for cars or truck-based LTVs.

   Footprint (in square feet) is represented in the model as the product of track width (the
average track width at the front and rear wheels) and wheelbase. The control variables in the
model include:  indicators of whether an incident occurred at night, in a rural county, on roads
with speed limits 55 miles per hour or above, and in States with relatively high fatality rates;
indicators of whether a vehicle is equipped with ESC, anti-lock brakes, all-wheel drive, curtain
airbags, curtain airbags that deploy in rollovers, torso airbags,  combination airbags that provide
torso and head protection, and light truck compatibility certification meeting Options 1 or 2;
vehicle age at the time of incident;  an indicator if the vehicle is new (i.e., MY=CY); eight
gender-specific driver age categories; driver gender; and indicators of calendar year.

   Separate logistic regressions were estimated for the three vehicle classes: passenger cars,
LTVs, and CUVs/minivans. Within each class in the  analysis, separate logistic regressions were
estimated across nine sets of crash types, including: first-event rollovers; collisions with fixed
objects, pedestrians/bicyclists/motorcyclists, heavy vehicles, passenger cars/CUVs/minivans
lighter than 3,157 pounds, passenger cars/CUVs/minivans 3,157 pounds or heavier, LTVs lighter
than 4,303 pounds, and LTVs 4,303 pounds or heavier; and all other crashes (mostly crashes
involving three or more vehicles).  A separate regression model was run for each of the nine
crash types within each of the three vehicle  types, for a total of 27 regression models.

   Consistent with the definition of vehicle  classes, the threshold weights for crash types
involving passenger cars/CUVs/minivans and LTVs were defined in both the 2012 report and
this analysis as  the median curb weight for the other vehicle in a fatal collision.  Similar to the
changes to the mass thresholds defining vehicle classes in this analysis, the mass thresholds for
crash types increased in the new analysis. The mass threshold for crashes with passenger
cars/CUVs/minivans increased 75 pounds (from 3,082 pounds), and the mass threshold for
crashes with LTVs increased 153 pounds (from 4,150 pounds). These increases are smaller than
the corresponding increases in the thresholds for vehicle classes, due to the presence of MY2002
and earlier vehicles as partner vehicles in two-vehicle crashes.

   For each vehicle class, a composite estimate of the change in societal fatality risk with respect
to curb weight was identified by weighting the estimated coefficients on curb weight for a given
crash type by the (adjusted) number of fatalities observed in the crash type for the vehicle class.
The adjustment to the number of fatalities observed in a given crash type for a given vehicle
class involves a downward revision to fatalities to take into account that the results will be used
to analyze effects of mass reduction in future vehicles, which will all be equipped with electronic
stability control (ESC), as required by NHTSA's regulations.  That is, although some vehicles in
the database did not have ESC (and hence are more likely to be in a crash than ESC-equipped
vehicles), all new vehicles are equipped with ESC; the lack of an adjustment would overstate the
expected volume of fatalities that changes in curb weight could influence.

-------
                                                       Assessment of Vehicle Safety Effects
   Table 8.3 presents the 2012 report's estimated percent increase in U.S. societal fatality rates
per ten billion VMT for each 100-pound reduction in vehicle mass, while holding footprint
constant, for each of the five vehicle classes:

Table 8.3 Results of 2012 NHTSA Final Report: Fatality Increase (%) per 100-Pound Mass Reduction While
                                  Holding Footprint Constant
MY2000-2007
CY 2002-2008
Cars< 3,106 pounds
Cars > 3,106 pounds
CUVs and minivans
Truck-based LTVs < 4,594 pounds
Truck-based LTVs > 4,594 pounds
Fatality Increase (%) Per 100-Pound Mass Reduction While
Holding Footprint Constant
Point Estimate
1.56
.51
-.37
.52
-.34
95% Confidence Bounds
+ .39 to +2.73
- .59 to +1.60
-1.55 to + .81
- .45 to +1.48
-.97 to + .30
   Table 8.4 presents the 2016 preliminary report's estimated percent increase in U.S. societal
fatality risk per ten billion VMT for each 100-pound reduction in vehicle mass, while holding
footprint constant, for each of the five classes of vehicles:

Table 8.4 Results of 2016 NHTSA Preliminary Report: Fatality Increase (%) per 100-Pound Mass Reduction
                               While Holding Footprint Constant
MY2003-2010
CY 2005-2011
Cars < 3,197 pounds
Cars > 3,197 pounds
CUVs and minivans
Truck-based LTVs < 4,947 pounds
Truck-based LTVs > 4,947 pounds
Fatality Increase (%) Per 100-Pound Mass Reduction While
Holding Footprint Constant
Point Estimate
1.49
.50
-.99
-.10
-.72
95% Confidence Bounds
- .30 to +3.27
- .59 to +1.60
-2. 17 to + .19
- 1.08 to +.88
- 1.45 to + .02
   The results indicate that societal fatalities per VMT would increase if the mass of passenger
cars (the two lightest vehicle classes in the analysis by median weight) were reduced. Mass
reduction in passenger cars below 3,197 pounds is estimated to increase societal fatality risk
when holding footprint constant; a 100-pound reduction in curb weight is estimated to increase
net fatalities by 1.49 percent. Mass reduction in passenger cars 3,197 pounds and above is
estimated to increase societal fatality risk when holding footprint constant; a 100-pound
reduction in curb weight is estimated to increase net fatalities by 0.50 percent.

   Conversely, the results indicate that societal  fatalities per VMT would decrease if the mass of
LTVs, CUVs and minivans were reduced. Mass reduction in LTVs 4,947 pounds and above is
estimated to decrease  societal fatality risk when holding footprint constant;  a  100-pound
reduction in curb weight is estimated to reduce  net fatalities by 0.72 percent.  Likewise, mass
reduction in CUVs and minivans (the second-heaviest vehicle class in the analysis by median
weight) is estimated to decrease societal fatality risk when holding footprint constant; a 100-
pound reduction in curb weight is estimated to reduce net fatalities by 0.99 percent.  Mass
reduction in LTVs below 4,947 pounds is  estimated to decrease societal fatality risk only slightly
when holding footprint constant; a 100-pound reduction in curb weight is estimated to decrease
net fatalities by 0.10 percent.

-------
                                                      Assessment of Vehicle Safety Effects
   None of the estimated effects is statistically significant at the 95-percent confidence level (i.e.,
the confidence bounds include both positive and negative values; the estimate for heavier LTVs
is very close, however (statistically significant at the 94-percent confidence level). Three of the
five estimated effects of mass reduction on societal fatalities are statistically significant at the 90-
percent confidence level, for lighter passenger cars, heavier LTVs, and CUVs and minivans,
indicating a strong likelihood that at least some of the estimated effects are significantly different
from zero.

   The principal difference between the results for heavier vehicles, especially truck-based
LTVs, and lighter vehicles, especially passenger cars, is that mass reduction has a different effect
in collisions with another light-duty vehicle in cars than in light trucks. When two vehicles of
unequal mass collide, the change in velocity ("delta V") is higher in the lighter vehicle, in the
same proportion as the mass ratio.  As a result, all else being equal, the fatality risk in the lighter
vehicle is also higher.

   Removing some mass from the heavy vehicle reduces delta V in the lighter vehicle, where
fatality risk is high, resulting in a large benefit, offset by a  small penalty because delta V
increases in the heavy vehicle, where fatality risk is low - adding up to a net societal benefit.
Removing some mass from the lighter vehicle results in a large penalty offset by a small benefit
- adding up to net harm.  These considerations drive the overall result: mass reduction in lighter
cars is associated with an increase in societal fatalities, mass reduction in the heavier LTVs is
associated with a decrease in societal fatalities, and mass reduction in the intermediate classes
has little effect.

   It is useful to compare the results from the 2012 and 2016 reports (as detailed in Table 8.Sand
Table 8.4). In general, the point estimates from the updated analysis are consistent with the
findings in the 2012 report. The ranges of the updated confidence bounds are similar size to the
corresponding values in the 2012 report for heavier passenger cars (a range of 2.19 percent in
both cases), lighter LTVs (1.96 percent in the updated analysis versus 1.93 percent in the 2012
report) and minivans (2.36 percent in both cases). This result may be unexpected, in light of the
decreased sample size for fatal incidents in the update relative to the 2012 report (i.e., a smaller
sample size tends to yield larger confidence bounds). The  range of the confidence bound for
lighter passenger cars is notably larger in the update (3.57 percent versus 2.34 percent), while the
range of the confidence bound for heavier LTVs is only somewhat larger in the update (1.47
percent versus 1.27 percent).

   The 2012 report presented one point estimate that was statistically  significant at the 95-
percent confidence level: the estimate for lighter passenger cars. The updated analysis yielded
no point estimates that are significant at the 95-percent  confidence level (the estimate for heavier
LTVs was just short of this threshold).  However, the updated analysis did yield three estimates
that would be statistically significant at the 90-percent confidence level, compared to one
estimate in the 2012 report: the estimates for lighter passenger cars, heavier LTVs, and CUVs
and minivans. Hence, although the updated analysis indicates a greater level of uncertainty
about the value of any given point estimate relative to the 2012 report (i.e., no estimated
coefficients  are significant at the 95-percent confidence level, versus one significant coefficient
in the 2012 report), the updated analysis also indicates a greater level  of certainty that at least
some of the  point estimates are of a particular sign (i.e., three estimated coefficients would be
significant at the 90-percent confidence level, versus one in the 2012 report).

-------
                                                      Assessment of Vehicle Safety Effects
   Two of the five updated point estimates are very close to the corresponding values in the 2012
report (the estimates for passenger car classes). This is consistent with the relatively small
change in the definition of the two passenger car classes in the update (i.e., the updated threshold
curb weight value is around 100 pounds heavier than in the 2012 report).  Furthermore, the
directionality of the changes in the point estimates for passenger cars are consistent with the
change in the threshold curb weight value (i.e., mass reduction for a heavier group of vehicles
should be more beneficial or less detrimental to society than for a lighter group of vehicles).

   The updated point estimates for LTVs are distinct from the corresponding values in the 2012
report. The directionality of the changes in the point estimates for LTVs is consistent with the
relatively large change in the threshold curb weight (around 350 pounds heavier in the update).
While the 2012 report indicated that mass reduction of lighter LTVs would lead to an increase in
net fatalities, the updated analysis indicates that, conditional on the observed increase in curb
weight for LTVs in general, mass reduction of lighter LTVs would lead to a decrease in net
fatalities.  Likewise, the 2012 report indicated that mass reduction of heavier LTVs would lead to
a decrease in net fatalities; the updated analysis indicates that, conditional on the observed
increase in curb weight for LTVs in general, this relationship has become stronger.

   The updated point estimates for CUVs and minivans are the most distinct from the
corresponding values in the 2012 report, but still of the same  sign. The directionality of the
change in the point estimate for CUVs and minivans is consistent with a general increase in
vehicle mass. However, there are factors limiting the inference one can draw from estimates in
this vehicle class. Chiefly, the range of curb weights for minivans is relatively small, which may
amplify the estimated impact of curb weight on fatality risk.

   The estimates in Table 8.3 and Table 8.4 of the model are formulated for each 100-pound
reduction in mass; in other words, if risk increases by 1 percent for 100 pounds reduction in
mass, it would increase by 2 percent for a 200-pound reduction, and 3 percent for a 300-pound
reduction.  Confidence bounds around the point estimates will grow wider by the same
proportions.

   The regression results are best suited to predict the effect of a small change in mass, leaving
all other factors, including footprint, the same.  With each additional change from the current
environment, the model may become somewhat less accurate and it is difficult to assess the
sensitivity to additional mass reduction greater than 100 pounds. The agencies recognize that the
light-duty vehicle fleet in the MYs 2022-2025 timeframe will be different from the MYs 2003-
2010 fleet analyzed for this study. Nevertheless, one consideration provides some basis for
confidence in applying the regression results to estimate the effects of mass reductions larger
than 100 pounds or over longer time periods. This was NHTSA's fifth evaluation of the effects
of mass reduction and/or downsizing, comprising databases ranging from MYs 1985 to 2010.
The results of the five studies are not identical, but they have been consistent up to a point.
During this time period, many makes and models have increased substantially in mass,
sometimes as much as 30-40 percent.N  If the statistical analysis has, over the  past years, been
NFor example, one of the most popular models of small 4-door sedans increased in curb weight from 1,939 pounds
  in MY 1985 to 2,766 pounds in MY 2007, a 43 percent increase. A high-sales mid-size sedan grew from 2,385 to
  3,354 pounds (41%); a best-selling pickup truck from 3,390 to 4,742 pounds (40%) in the basic model with 2-
  door cab and rear-wheel drive; and a popular minivan from 2,940 to 3,862 pounds (31%).

-------
                                                      Assessment of Vehicle Safety Effects
able to accommodate mass increases of this magnitude, perhaps it will also succeed in modeling
the effects of mass reductions on the order of 10-20 percent, if they occur in the future.

   NHTSA's 2012 report acknowledged another source of uncertainty, namely that the baseline
statistical model can be varied by choosing different control variables or redefining the vehicle
classes or crash types, for example. Alternative models produce different point estimates. The
principal comments on the preliminary version of the 2012 report were suggestions or
demonstrations of other ways to analyze NHTSA's database, especially by Farmer and Green in
their peer reviews, Van Auken (DRI) in his most recent analyses, and Wenzel in his assessment
of NHTSA's report. The analyses and findings of Wenzel's and Van Auken's reports are
summarized below. These reports, among other analyses, define and run specific alternative
regression models to analyze NHTSA's 2012 databases.0

   From these suggestions and demonstrations, NHTSA garnered 11  more or less plausible
alternative techniques that could be construed as sensitivity tests of the baseline model; these
alternative model structures were evaluated in the 2011, 2012 and 2016 reports.1"  The models use
NHTSA's databases and regression-analysis approach, but differ from the baseline model in one
or more terms  or assumptions. All of them try to control  for fundamentally the same  driver,
vehicle, and crash factors, but differ in how they define these factors  or how much detail or
emphasis they provide for some of them.  NHTSA applied the 11 techniques to the latest
databases to generate alternative  estimates of the societal effect of 100-pound mass reductions in
the five classes of vehicles. The range of estimates produced by the sensitivity tests gives an
idea of the uncertainty inherent in the formulation  of the models, subject to the caveat that these
11 tests are, of course, not an exhaustive list of conceivable alternatives.

   Each model in the sensitivity analysis estimates fatality rates as a function of curb  weight,
vehicle size, driver-specific attributes and incident-specific attributes. The baseline model
represents vehicle size in terms of footprint (i.e., the product of wheelbase and track width,
measured in square feet), and is calibrated with respect to FARS data (the fatal outcomes in the
logistic regressions) and induced exposure data incorporating non-culpable incidents  across a
sample of 13 states; the FARS data are a census of fatal incidents, while  the induced exposure
data are weighted to represent all VMT for each make-model-model year combination in each
calendar year in the sample.

   One alternative model represents induced exposure through the subset of non-culpable cases
in the sample involving stopped vehicles (referred to here as the stopped vehicle model).  This
alternative was proposed under the hypothesis that restricting the analysis to  stopped vehicles
would minimize any bias due to uncertainty regarding which driver was  at fault in the two-
vehicle crash,  and improve the degree to which the induced exposure data represent baseline
accident risk.  Furthermore, DRI assumed that the  set of non-culpable incidents may induce bias
because relatively skilled drivers may be more likely to avoid crashes altogether, and  hence
relatively skilled drivers would be under-represented. If this bias is present, the resulting
estimates would over-represent the behavior of relatively unskilled drivers.
0 Wenzel (2012a), Van Auken and Zellner (2012b, 2012c, 2012d).
p See Kahane (2012), pp. 14-16 and 109-128 for a further discussion of the alternative models and the rationales
  behind them.

-------
                                                       Assessment of Vehicle Safety Effects
   The other alternative model represents vehicle size in terms of track width and wheelbase
separately (referred to here as the split footprint model). DRI proposed this alternative under the
hypothesis that vehicle size could be accounted for independently of curb weight more
effectively by representing distinct effects of track width (e.g., rollover resistance) and
wheelbase (e.g., crush space in frontal impacts).  This alternative can be applied using either the
baseline induced exposure data (as represented in the analytical results below), or combined with
the application  of stopped vehicle data.

   The sensitivity analyses examined the stopped vehicle and split footprint alternatives to re-
evaluate the limitations of the alternatives  that were raised in the 2012 report, to confirm whether
the limitations still apply. The primary limitations of the  stopped vehicle model raised in the
2012 report that apply to the data in the 2016 report are:

    •  Restricting the analysis to stopped  vehicles results in a serious loss of sample size;
    •  The stopped vehicle cases represent the distribution of driver age disproportionately;
    •  The stopped vehicle cases represent the share of incidents on roads with speed limits 55
       miles per hour or above disproportionately; and
    •  Comments from previous (1999 and 2003) peer review support the use of the baseline
       model over the stopped vehicle model.
   Each of the above limitations applies to the analysis in the 2016 report. Restricting the
analysis to stopped vehicles results in a loss of approximately three-fourths of observations in the
sample; estimates calibrated with respect to a restricted sample size are subject to greater
uncertainty (i.e., larger confidence bounds) than those calibrated with respect to a larger  set of
data. The stopped vehicle database includes 670,230 observations, which is a large dataset by
general standards.  However, driver-, crash- and vehicle-specific factors explain such a large
share of variability in fatality rates that it is preferable to preserve sample size in  an effort to
estimate effects specific to  curb weight and vehicle size, all else being equal.

   Consistent with the 2012 report, the stopped vehicle data in the 2016 report represent  drivers
with ages associated with lower risk  (i.e., drivers between 30 and 60 years of age) at a higher rate
than the non-culpable data, and conversely represent drivers with ages associated with higher
risk (chiefly, drivers below the age of 30) at a lower rate than the non-culpable data. Similarly,
as in the 2012 report, the stopped vehicle data include a smaller share of incidents: on roads with
speed limits of 55 miles per hour or above; on rural roads; at night; and involving male drivers.

   However, the non-culpable data are constrained by the relative accuracy of police
identification of at-fault drivers.  If the non-culpable cases actually include a sufficient share of
culpable cases,  the data would not meaningfully represent baseline risk. Hence, the findings of
analysis calibrated with respect to the non-culpable data are strictly  conditional on the validity of
the assignment of culpability. Peer review indicated two conflicting views: (1) that stopped
vehicle data under-represent risky drivers because risky drivers do not stay stopped long enough
to be involved in collisions; and  (2) that non-culpable vehicle data over-represent drivers because
safe drivers avoid incidents more frequently. It is not clear whether the non-culpable vehicle
sample or the stopped vehicle sample better represents the overall distribution of drivers  and
vehicles on the nation's roadways, and therefore which sample is more appropriate to use to
create the induced exposure records.

-------
                                                       Assessment of Vehicle Safety Effects
   Peer review comments on the preliminary version of the 2012 report suggested that a suitable
representation of induced exposure would involve distributions of VMT by vehicle-, crash- and
driver-specific factors that represent the population of drivers and vehicles on the road at any
given time.

   The limitations of the split footprint model raised in the 2012 report that apply to the data in
the 2016 report are:

   •   Track width and wheelbase are generally highly correlated with one another and with
       curb weight for the range of vehicles in the analysis, raising the threat of
       multicollinearity;
   •   The CAFE model is footprint-based, and hence working directly with footprint is
       preferable to decomposing it; and
   •   While the estimated relationship between track width and fatality risk in certain types of
       crashes is consistent with crash physics, the relationship between wheelbase and fatality
       risk is not.
   The threat of multicollinearity can be evaluated in a direct manner by  comparing correlations
among model inputs. Multicollinearity is a significant concern even in the baseline model,
through strong correlations between curb weight and footprint; correlations within vehicle
classes range from around 0.73 to 0.89, (with the exceptions of correlations of around 0.24 for
large pickups and 0.49 for minivans when examined separately from other LTVs and CUVs,
respectively).

   Critically, for all vehicle classes in the analysis, curb weight is correlated either nearly as high
or higher with track width as with footprint and track width and wheelbase are also highly
correlated with one another (ranging from around 0.64 to 0.80, with the exceptions of smaller
correlations for large pickups and minivans). Viewed from another angle, wheelbase is almost
perfectly correlated with footprint (with correlations ranging from around 0.95 to 0.97).

   Considered in concert, the split footprint model essentially incorporates the full correlation
issues from the baseline model (curb weight highly correlated with another independent variable)
and adds a further correlation issue (the variable that is highly  correlated with curb weight is also
highly correlated with a separate independent variable). Ultimately,  it is difficult to support the
preference of a model with two correlated independent variables representing vehicle size when
a single variable (footprint) tracks the two variables closely. The ability of the model to tease out
separate, representative effects for three highly correlated variables is questionable; what may
appear to be a distinct effect once two dimensions of vehicle size are accounted for may in fact
be an artifact of unfortunate statistical properties.

   In the 2016 NHTSA baseline model, a one-inch reduction in track width is associated with
increases in rollover fatality risks, as expected: a 30 percent increase in rollover fatality risk in
cars, and an 8 percent increase in rollover fatality risk in light trucks and CUVs/minivans.
However, a one-inch reduction in wheelbase is not consistently associated with large increases in
fatality risks in crashes with objects or other light-duty vehicles. This may be because wheelbase
is not as good a proxy for frontal crush space, as say frontal overhang*2 in frontal impacts; and
because a large fraction of fatalities in two-vehicle crashes are not frontal impacts that would be
Q Frontal overhang is the distance from the front of the front bumper to the front wheel axle.

-------
                                                       Assessment of Vehicle Safety Effects
influenced by wheelbase or frontal overhang (i.e. they are the result of side impacts).  So the
regression coefficients for track width are consistent with crash theory, while the coefficients for
wheelbase are not, possibly because they are masked by other types of crashes in which frontal
crush space is not expected to protect occupants.

   Table 8.5  shows the baseline and alternative results, ordered from the lowest to the highest
estimated increase in societal risk for cars weighing less than 3,197 pounds:
Table 8.5 Societal Fatality Increase (%) Per 100-Pound Mass Reduction While Holding Footprint* Constant

Baseline estimate
95% confidence bounds (sampling error) Lower:
Upper:
Cars
< 3,197
1.49
-.30
3.27
Cars
> 3,197
.50
-.59
1.60
CUVs&
Minivans
-.99
-2.17
.19
LTVst
< 4,947
-.10
-1.08
.88
LTVst
> 4,947
-.72
-1.45
.02
11 Alternative Models
1. W/O CY control variables
2. Track width/wheelbase w. stopped vehicle data
3. By track width & wheelbase
4. Incl. muscle/police/ AWD cars/big vans
5. W/O non-significant control variables
6. CUVs/minivans weighted by 2010 sales
7. With stopped-vehicle data
8. Limited to drivers with BAC=0
9. Control for vehicle manufacturer
10. Control for vehicle manufacturer/nameplate
11. Limited to good drivers*
.53
.88
.92
1.44
1.47
1.49
1.58
2.22
2.39
2.65
2.82
.10
-.43
.48
.63
.54
.50
-.43
1.38
1.37
2.96
1.86
-1.13
-.66
-1.15
-.99
-.84
-.27
-.61
-.92
.00
-.43
-.97
-.10
-.85
-.66
-.05
-.13
-.10
-.07
.31
.32
.30
.37
-.53
-2.14
-.97
-.94
-.70
-.72
-1.80
-.91
-.09
.00
-.62
Notes:
       * While holding track width and wheelbase constant in alternative model nos. 1 and 3.
       f Excluding CUVs and minivans.
       {Blood alcohol content=0, no drugs, valid license, at most 1 crash and 1 violation during the past 3 years.


   For example, in cars weighing less than 3,197 pounds,  there are an equal number of models
with estimated effects of 100-pound mass reduction above and below the baseline value, a 1.49
percent increase in societal fatalities. The estimates range from a relatively small increase of
0.53 percent in the first alternative model up to a 2.82 percent increase in the last model, nearly
double the baseline effect.  Each of the 11 alternative point estimates for cars < 3,197 pounds is
within the range of the 95 percent sampling-error confidence bounds for the baseline estimate: -
0.30 to 3.27 percent.

   The sensitivity tests illustrate both the fragility and the robustness of the baseline estimate.
On the one hand, the variation among the alternative estimates is quite large relative to the
baseline estimate: in the preceding example of cars < 3,197 pounds, from approximately one-
third of the baseline value to almost double the baseline. In fact, the difference in estimates is a
reflection of the small statistical effect that mass reduction has on societal risk, relative to other
factors. Thus, sensitivity tests which vary vehicle, driver, and crash factors can appreciably
change the estimate of the effect of mass reduction on societal risk in relative terms.

-------
                                                     Assessment of Vehicle Safety Effects
   On the other hand, the variations are not all that large in absolute terms. The ranges of the
alternative estimates, at least these alternatives, are about as wide as the sampling-error
confidence bounds for the baseline estimates.  As a general rule, in the alternative models, as in
the baseline models, mass reduction tends to be relatively more harmful in the lighter vehicles,
and more beneficial in the heavier vehicles.  Thus, in all models, the estimated effect of mass
reduction is a societal fatality increase for cars < 3,197 pounds, and in all models except one, a
societal fatality reduction for LTVs > 4,947 pounds. None of these models suggest mass
reduction in small cars would be beneficial.  All suggest mass reduction in heavy LTVs would be
beneficial or,  at least, close to neutral. In general, any judicious combination of mass reductions
that maintain  footprint and are proportionately higher in the heavier vehicles is unlikely to have a
societal effect large enough to be detected by statistical analyses of crash data. NHTSA
conducted a sensitivity analysis to estimate the fatality impact of the alternative models using the
coefficients for these 11 test cases. The results for these sensitivity runs can be found in Table 4-
2 of NHTSA's 2016 preliminary report. The discussion of the 2016 preliminary report concludes
with a review of the limitations of the analysis, and corresponding implications for the
interpretation and application of the results.  The presence of non-significant results in this
analysis is not due to a paucity  of data (except, perhaps, the paucity of very small or very light
cars and LTVs during MY2003-2010) or other weaknesses in the data, but because the  societal
effect of mass reduction while maintaining footprint, if any, is small.  By contrast, statistical
analyses of the effect of mass reduction allowing historically commensurate reductions of
footprint (downsizing) show larger, statistically significant increases in fatality risk in passenger
cars (see Alternative regression model 6 in Table 8.6 from the 2016 LBNL Phase2 study
presented in the following sub-sections).

   The composite effects are limited in significance, with estimated effects for three of five
vehicle classes significant at the 90-percent confidence level. However, this does not indicate
that the non-significant estimated composite effects should be ignored. We include and apply
non-significant estimates because the regulatory analysis must provide the best estimate of the
expected effect of mass reduction. Our best estimate is the estimated  composite effect (i.e., an
estimate of zero would be a worse fit to the data); the confidence bounds serve to indicate the
range of uncertainty. One reason that the regulatory analysis must have such estimates is that it,
too, is ultimately an intermediate computational tool in estimating the overall health and societal
impact of CAFE and GHG regulation.

   The estimates of this report are based on statistical analyses of historical data, which puts
some limitations on their value for predicting the effects of future mass reductions. Analyses of
historical data necessarily lag behind the latest developments in vehicles and in driving patterns
because it takes years for sufficient crash data to accumulate. It is important to note that while
the MY2003-2010 database represents more modern vehicles with technologies more
representative of vehicles on the road today than previous reports, it still does not represent the
newer vehicles that will be on the road in the 2022-2025 timeframe. The vehicles manufactured
in the 2003-2010 timeframe were not subject to a footprint-based fuel-economy standard;
vehicles actually became heavier on the average, not lighter during MY2003-2010 and when
they became heavier it was commonly to provide additional features.  NHTSA and EPA expect
that the attribute-based standard will affect the design of vehicles such that manufacturers may
reduce mass while maintaining footprint more than has occurred prior to model year 2010.

-------
                                                      Assessment of Vehicle Safety Effects
Therefore, it is possible that the analysis for 2003-2010 vehicles may not be fully representative
of those vehicles that interact with the existing fleet in 2022 and beyond.

   Statistical analyses can control for many factors such as a driver's age and gender, but there
are other factors they do not control for, such as driver characteristics that cannot be quantified
with available demographic variables or unobserved factors relating to how a particular vehicle
was being driven at the time the crash occurred (e.g., travel speed, attention). Furthermore, the
analyses of this report are "cross-sectional": they compare the fatality rates for vehicles weighing
100 pounds  less than other models in the same vehicle class, rather than directly comparing the
fatality rates for a specific make and model before and after a mass reduction had been
implemented for the purpose of improving fuel economy. After substantial materials substitution
has become  more widespread, it may become feasible to improve the ability to directly compare
the effects of mass reductions at the vehicle-model level.  However, such models would still be
limited in their ability to represent other design changes that influence fatalities beyond mass
reduction.

8.2.4.6Report by Tom Wenzel,LBNL, "An Assessment of NHTSA's Report 'Relationships
between Fatality Risk, Mass, and Footprint in Model Year 2003-2010 Passenger Cars and
LTVs,'"2016

   DOE contracted with Tom Wenzel of Lawrence Berkeley National Laboratory to conduct an
assessment of NHTSA's updated 2016 study of the effect of mass and footprint reductions on
U.S. fatality risk per vehicle miles traveled (LBNL 2016 "Phase 1" preliminary report), and to
provide an analysis of the effect of mass and footprint reduction on casualty risk per police-
reported crash, using independent data from thirteen states (LBNL 2016 "Phase 2" preliminary
report). Both reports will be reviewed by NHTSA, EPA, and DOE staff, as well as by  a panel of
reviewers.R  The final versions of the reports will reflect responses to comments made in the
formal review process, as well as changes made to the  VMT weights developed by NHTSA for
the final rule, and inclusion of 2012 data for  13 states that were not available for the analyses in
the preliminary versions included in the NPRM docket.

   The 2016 LBNL Phase 1 report27 replicates Volpe's 2016 analysis for NHTSA, using the
same data and methods, and in many cases using the same SAS programs, in order to confirm
NHTSA's results.  The LBNL report confirms  NHTSA's 2016 finding that, holding footprint
constant, each 100-lbs  of mass reduction is associated with a 1.49 percent increase in fatality risk
per vehicle miles travelled (VMT) for cars weighing less than 3,197 pounds, a 0.50 percent
increase for  cars weighing more than 3,197 pounds, a 0.10 percent decrease in risk for light
trucks  weighing less than 4,947 pounds, a 0.71 percent decrease in risk for light trucks  weighing
more than 4,947 pounds,  and a 0.99 percent decrease in risk for CUVs/minivans.s  Holding mass
constant, each square foot reduction in vehicle footprint is associated with a 0.28 percent
REPA sponsored the peer review of the LBNL 2011 Preliminary Phase 1 and 2 Reports.
s Only the changes in fatality risk for lighter cars, heavier trucks, and CUVs/minivans are statistically significant at
  the 95% significance level using the standard errors output by SAS. The relationship between mass reduction and
  fatality risk for these three vehicle types also is statistically significant at the 90% level of significance based on
  NHTSA's estimate of uncertainty using a jack knife method; none of the estimates are statistically significant at
  the 95% level of significance based on NHTSA's jack knife uncertainty method.

-------
                                                      Assessment of Vehicle Safety Effects
increase in risk in cars, a 0.38 percent increase in light trucks, and a 1.18 percent increase in
CUVs/minivans.T Wenzel tested the sensitivity of these estimates to changes in the measure of
risk and the control variables and data used in the regression models. Wenzel also concluded
that there is a wide range in fatality risk by vehicle model for models that have comparable mass
or footprint, even after accounting for differences in drivers' age and gender, safety features
installed, and crash times and locations.  This section summarizes the results of the 2016 Wenzel
assessment of the 2016 NHTSA preliminary analysis.

   The 2016 LBNL Phase  1 report notes that many of the control variables NHTSA includes in
its logistic regressions are statistically significant, and have a much larger estimated effect on
fatality risk than vehicle mass. For example, installing torso side airbags, electronic stability
control, or an antilock braking system in a car is estimated to reduce fatality risk by at least 7
percent; cars driven by men are estimated to have a 40 percent higher fatality risk than cars
driven by women; and cars driven at night, on rural roads, or on roads with a speed limit higher
than 55 mph are estimated to have a fatality risk over 100 times higher than cars driven during
the daytime on low-speed non-rural  roads. While the estimated effect of mass reduction may
result in a statistically-significant increase in risk in certain cases, the increase is small and is
overwhelmed by other known vehicle, driver, and crash factors.

   As was true in 2012, NHTSA in 2015 notes these findings are additional evidence that
estimating the effect of mass reduction is a complex statistical  problem, given the presence of
other factors that could have large effects. The preceding examples are limited to technologies
emerging in the 2005-2011 timeframe but that will be in all model year 2017-2025 vehicles (side
airbags, electronic stability control) or factors that are simply unchangeable circumstances in the
crash environment outside the control of CAFE or other vehicle regulations (for example, that
about half of the drivers are males and that much driving is at night or on rural roads).

   LBNL tested the sensitivity of the NHTSA estimates of the  relationship between vehicle
weight and risk using 33 different regression analyses that changed the measure of risk, the
control variables used, or the data used in the regression models. LBNL analyzed alternative
models 1 through 19 in its  2012 assessment of the NHTSA 2012 report; the results from these
models using data updated through 2011 are shown in Table 8.6. Table 8.6 also shows the
results of the 14 new alternative regression models LBNL conducted as part of its 2016
assessment.  Models 20 through 23  explore two changes to how light trucks are classified:
excluding light trucks with a GVWR rating over 10k pounds, and treating small (1/2-ton
capacity) pickups and SUVs as a separate class distinct from large (3/4- and 1-ton capacity)
pickups.  As noted in the Table 8.6 footnotes, the median weight was recalculated for each
alternative truck category.  Models 24 through 27 test the sensitivity to which cars are included;
Models 28 through 30 add a two-piece variable for CUV/minivan curb weight, based on the
median CUV/minivan curb weight, as was done for cars and light trucks in the NHTSA baseline
model; and two-piece variables for footprint for all vehicle types, based on the median footprint
by vehicle type.  Finally, Models 31 to 33 replace NHTSA's VMT weights with weights
developed from annual odometer readings in Texas.
T Based on the standard errors output by SAS, only the increases in risk from footprint reduction in light trucks and
  CUVs/minivans are statistically significant at the 95% confidence level.

-------
                                                             Assessment of Vehicle Safety Effects
 Table 8.6 Societal Fatality Increase (%) Per 100-Pound Mass Reduction While Holding Footprint* Constant
                                          from Wenzel Study


Regression model
Baseline model
1. Weighted by current distribution of fatalities
2. Single regression model across all crash types
3. Fatal crashes per VMT
4.Fatalities per induced exposure crash
5. Fatalities per registered vehicle-year
6. Allow footprint to vary with mass2
7. Account for 14 vehicle manufacturers
S.Account for 14 manufacturers + 5 luxury brands
9.Account for initial vehicle purchase price
10. Exclude CY variables
1 1. Exclude crashes with alcohol/drugs
12. Exclude crashes with alcohol/drugs, and bad drivers
13. Account for median household income
14.1nclude sports, police, and AWD cars, and full vans
15. Use stopped instead of non-culpable vehicles
16. Replace footprint with track width & wheelbase
IV.Above two models combined (15 & 16)
IS.Reweight CUV/minivans by 2010 sales
19. Exclude non-significant control variables
20.Exclude LTs over 10k GVWR3
21. Small pickups and SUVs only3
22. Large pickups only3
23. Large pickups only, exclude those > 10k GVWR3 (20 & 22)
24. Include AWD, but not muscle or police, cars
25. Include muscle and police, but not AWD, cars
26. Exclude 3 high-risk car models
27. Include AWD cars, exclude 3 high-risk car models (24 & 26)
28. 2-piece variable for CUV weight4

29. 2-piece variable for PC and LT footprint5
30. 2-piece variable for weight and for footprint4-5 (28 & 29)

31. Remove kinks in NHTSA VMT schedules
32. Use Texas rather than Polk odometer ratios
33. Both adjustments to NHTSA VMT (3 1 and 32)
Cars
<3,197
Ibs
.49%
.37%
.36%
.67%
.14%
.45%
.71%
2.39%
2.65%
1.42%
0.53%
2.08%
2.72%
1.42%
1.44%
1.58%
0.93%
0.88%
.49%
.47%
.49%
.49%
.49%
.49%
.29%
.66%
.38%
.15%
.49%

.31%
.31%

.47%
.21%
.19%
>3,197
Ibs
0.50%
0.46%
0.46%
0.58%
-0.85%
2.90%
0.68%
1.37%
2.96%
0.70%
0.10%
1.09%
1.57%
-0.11%
0.62%
-0.42%
0.48%
-0.43%
0.50%
0.54%
0.50%
0.50%
0.50%
0.50%
0.77%
0.40%
0.29%
0.53%
0.50%

0.72%
0.72%

0.49%
0.15%
0.13%
Light trucks1
<4,947
Ibs
-0.10%
-0.13%
-0.13%
-0.02%
-1.66%
-0.56%
0.26%
0.32%
0.30%
-0.39%
-0.10%
0.21%
0.42%
-0.08%
-0.05%
-0.09%
-0.66%
-0.85%
-0.10%
-0.13%
0.06%
-0.01%
-4.27%
-6.49%
-0.10%
-0.10%
-0.10%
-0.10%
-0.10%

-0.75%
-0.75%

-0.10%
-0.25%
-0.26%
>4,947
Ibs
-0.71%
-0.56%
-0.56%
-0.72%
-1.06%
-1.24%
-0.55%
-0.09%
0.00%
-0.99%
-0.52%
-0.83%
-0.55%
-0.62%
-0.94%
-1.80%
-0.97%
-2.13%
-0.71%
-0.70%
-0.80%
-0.24%
0.52%
1.31%
-0.71%
-0.71%
-0.71%
-0.71%
-0.71%

-0.89%
-0.89%

-0.72%
-0.87%
-0.87%

CUV/
minivan
-0.99%
-1.30%
-1.31%
-1.28%
-0.16%
-0.42%
-0.25%
0.00%
-0.43%
-1.65%
-1.13%
-1.01%
-1.00%
-1.43%
-0.99%
-0.61%
-1.15%
-0.66%
-0.27%
-0.84%
-0.99%
-0.99%
-0.99%
-0.99%
-0.99%
-0.99%
-0.99%
-0.99%
-0.31%
-1.21%
-1.07%
-0.20%
-1.21%
-0.99%
-0.99%
-1.00%
Notes:
 Red font indicates estimate is statistically significant at 95% confidence interval.
 Gray shading indicates estimate is not changed from baseline regression model in alternative regression model.
1 Light trucks includes pickups and truck-based SUVs, and excludes car-based CUVs and minivans.
2 In model 6 footprint is allowed to vary with mass.
3 The median mass used for Models 20-23 is: 4,870 pounds for Model 20; 4,704 pounds for Model 21; 6,108 pounds
  for Model 22; and 6,062 pounds for Model 23.
4 The median mass for CUVs/minivans used for Models 28 and 30 is 3,939 pounds.

-------
                                                       Assessment of Vehicle Safety Effects
5 The median footprints used for Models 29 and 30 are 44.3 square feet for cars, 56.9 square feet for light trucks, and
  49.0 square feet for CUVs/minivans.
   Table 8.6 indicates that, for cars < 3,197 pounds, all alternative models estimate that mass
reduction is associated with an increase in societal fatality risk, ranging from a 0.53 percent
increase (Model 10) to a 2.72 percent increase (Model 12). 19 of the 33 alternative models
estimate a smaller increase in risk, and 8 estimate a larger increase in risk, than the NHTSA
baseline model (the remaining 6 alternative models, shaded in grey in Table 8.6, do not make
changes to the regression model for cars).  For cars > 3,197 pounds, all but four of the alternative
models estimate that mass reduction is associated with an increase in societal fatality risk,
ranging from a 0.85 percent decrease (Model 4) to a 2.96 percent increase (Model 8).  13 of the
33 alternative models estimate a smaller increase, or a decrease, in risk, and 14 estimate a larger
increase in risk, than the NHTSA baseline model (six alternative models do not make changes to
the regression model for cars).

   For light trucks < 4,947 pounds, Table 8.6 indicates that only six of the 31 applicable
alternative modelsu estimate that mass reduction is associated with an increase in fatality risk:
ranging from a 1.66 percent decrease in risk (Model 4) to a 0.42 percent increase in risk (Model
12).  12 of the 31 applicable alternative models estimate a larger decrease  in risk, 11 estimate a
smaller decrease, or an increase, in risk, and two estimate the same change in risk, compared to
the NHTSA baseline model (six alternative models do not make changes  to the regression model
for light trucks). In the two models restricted to analyses of large pickups, mass reductions in
large pickups < 6,108 pounds (Model 22) and < 6,062 pounds (Model 23) are associated with
decreases in fatality risk an order of magnitude larger than in the baseline NHTSA model (4.3
percent and 6.5 percent decreases in  risk, respectively).  The classification of relatively light (i.e.,
below the median) trucks in Models  22 and 23 is distinct to the classification of relatively light
trucks in the other models; NHTSA advises caution in the interpretation and comparison of
estimates in Models 22 and 23 with other models.

   For light trucks > 4,947 pounds, none of the 31 applicable alternative modelsv estimate that
mass reduction is associated with an increase in fatality risk, and range from a 2.13 percent
decrease in risk (Model 17) to no change in risk (Model 8). 15 of the 31 applicable alternative
models estimate a larger decrease in  risk, 9 estimate a smaller decrease in risk, and one no
change in risk,  compared to the NHTSA baseline model (six alternative models do not make
changes to the regression model for light trucks).  In the two models restricted  to analyses of
large pickups, mass reductions in large pickups > 6,108 pounds (Model 22) and > 6,062 pounds
(Model 23) are associated with  increases in fatality risk (of 0.52 percent and 1.31 percent,
respectively), compared to the decrease in the baseline model. The classification of relatively
heavy (i.e., above the median) trucks in Models 22 and 23 is distinct to the classification of
u Not including Models 22 and 23, which apply to large pickups only, and use much higher median weights (6,108
  and 6,062 pounds, respectively) to define lighter and heavier large pickups than in the baseline model.
v Not including Models 22 and 23, which apply to large pickups only, and use much higher median weights (6,108
  and 6,062 pounds, respectively) to define lighter and heavier large pickups than in the baseline model.

-------
                                                      Assessment of Vehicle Safety Effects
relatively heavy trucks in the other models; as before, NHTSA advises caution in the
interpretation and comparison of estimates in Models 22 and 23 with other models.

   For CUVs/minivans, all but one of the 31 applicable alternative modelsw estimate that mass
reduction is associated with a decrease in fatality risk, and range from a 1.65 percent decrease in
risk (Model 9) to no change in risk (Model 7). 11 of the 31 applicable alternative models
estimate a larger decrease in risk, and nine estimate a smaller decrease in risk, and two estimate
no change in risk, than the NHTSA baseline model (9 alternative models do not make changes to
the regression model for CUVs/minivans). In the two models that estimate the effect of mass
reduction on risk separately for lighter- and heavier-than-average  CUVs/minivans, mass
reduction in lighter (< 3,939 pounds) CUVs/minivans is associated with smaller decreases in
fatality risk (0.31 percent and 0.20 percent decreases in Models 28 and 30, respectively) than
mass reduction in heavier (> 3,939 pounds) CUVs/minivans (1.21 percent decrease in both
models).

   LBNL noted that if the relationship between mass reduction and societal fatality risk is strong,
one would expect to observe a relatively low sensitivity of estimated effects from NHTSA's
baseline model when substituting alternative induced exposure data, excluding certain cases, and
including supplementary independent variables. However this is not the case; the baseline
results can be sensitive, especially for cars, to changes in the variables and data used. For
instance, accounting for vehicle manufacturer (Model 8), or removing crashes involving alcohol,
drugs,  or bad drivers (Model 12), substantially increases the detrimental effect of mass reduction
in cars on risk.  On the other hand, the DRI measures (using stopped instead of non-culpable
vehicles and replacing footprint with wheelbase and track width, Model 17),  including AWD
cars but excluding three high-risk sporty compact cars (Model 27), and using VMT weights
based on Texas odometer data (Model 33) substantially decreases the detrimental effect of mass
reduction in cars on risk.

   The differences among the point estimates of the alternative regression models in Table 8.6
are within the uncertainty bounds NHTSA estimated using a jack  knife method. However,
because the Volpe model uses the point estimates,  and not the uncertainty bounds, using the
estimates from one of the alternative models could result in large changes in the estimated
change in fatalities from mass reduction.  For example, if NHTSA used the estimated
relationship between mass reduction for lighter cars and societal fatality risk from Model 17
(0.88 percent reduction) rather than the estimate from the baseline model (1.49 percent), the
Volpe  model would enable manufacturers to make much larger reductions in mass without
compromising safety.

   Using two or more variables that are strongly correlated in the  same regression model
(referred to as multicollinearity) can lead to inaccurate results. However, the correlation between
vehicle mass and footprint may not be strong enough to cause serious concern. The Pearson
correlation coefficient r between vehicle mass and footprint ranges from 0.95 for four-door
sedans and SUVs,  to 0.19 for minivans.x  The variance inflation factor (VIF) is a more formal
measure of multicollinearity of variables included in a regression  model. Allison28 "begins to get
w Not including Models 28 and 30, which estimate the effect of mass reduction on risk separately for lighter (<
  3,939 pounds) and heavier (> 3,939 pounds) CUVs/minivans.
x Removing one minivan model, the Kia Sedona, improves the correlation for minivans to 0.50

-------
                                                      Assessment of Vehicle Safety Effects
concerned" with VIF values greater than 2.5, while Menard29 suggests that a VIF greater than 5
is a "cause for concern," while a VIF greater than 10 "almost certainly indicates a serious
collinearity problem;" however, O'Brien30 suggests that "values of VTF of 10, 20, 40 or even
higher do not, by themselves, discount the results of regression analyses."  When both weight
and footprint are included in the regression models,  the highest VIF associated with any variable
exceeds 5 for four-door cars, small pickups, SUVs, and CUVs, exceeds 2.5 for two-door cars and
minivans, and is 1.5 for large pickups. NHTSA included several analyses to address possible
effects of the near-multicollinearity between mass and footprint.

   First, NHTSA ran a sensitivity case where footprint is not held constant, but rather allowed to
vary as mass varies (i.e., NHTSA ran a regression model which includes mass but not footprint);
this is Model 6 in Table 8.6.Y If the multicollinearity was so great that including both variables
in the same model gave misleading results, removing footprint from the model would give much
different results than keeping it in the model. NHTSA's sensitivity test estimates that when
footprint is allowed to vary with mass, the effect of mass reduction on risk increases for all
vehicles types: from a 1.49 percent increase to  a  1.71 percent increase for lighter cars, and from a
0.50 percent increase to a 0.68 percent increase for heavier cars; from a 0.10 percent decrease to
a 0.26 percent increase for lighter light trucks,  and from a 0.71 percent decrease to a 0.55 percent
decrease for heavier light trucks; and from a 0.99 percent decrease to a 0.25 percent decrease for
CUVs and minivans.

   Second, NHTSA conducted a stratification analysis of the effect of mass reduction on risk by
dividing vehicles into deciles based on their footprint, and running a separate regression model
for each vehicle and crash type, for each footprint decile (3 vehicle types times 9 crash types
times 10 deciles equals 270 regressions).2 This analysis estimates the effect of mass reduction
on risk separately for vehicles with similar footprint. The analysis indicates that reducing
vehicle mass does not consistently increase risk across all footprint deciles for any combination
of vehicle type and crash type. Risk increases  with decreasing mass in a majority of footprint
deciles for 12 of the 27 crash and vehicle combinations, but few of these increases are
statistically significant.  On the other hand, risk decreases with decreasing mass in a majority of
footprint deciles for 5 of the 27 crash and vehicle combinations; in some cases these risk
reductions are large and statistically significant.AA If reducing vehicle mass while maintaining
footprint inherently leads to an increase  in risk, the coefficients on mass reduction should be
more consistently positive, and with a larger R2,  across the 27 vehicle/crash combinations, than
shown in the analysis.  These findings are consistent with the conclusion of the basic regression
analyses; namely, that the effect of mass reduction while holding footprint constant, if any, is
small.

   LBNL noted that one limitation of using logistic  regression to estimate the effect of mass
reduction on risk is that a standard statistic to measure the extent to which the variables in the
model explain the  range in risk, equivalent to the R2 statistic in a linear regression model, does
not exist.  (SAS does generate a pseudo-R2 value for logistic regression models; in almost all of
the NHTSA regression models this value is less than 0.10).  For this reason LBNL conducted an
YKahane (2012), pp. 93-94.
z Ibid, pp. 73-78.
AA And in 10 of the 27 crash and vehicle combinations, risk increased in 5 deciles and decreased in 5 deciles with
  decreasing vehicle mass.

-------
                                                      Assessment of Vehicle Safety Effects
analysis of risk versus mass by vehicle model, for 246 models with at least 10 billion VMT, or at
least 100 fatalities (90 car models, 113 light truck models, and 43 CUV/minivan models); these
246 models represent nearly 90 percent of all fatalities, vehicle registration-years, and VMT.
Figure 8.1 shows the relationship between vehicle mass and actual, or unadjusted, societal risk
per VMT, by vehicle type and model; the curb weight for each model is averaged over model
years 2003 to 2010. For most vehicle types, risk decreases as mass increases; however, risk does
not appear to change as small pickup mass increases, and risk actually increases with increasing
mass of large pickups. And the correlation between mass and risk is quite low, ranging from an
R2 of 0.25 for large pickups to essentially zero for SUVs. LBNL then estimated adjusted risk,
after accounting for all of the variables in the baseline regression model except for vehicle
weight and footprint. First LBNL calculated the predicted risk for each induced exposure case,
based on its vehicle attributes, driver characteristics, and crash circumstances.  Then
standardized risks for each vehicle model were estimated for a 50-year old male driving a 4-year
old vehicle in the day, in a non-rural county, in a low-risk state, on a high-speed road.  The
standardized risk was then multiplied by the ratio of actual risk to predicted risk (a measure of
the residual risk  not controlled for by the NHTSA baseline model) to estimate adjusted risk per
VMT for each vehicle model, which controls all vehicle, driver and  crash variables other than
weight or footprint.
       320 -
          2000 2400 2800  3200 3600 4000 4400 4800  5200 5600 6000 6400 6800  7200 7600
                                         Curb weight (Ibs)

 Figure 8.1 Actual (Unadjusted) U.S. Societal Fatality Risk per VMT vs. Curb Weight, By Vehicle Type and
                                          Model

-------
                                                       Assessment of Vehicle Safety Effects
          2000  2400  2800 3200  3600  4000 4400  4800  5200  5600 6000  6400  6800 7200
                                          Curb weight (Ibs)

Figure 8.2 Adjusted U.S. Societal Fatality Risk per VMT vs. Curb Weight, by Vehicle Type and Model, After
           Accounting for All Driver, Crash, and Vehicle Variables except Mass and Footprint

   As shown in Figure 8.2, after accounting for all the control variables except vehicle mass and
footprint, adjusted risk does decrease as mass increases, at least for all vehicle types except
SUVs and large pickups. However, risk and mass are not strongly correlated, with the R2
ranging from 0.40 for two-door cars  and 0.36 for four-door cars, to essentially zero for SUVs and
large pickups. This means that, on average, risk decreases as mass increases, but the variation in
risk among individual vehicle models is stronger than the trend in risk from light to heavy
vehicles.

   Figure 8.2 indicates that some vehicles on the road today have the same, or lower, fatality risk
than models that weigh substantially more, and are substantially larger in terms of footprint.
After accounting for differences in driver age and gender, safety features installed,  and crash
times and locations, there are numerous examples of different models with similar weight and
footprint yet widely varying fatality risk. The variation of fatality risk among individual models
may reflect differences in vehicle design, differences in the drivers who choose such vehicles
(beyond what can be explained by demographic variables such as  age and gender),  and statistical
variation of fatality rates based on limited data for individual models.

   The figure shows that when the data are aggregated at the make-model level, the combination
of differences in vehicle design,  vehicle selection, and statistical variations has more influence
than mass on fatality rates. The figure  perhaps also suggests that,  to the extent these variations in
fatality rates are due to differences in vehicle design rather than vehicle selection or statistical
variations, there is potential for lowering fatality rates through improved vehicle design.  This is
consistent with NHTSA's 2012 opinion that some of the changes in its regression results

-------
                                                      Assessment of Vehicle Safety Effects
between the 2003 study and the 2012 study are due to the redesign or removal of certain smaller
and lighter models of poor design.

   In its 2012 report NHTS A estimated the effect of four scenarios of mass reduction in the
recent vehicle fleet on the overall number of fatalities, using the relationships between mass
reduction and societal fatality risk estimated in the NHTSA baseline model.  LBNL recreated
this methodology using the updated 2016 NHTSA baseline model, for the four scenarios NHTSA
analyzed in 2012 plus two additional  scenarios:

   •  Scenario 1: 100-lb reduction in all vehicles;
   •  Scenario 2: proportionate 2.5 percent mass reduction in all vehicles;
   •  Scenario 3: mass reduction of 5.0 percent in heavier light trucks, 2.5 percent in all other
       vehicle types except cars, whose mass is kept constant;
   •  Scenario 4: a safety-neutral scenario (2012: 0.5 percent mass reduction in lighter cars, 2.1
       percent in heavier cars, 3.1 percent in CUVs/minivans, 2.6 percent in lighter light trucks,
       and 4.6 percent in heavier light trucks; 2016: 2.0 percent mass reduction in cars, 2.5
       percent in lighter light trucks and CUVs/minivans; and 3.0 percent in heavier light
       trucks);
   •  Scenario 5: reduce mass of light trucks to the median mass of cars; and
   •  Scenario 6: mass reduction estimated in 2015 NRC committee report (reduce mass in
       small cars by 5 percent, midsize cars 10 percent, large cars  15 percent, and all light
       trucks, including CUVs/minivans, 20 percent; LBNL translated the mass reductions for
       cars into 5 percent for lighter-than-average cars and 12.5 percent for heavier-than-average
       cars).
   Table 8.7 shows that the relationship between mass reduction and risk estimated in 2012
resulted in an annual 224 increase in fatalities under the mass reduction scenario called for in the
2015 NRC report (Scenario 6).  However, using the updated relationships from the 2016 NHTSA
baseline, this fleet mass reduction scenario is estimated to result in 220 lives saved, and over
1,300 lives saved using the relationships estimated after including the two DRI measures
(stopped vehicle induced exposure and split-footprint model).
  Table 8.7 Estimated Annual Change in Fatalities from Six Different Fleetwide Mass Reduction Scenarios,
     Using Coefficients Estimated By 2012 and 2016 NHTSA Baseline Models and 2016 DRI Measures
Coefficients used
2012 NHTSA baseline
2016 NHTSA baseline
2016 DRI measures
Scenario 1
157
55
-114
Scenario 2
108
22
-152
Scenario 3
-8
-53
-282
Scenario 4
0
0
-174
Scenario 5
-150
-404
-1,901
Scenario 6
224
-220
-1,306
8.2.4.7 Fleet Simulation Model

   NHTSA has traditionally used real world crash data as the basis for projecting the future
safety implications for regulatory changes. However, since lightweight vehicle designs are
introducing fundamental changes to the structure of the vehicle, there is some concern that the
historical safety trends may not apply. To address this concern, NHTSA developed an approach
to utilize the lightweight vehicle  designs to evaluate safety in a subset of real world
representative crashes. The methodology focused on frontal crashes due to the availability of
existing vehicle and occupant restraint models. Representative crashes were simulated between

-------
                                                     Assessment of Vehicle Safety Effects
baseline and lightweight vehicles against a range of vehicles and roadside objects using two
different size belted driver occupants (adult male and small female) only.  No passenger(s) or
unbelted driver occupants were considered in this fleet simulation. The occupant injury risk
from each of the simulations were calculated and summed to obtain combined occupant injury
risk. The combined occupant injury risk was weighted according to the frequency of real world
occurrences to develop overall societal risk for baseline and light-weighted vehicles. Note here,
the generic restraint system developed and used in the baseline occupant simulations were also
used in the light-weighted vehicle occupant simulations as the purpose of this fleet simulation is
to understand changes in societal injury risks due to mass reduction for different class of vehicles
in frontal crashes. No modifications to the restraint systems were done for light-weighted
vehicle occupant simulations. Any modifications to the restraint systems to improve occupant
injury risks or societal injury risks in the light-weighted vehicle, would have conflated the results
without identifying the effects of mass reduction only. The following sections provide an
overview of the fleet simulation study -

   NHTSA contracted with George Washington University to develop a fleet simulation model31
to study the impact and relationship of light-weighted vehicle design with injuries and fatalities.
In this study, there were eight vehicles as follows:

    •   2001 model year Ford Taurus finite element model baseline and two simple design
       variants included a 25 percent lighter vehicle while maintaining the same vehicle front
       end stiffness and 25 percent overall stiffer vehicle while maintaining the same overall
       vehicle mass32.
    •   2011 model year Honda Accord finite element baseline vehicle and its 20 percent light-
       weight vehicle designed by Electricore.  (This mass reduction study was sponsored by
       NHTSA33).
    •   2009/2010 model year Toyota Venza finite element baseline vehicle and two design
       variants included a 20 percent light-weight vehicle model (2010 Venza) (Low option
       mass reduction vehicle funded by EPA and International Council on Clean
       Transportation (ICCT)) and a 35 percent light-weight vehicle (2009 Venza) (High option
       mass reduction vehicle funded by California Air Resources Board34).
   The light weight vehicles were designed to have similar vehicle crash pulses to the baseline
vehicles. Over 440 vehicle crash simulations were conducted for the range of crash speeds and
crash configurations to generate the crash pulse and intrusion data points shown in Figure 8.3.
The crash pulse data and intrusion data points will be used as inputs in the occupant simulation
models.

-------
                                                      Assessment of Vehicle Safety Effects
        Speed
        (mph)
15
20
25
30
35
40
15
20
25
30
35
             Crash
            Partner
                    Fixed Object
                    «*»
                      Vehicle
                                   Crash Type
                                 Configuration
                                     Full Engagement      % Offset
                                               Center
                                                Pole
                                     Fo// Engagement    % Offset
Occupant
  Sizes
Midsize Male
(Target)
Small Female
(Target)
                                                         Midsize Male Midsize Male
                                                           (Partner)     (Target)
                                                         Small Female Small Female
                                                           (Partner)     (Target)
                             Figure 8.3 Vehicle Crash Simulations

   For the vehicle to vehicle impact simulations, four finite element models were chosen to
represent the fleet as shown in Table 8.8. The partner vehicle models were selected to represent
a range of vehicle types and weights.  It was assumed that the vehicle models would reflect the
crash response for all vehicles of the same type, e.g. mid-size car. Only the safety or injury risk
for the driver in both the target vehicle and in partner vehicle was evaluated in this study.
                  Table 8.8 Base Vehicle Models Used in the Fleet Simulation Study
      Vehicle Model (NCAC)

      http://www.ncac.gwu.edu/vml/models.html
                                           FE Weight

                                           No. Parts/Elements
      Taurus

      (MY2000 -2007)
                                                                   1505kg

                                                                    802/
                                                                 973,351
      Yaris

      (MY2005-2013)
                                                                   1100kg
                                                              917/1,514,068
      Explorer

      (MY2002 - 2005)
                                                                   2025 kg
                                                               923/714,205
      Silverado

      (MY2007-2013)
                                                                   2270 kg
                                                               719/963,482

-------
                                                      Assessment of Vehicle Safety Effects
   As noted earlier, the vehicle simulations generated vehicle deformations and acceleration
responses that were utilized to drive occupant restraint simulations and predict the risk of injury
to the head, neck, chest, and lower extremities. In all, over 1,520 occupant restraint simulations
were conducted to evaluate the risk of injury for mid-size male and small female drivers.

   The computed societal injury risk (SIR) for a target vehicle v in frontal crashes is an
aggregate of individual serious crash injury risks weighted by real-world frequency of
occurrence (v) of a frontal crash incident.  A crash incident corresponds to a crash with different
partners (Npartner) at a given impact speed (Pspeed), for a given driver occupant size  (Loccsize),
in the target or partner vehicle (T/P), in a given crash configuration (Mconfig), and in  a single- or
two-vehicle crash (Kevent).  CIR (v) represents the combined injury risk (by body region) in a
single crash incident, (v) designates the weighting factor,  i.e., percent of occurrence, derived
from National Automotive Sampling System Crashworthiness Data System (NASS CDS) for the
crash incident. A driver age group of 16 to 50 years old was chosen to provide a population with
a similar, i.e., more consistent, injury tolerance. Figure 8.4 shows how overall change in the
societal risk is computed.
^Kevent ^Loccsize ^Mconfig ^Npartner
2ifc=l   Ll=l    Lm=l    ZJ,,=Q
                                                         ^Pspeed
                                                                                klmnop
                                                                                      /-. •>
                                                                                      (V)
                                                                      Overall
                                                                   Societal Risk
                                                                     for Target
               CIR of
             partner in
                VTV
               Figure 8.4 Diagram of Computation for Overall Change in Societal Risk

   The fleet simulation was performed using the best available engineering models, with base
vehicle restraint and airbag settings, to estimate societal risks of future lightweight vehicles.  The
range of the predicted risks for the baseline vehicles is from 1.25 percent to 1.56 percent, with an
average of 1.39 percent, for the NASS frontal crashes that were simulated.  The change in driver
injury risk between the baseline and light-weighted vehicles will provide insight into the estimate
of modification needed in the restraint and airbag systems of lightweight vehicles.  If the
difference extends beyond the expected baseline vehicle restraint and airbag capability, then
adjustments to the structural designs would be needed. The results from the fleet simulation
study show the trend of increased societal injury risk for light-weighted vehicle designs, as

-------
                                                      Assessment of Vehicle Safety Effects
compared to their baselines, occurs for both single vehicle and two-vehicle crashes. Results are
listed in Table 8.9.

   In general, the societal injury risk, in the frontal crash simulation, associated with the small
size driver is elevated when compared to that of the mid-size driver. However, both occupant
sizes had reasonable injury risk in the simulated impact configurations that are representative of
the regulatory and consumer information testing. NHTSA examined three methods for combing
injuries to different body regions. One observation was that the baseline mid-size CUV model
was more sensitive to leg injuries.

Table 8.9 Overall Societal Risk Calculation Results for Model Runs, with Base Vehicle Restraint and Airbag
                   Settings Being the same for All Vehicles, in Frontal Crash Only
Target Vehicle
Weight (Ibs)
reduction
% mass reduction
Societal Risk 1
Delta Increase
Societal Risk II
Delta Increase
Societal Risk IIP
Delta Increase
Passenger
Car Baseline
3681


1.56%

1.43%

1.44%

Passenger
CarLW
2964
716
19%
1.73%
0.17%
1.57%
0.14%
1.59%
0.15%
CUV Baseline
3980


1.36%

1.14%

CUV Low
Option
3313
668
17%
1.46%
0.10%
1.20%
0.06%
CUV High
Option
2537
1444
36%
1.57%
0.21%
1.30%
0.16%

Societal Risk 1 - Target + Partner Combined AIS3+ risk of Head, Neck, Chest & Femur
Societal Risk II - Target + Partner Combined AIS3+ risk of Head, Neck, and Chest
Societal Risk IIP - Target + Partner Combined AIS3+ risk of Head, Neck, and Chest with A-Pillar Intrusion Penalty
   This study only looked at lightweight designs for a midsize sedan and a mid-size CUV and
did not examine the safety implications for heavier vehicles.  The study was also limited to only
frontal crash configurations and considered just mid-size CUVs whereas the statistical regression
model considered all CUVs and all crash modes.

   The change in safety risk from the MY2010 fleet simulation study was direct!onally
consistent with the results for passenger cars from NHTSA 2012 regression analysis studyBB,
which covered data for MY2000-MY2007. The NHTSA 2012 regression analysis study was
updated in 2016 to reflect newer MY 2003 to MY 2010. Comparing the fleet simulation overall
societal risk to the to the 2016 update of the NHTSA 2012 regression analysis, the risk
assessment from the fleet simulation is similarly directionally consistent with the passenger car
risk assessment from NHTSA 2016 regression analysis.  As noted above, the fleet simulations
were performed only in frontal crash mode and did not consider other crash modes including
BB The 2012 Kahane study considered only fatalities, whereas, the fleet simulation study considered severe (AIS 3+)
  injuries and fatalities (DOT HS 811 665).

-------
                                                     Assessment of Vehicle Safety Effects
rollover crashes. (The risk assessment for CUV in the regression model combined CUVs and
minivans in all crash modes and included belted and unbelted occupants)

   This fleet simulation study does not provide information that can be used to modify the
coefficients derived by NHTSA 2016 regression analysis study due to the restricted types of
crashes00 and vehicle designs. The fleet simulation modeling study does not affect the agencies'
assessment of the amount of mass reduction that may be implemented with a neutral effect on
safety. As explained earlier, the fleet simulation study assumed restraint equipment to be as in
the baseline model, in which the restraints/airbags are not redesigned to be optimal with light-
weighting.

8.2.5   Based on this Information, What do the Agencies Consider to be the Current State
of Statistical Research on Vehicle Mass and Safety?

   The agencies believe that statistical analysis of historical crash data continues to be an
informative and important tool in assessing the potential safety impacts of the proposed
standards. The newest studies described in this chapter affirm that the effect of mass reduction
while maintaining footprint is a complicated topic, and there are still open questions of whether
future vehicle designs will reduce the historical correlation between weight and size. It is
important to note that while the updated database (with MY2003-MY2010) represents more
current vehicles with technologies more representative of vehicles on the road today, that
database cannot fully represent what vehicles will be on the road in the MYs 2017-2025
timeframe.  As was also true with the 2000-2007 model year data, the vehicles manufactured in
model years 2003-2010 were not subject to footprint-based fuel economy standards. As
explained earlier, the agencies expect that the attribute-based standards will likely facilitate the
design of vehicles such that manufacturers may reduce mass while maintaining footprint.
Therefore, it is possible that the analysis for MYs 2003-2010 vehicles may not be fully
representative of the vehicles that will be on the road in 2017 and beyond.

   We recognize that statistical analysis of historical crash data may not be the only way to think
about the future relationship between vehicle mass and safety.  However, we recognize that other
assessment methods are also subject to uncertainties, which makes statistical analysis of
historical data an important starting point if employed mindfully and recognized for how it can
be useful and what its limitations may be.

   Before the 2017-2025 final rule, NHTSA had funded an independent review of statistical
studies and held a mass-safety workshop in 2011 in order to help the agencies sort through the
ongoing debates over how statistical analysis of the historical relationship between mass and
safety should be interpreted. After the final rule, NHTSA held a follow-up workshop in May
2013. Previously, the agencies had assumed that differences in results were due in part to
inconsistent databases.  By continuing to create updated common databases and making them
publicly available, we are hopeful that this aspect of the problem has been resolved.

   At the 2013 workshop, it was reported by UMTRI that the 2011  independent review of 18
statistical reports suggested that differences in data were probably less significant than the
agencies may have thought. UMTRI stated that statistical analyses of historical  crash data
should be examined more closely for potential multicollinearity issues that exist in some of the
 ; The fleet simulation considered only frontal crashes.

-------
                                                     Assessment of Vehicle Safety Effects
current analyses.  The agencies will continue to monitor issues with multicollinearity in our
analyses, and hope that outside researchers will do the same.

   Finally, based on the findings of the independent review, the agencies continue to be
confident that NHTSA's regression (Kahane)) analytical technique is one of the best for the
purpose of analyzing potential safety effects of future CAFE and GHG standards. UMTRI
concluded that the approach is valid, and NHTSA continued and refined that approach for the
2011 and 2012 analyses; the 2016 NHTSA/Volpe preliminary report continues NHTSA 2012
approach but with newer data, and finds directionally similar (although fewer statistically
significant) relationships between vehicle mass, size, and footprint. Based on these findings, the
agencies continue to believe that in the future, fatalities due to mass reduction will be best
reduced if mass reduction is concentrated in the heaviest vehicles. Analyses should be
continually updated to determine how the effect of mass reduction on safety changes over time.

   Both agencies continue to agree that there are several identifiable safety trends already  in
place or expected to occur in the foreseeable future that may influence the historical relationship
between mass and safety. For example, there are several important new safety standards that
have already been issued and have been phasing in after MY2010 and some potential safety
standards, as shown in Table 8.10. In addition, there are several safety requirements on the
horizon, such as automated  braking, that could further influence the overall historical
relationship between  mass and safety.
               Table 8.10 Additional Safety Requirements Post 2010 (FMVSS, IIHS)35'36
Final Rules
FMVSS No. 126
FMVSS No. 214
FMVSS No. 216
FMVSS No. 226
FMVSS No. Ill
IIHS small overlap
Specifics
(49 CFR § 571.126) requires electronic
stability control in all new vehicles
Side Impact Protection, (49 CFR § 571.214)
new vehicles being equipped with head-
curtain air bags by MY2014.
(49 CFR Parts 571 and 585) Vehicle roof
structure must withstand 3.0 times vehicle
weight - up from 1.5 times, applicable up to
10k Ibs from 6klb vehicles
(49 CFR Parts 571, 585) reduce partial and
complete ejection of vehicle occupants
through side windows in crashes, particularly
rollovers, applies to vehicles 
-------
                                                     Assessment of Vehicle Safety Effects
2025 CAFE and GHG standards for the midterm evaluation of our assumptions that mass
reduction could be used to meet the standards in a cost-effective way without adversely affecting
safely. Section 8.4.1 below discusses the methodology used by the agencies in more detail.

   While the results of the 2016 safety effects analysis are less statistically significant than the
results in the MYs 2017-2025 final rule, the agencies still believe that any statistically significant
results warrants careful consideration of the assumptions about appropriate levels of mass
reduction, and have acted accordingly in conducting this draft technical analysis.

8.3    How do the Agencies Think Technological Solutions Might Affect the
Safety Estimates Indicated by the Statistical Analysis?

   As mass reduction continues to be an important technology option for manufacturers in
meeting future CAFE and GHG standards, manufacturers may invest more and more resources
in developing increasingly lightweight vehicle designs that meet their needs for
manufacturability and the public's need for vehicles that are also safe, useful, affordable, and
enjoyable to drive. There are many different ways to reduce mass, and a considerable amount of
information is available today on lightweight vehicle designs currently in production and that
may be able to be put into production in the MYs 2022-2025 timeframe.  Discussion of
lightweight material designs from NHTSA's workshop is presented below.

   Besides "lightweighting" technologies themselves, though, there are a number of
considerations when attempting to evaluate how future technological developments might affect
the safety estimates indicated by the  historical  statistical analysis. As discussed in the first part
of this section, for example, careful changes in design and/or materials used might mitigate some
of the potential increased risk from mass reduction for vehicle self-protection, through improved
distribution of crash pulse energy, etc.  At the same time, these lightweighting techniques can
sometimes lead to other problems, such as increased crash forces on vehicle occupants that have
to be mitigated, or greater aggressiveness against other vehicles in crashes.  Manufacturers may
develop new and better restraints - air bags, seat belts, etc. - to protect occupants in lighter
vehicles in crashes, but NHTSA's current safety standards for restraint systems are designed
based on the current fleet, not the yet-unknown future fleet. The agency will need to monitor
trends in the crash data to see whether changes to the safety standards (or new safety standards)
become advisable. Manufacturers are also increasingly investigating a variety  of crash
avoidance technologies - forward  collision warning, auto braking, lane departure warning, lane
departure prevention, adaptive headlights, blind spot detection, and vehicle-to-vehicle (V2V)
communications - that, as they become more prevalent in the fleet, are expected to reduce the
number of overall crashes, and thus crash fatalities.  Until these technologies are present in the
fleet in greater numbers,  however, it will be difficult to assess whether they can mitigate the
observed relationship between vehicle mass and safety in the historical data.

-------
                                                     Assessment of Vehicle Safety Effects
8.3.1   Workshops on Technological Opportunities and Constraints to Improving Safety
under Mass Reduction

8.3.1.1 2011 Workshop on Vehicle Mass, Size and Safety

   As stated above in Section 8.2.3, on February 25, 2011, NHTSA hosted a workshop on mass
reduction, vehicle size, and fleet safety at the headquarters of the U.S. Department of
Transportation in Washington, DC. The purpose of the workshop was to provide the agencies
with a broad understanding of current research in the field and provide stakeholders and the
public with an opportunity to weigh in on this issue. The agencies also created a public docket to
receive comments from interested parties that were unable to attend.  The presentations were
divided into two sessions that addressed the two expansive sets of issues.  The first session
explored statistical evidence of the roles of mass and size  on safety, and is summarized in
Section 8.2.3. The second session explored the engineering realities of structural
crashworthiness, occupant injury and advanced vehicle  design, and is summarized here. The
speakers in the second session included Stephen Summers of NHTSA, Gregg Peterson of Lotus
Engineering, Koichi Kamiji of Honda, John German of the International Council on Clean
Transportation (ICCT), Scott Schmidt of the Alliance of Automobile Manufacturers, Guy
Nusholtz of Chrysler, and Frank Field of the Massachusetts Institute of Technology.

   The second session explored what degree of mass reduction and occupant protection are
feasible from technical, economic, and manufacturing perspectives. Field emphasized that
technical feasibility alone does not constitute feasibility in the  context of vehicle mass reduction.
Sufficient material production capacity and viable manufacturing processes are essential to
economic feasibility. Both Kamiji and German noted that both good materials and good designs
will be necessary to reduce fatalities. For example, German cited the examples of hexagonally
structured aluminum columns, such as used in the Honda Insight that can  improve crash
absorption at lower mass, and of high-strength steel components that can both reduce weight and
improve safety. Kamiji made the point that widespread mass reduction will reduce the kinetic
energy of all crashes which should produce some beneficial effect.

   Summers described NHTSA's plans for a model to estimate fleet wide safety effects based on
an array of vehicle-to-vehicle computational crash simulations of current and anticipated vehicle
designs. In particular, three computational models of lightweight vehicles are under
development. They are based on current vehicles that have been modified or redesigned to
substantially reduce mass. The most ambitious was the "high development" derivative of a
Toyota Venza developed by Lotus Engineering and discussed by Mr. Peterson.  The Lotus light-
weighted Venza structure contains about 75 percent aluminum, 12  percent magnesium, 8 percent
steel,  and 5 percent advanced composites.  Peterson expressed confidence that the design had the
potential to meet federal safety standards. Nusholtz emphasized that computational crash
simulations involving more advanced materials were less reliable than those involving traditional
metals such as aluminum and steel.

   Nusholtz presented a revised  data-based fleet safety model in which important vehicle
parameters were modeled based on trends from current NCAP crash tests. For example, crash
pulses and potential intrusion for a particular size vehicle were based on existing distributions.
Average occupant deceleration was used to estimate injury risk.  Through a range of simulations
of modified vehicle fleets, he was able to  estimate the net  effects of various design strategies for
lighter weight vehicles, such as various scaling approaches for vehicle stiffness or intrusion.  The

-------
                                                     Assessment of Vehicle Safety Effects
approaches were selected based on engineering requirements for modified vehicles. Transition
from the current fleet was considered.  He concluded that protocols resulting in safer transitions
(e.g., removing more mass from heavier vehicles with appropriate stiffness scaling according to a
3/2 power law) were not generally consistent with those that provide the greatest reduction in
GHG production: i.e., that the most effective mass reduction in terms of reducing GHG
emissions was not necessarily the safest.

   German discussed several important points on the future of mass reduction.  Similar to
Kahane's discussion of the difficulties of isolating the impact  of mass reduction, German stated
that other important variables, such as vehicle design and compatibility factors, must be held
constant in order for size or weight impacts to be quantified in statistical analyses. He presented
results that the safety impacts of size and weight are small and difficult to quantify when
compared to driver, driving influences, and vehicle design influences. He noted that several
scenarios, such as rollovers, greatly favored the occupants of smaller  and lighter cars once a
crash occurred. He pointed out that if size and design are maintained, lower weight should
translate into a lower total crash force.  He thought that advanced material designs have the
potential to "decouple" the historical correlation between vehicle size and weight, and felt that
effective design and driver attributes may start to dominate size and weight issues in future
vehicle models.

   Other presenters noted industry's perspective of the effect of incentivizing mass reduction.
Field highlighted the complexity of institutional changes that may be  necessitated by mass
reduction, including redesign of material and component supply chains and manufacturing
infrastructure.  Schmidt described an industry  perspective on the complicated decisions that must
be made in the face of regulatory change, such as evaluating goals, gains, and timing.

   Field and Schmidt noted that the introduction of technical innovations is generally an innate
development process involving both tactical and strategic considerations that balance desired
vehicle attributes with economic and technical risk.  In the absence of challenging regulatory
requirements, a substantial technology change is often implemented in stages, starting with lower
volume pilot production before a commitment is made to the infrastructure and supply chain
modifications which are necessary for inclusion on a high-volume production model. Joining,
damage characterization, durability, repair, and significant uncertainty in final component  costs
are also concerns.  Thus, for example, the widespread implementation of high-volume composite
or magnesium structures might be problematic in the short or medium term when compared to
relatively transparent aluminum or high strength steel implementations. Regulatory changes will
affect how these tradeoffs are made and these risks are managed.

   Koichi  Kamiji presented data showing in increased use of high strength steel in their Honda
product line to reduced vehicle mass and increase vehicle safety. He  stated that mass reduction
is clearly a benefit in 42 percent of all fatal crashes because absolute energy is reduced. He
followed up with slides showing the application of certain optimized  designs can  improve  safety
even when controlling for weight and size.  A philosophical theme developed that explored the
ethics of consciously allowing the total societal harm associated with mass reduction to approach
the anticipated benefits of enhanced safety technologies. Although some participants agreed that
there may eventually be specific fatalities that would not have occurred without downsizing,
many also agreed that safety strategies will have to be adapted to the reality created by consumer

-------
                                                    Assessment of Vehicle Safety Effects
choices, and that "We will be ok if we let data on what works - not wishful thinking - guide our
strategies."

8.3.1.2 2013 Workshop on Vehicle Mass, Size and Safety

   As stated above in Section 8.2.4,  on May 13-14, 2013, NHTSA hosted a follow-on
symposium to continue to explore the relevant issues and concerns with mass, size, and safety
tradeoffs.  The first day of the two-day symposium addressed "engineering realities," specifically
the feasible amount of mass reduction and the implications for  structural crashworthiness,
occupant injury, and advanced vehicle design.

   The first-day speakers included Greg Kolwich of FEV, Inc.  (Forschungsgesellschaft fur
Energietechnik und Verbrennungsmotoren (FEV)), Gregg Peterson of Lotus Engineering, Jackie
Rehkopf of Plasan Carbon Composites, Doug Richman of Kaiser, Stephen Ridella of NHTSA,
Scott Schmidt of the Alliance of Automobile Manufacturers, Harry Singh of EDAG Engineering
GmbH. (Engineering and Design Aktiengesellschaft (EDAG)), Chuck Thomas of Honda, and
Blake Zuidema of Arcelor Mittal.

   Peterson discussed continued analysis  of the "high development" and "low development"
options for mass reduction of a Toyota Venza as published in 2012. Lotus Engineering's further
review of the 2010 "high development" study, through CAE and crash analyses, revealed that
some design changes would be required for the aluminum intensive design. The amount of mass
reduction from the body-in-white was likely to decrease but it was felt that much of this could be
offset with mass reduction elsewhere in the vehicle. Joining durability and cycle time were
important  considerations, as was the need to evaluate capital expenditures to implement various
material and structural options.

   Kolwich described an effort to provide detail design,  structural simulation, and cost analysis
to the low development Venza model in an attempt to provide a reasonable mix of
manufacturability, cost, and increased fuel economy. Optimization of material, geometry, and
gauge (thickness) were considered.  FEV  believes a cost-neutral 18 percent mass reduction is
possible but noted that the modeling includes no verification of the redesigned vehicle's dynamic
characteristics.

   Singh described a similar effort to redesign the 2011 Honda Accord.  The economic constraint
was a limit of a 10 percent increase in estimated manufacturing costs. They investigated
combinations of steel, aluminum, magnesium, plastic, and composites applications  and
alternative joining and manufacturing technologies. They employed topology optimization of
the structural elements while maintain interior volume and other functionality.  They required the
revised structure to maintaining an equivalent rating in existing regulatory and consumer testing
programs (e.g., roof crush, side impact, etc.).

   A review of the EDAG design by Honda and presented by Thomas acknowledged that may of
the concepts have tremendous potential and are under consideration, but the estimated 332 kg
(22 percent) in mass reduction might be overly optimistic. He identified some possible
deficiencies against internal testing and performance standards, such as  drivability and noise,
vibration,  and harshness (NVH) that might require remediation of up to 50 kg.  He also noted the
economic  reality that manufacturers must leverage platforms across several vehicle models to
maintain a competitive array of vehicles.  This platform  commonality is inherently non-optimal.

-------
                                                     Assessment of Vehicle Safety Effects
After that adjustment and the associated reinstatement of engine horsepower and other structural
enhancements, the feasible mass reduction might be as little as 175 kg.

   Schmidt discussed two top concerns of automakers for mass reduction approaches.  First,
substantial mass reduction will require comprehensive platform redesign.  This has practical
economic concerns in terms of infrastructure investment and the maintenance of stable
economically viable global supply chains for advanced materials. Second, fleet-wide safety
considerations of mass reduction need to be estimated carefully, especially in light of the
possible effect on baseline mass of any new global safety regulation. He reiterated the theme
that these concerns must be addressed in the context of maintaining current levels of
performance and comfort.

   Zuidema presented the perspectives of the steel industry. Through optimizing grade, gauge,
and geometry, it is believed that advanced high strength steel applications  can provide significant
mass reduction of many components while minimizing required infrastructure changes. There
are numerous new grades being developed that have combinations of ultimate strength and
ultimate elongation that can be used to address the specific requirements of particular
components.  These often result in a minimum cost solution for any strength critical application
and many stiffness-controlled structures. He also noted that life cycle CO2 emissions (i.e.,
accounting for the emissions is material production) and recyclability considerations make steels
even more attractive.

   Richman represented the Aluminum Association and talked about the ability of aluminum to
meet the needs of automotive mass reduction.  He noted the differences in stiffness-controlled
load cases (e.g., vibration and handling) and strength-controlled load cases (e.g., crash). He cited
a German university study (see his slide 14) that implies steel could generate an 11 percent mass
reduction for the vehicle considered while aluminum could generate a 40 percent reduction.
Practical  considerations, such as maintaining a crush zone of approximately 650 mm and
economics as applied by the industry broadly will determine the ultimate multi-material mix in
any vehicle design.

   Rehkopf discussed carbon fiber composites applications in current and  future vehicles.
Composites can be designed to produce  complex geometries with fiber orientations optimized to
give strength and stiffness only where required. The consolidation of numerous parts into one
can reduce both manufacturing time and mass. Analytical capabilities, material costs, and
production improvements (e.g., faster curing resins for reduced cycle time) are continually
bringing down manufactured part costs.  Currently, carbon fiber vehicle components are most
cost competitive when the production rate is under 50,000 per year.

   Ridella presented planned NHTSA research on the introduction of lightweight vehicles into
the vehicle  fleet. NHTSA has developed crash models of several vehicles from recent model
years. The recent mass reduction studies (Venza by FEV, Accord by EDAG, modified Taurus
model). A matrix of computer crash simulations will be performed  across a fleet of various
existing crash models and the new lightweight models. The frontal  crash simulations will be run
at multiple speeds (15 to 40 mph for fixed object crashes, 15 to 35 mph for vehicle-to-vehicle
crashes),  multiple geometries (pole impact, full engagement, offset engagement), and with
multiple occupants (midsize male, small female). Crash pulses extracted from the vehicle models
will be inputs for injury models. Preliminary findings of societal injury risk (defined as
combined likelihood of AIS3 or higher injury by various criteria to target and partner vehicle

-------
                                                     Assessment of Vehicle Safety Effects
occupants) did rise by 5 to 21 percent in the lighter vehicles. The final report was expected out
several months later.

   A panel discussion from the first day panelists focused on the realities of mass reduction as a
moving target both in terms of technology development and in terms of the existing baseline for
each incumbent vehicle design. In any regular redesign cycle, technologies are often frozen two
years before model year release and then remain substantially unchanged for five to seven years.
Thus, as technology advances before the next design cycle, there is likely to be a fair amount of
low hanging fruit. Thomas estimated that 10 percent mass reduction may be a realistic estimate
of the mass reduction broadly feasible by 2025. Peterson concurred, noting that the Lotus studies
were not subject to all the constraints that arise in the full process required to design a vehicle for
high volume manufacturing.  Kolwich felt it may be possible to extract only 4 percent from the
body but as much as  14 percent from the rest of the vehicle. The influence of non-structural
mass (e.g., interior, HVAC) has implications.

   The point was made that footprint-based regulations may have fewer unintended
consequences than mass-based regulation. Ridella cautioned that tradeoffs by all the
stakeholders must be  considered carefully, especially in their impact on overall safety.  The
practical consideration of reliable repair of advanced material components was raised.

8.3.2   Technical Engineering Projects

The agencies conducted several technical/engineering projects described below to estimate the
potential for advanced materials and improved designs to reduce mass in the MY 2017-2025
timeframe, while continuing to meet safety regulations and maintain functionality and
affordability of vehicles.  Another NHTSA-sponsored study will estimate the effects of these
design changes on overall fleet safety.  The detailed discussions about these studies can be found
in the 2012 FRM Joint TSD Section 3.3.5.5. After reviewing comments from Honda regarding
the first of these studies discussed below, NHTSA sponsored a subsequent study to modify the
results of the first study.

8.3.2.1 Honda Accord Study

   NHTSA awarded a contract in December 2010 to Electricore, with EDAG and George
Washington University  (GWU) as subcontractors, to study potential for mass reduction of a mid-
size car - specifically, a Honda Accord — while maintaining the functionality of the baseline
vehicle (the LWV study). The project team was charged to maximize the amount of mass
reduction with the technologies that are considered feasible for 200,000 units per year production
volume during the time frame of this rulemaking while maintaining the retail price in parity
(within ±10 percent variation) with the baseline vehicle. When selecting materials, technologies
and manufacturing processes, the Electricore/EDAG/GWU team utilized, to the extent possible,
only those materials,  technologies and design  which are currently used or planned to be
introduced in the near term (MY2012-2015) on low-volume production vehicles. This approach,
commonly used in the automotive industry, is employed by the team to make sure that the
technologies used in the study will be feasible for mass production for the time frame of this
rulemaking. The Electricore/EDAG/GWU team took a "clean sheet of paper" approach and
adopted collaborative design, engineering and CAE process with built-in feedback loops to
incorporate results and outcomes from each of the design steps into the overall vehicle design
and analysis.  The team tore down and benchmarked 2011 Honda Accord and then undertook a

-------
                                                     Assessment of Vehicle Safety Effects
series of baseline design selections, new material selections, new technology selections and
overall vehicle design optimization.  Vehicle performance, safety simulation and cost analyses
were run in parallel to the design and engineering effort to help ensure that the design decisions
are made in-line with the established project constrains.

   Multiple materials were used for this study. The body structure was redesigned using a
significant amount of high strength steel. The closures and suspension were designed using a
significant amount of aluminum. Magnesium was used for the instrument panel cross-car beam.
A limited amount of composite material was used for the seat structure.

   Safety performance of the light-weighted design was compared to the safety rating of the
baseline MY2011 Honda Accord for seven consumer information and federal safety crash tests
using LS-DYNA.DD These seven tests are the NCAP frontal test, NCAP lateral MDB test,
NCAP lateral pole test, IIHS roof crush, IIHS lateral MDB, IIHS front offset test, and FMVSS
No. 301 rear impact tests.  These crash simulation analyses did not include use of a dummy
model. Therefore only the crash pulse and intrusion were compared with the baseline vehicle
test results. The vehicle achieved equivalent safety performance in all seven self-protection tests
comparing to MY2011 Honda Accord with no damage to the fuel tank.  Vehicle handling is
evaluated using MSC/ADAMSEE modeling on five maneuvers, fish-hook test, double lane
change maneuver, pothole test, and 0.7G constant radius turn test and 0.8G forward braking test.
The results from the fish-hook test show that the light-weighted vehicle  can achieve a five-star
rating for rollover, same as baseline vehicle. The  double  lane change maneuver tests show that
the chosen suspension geometry and vehicle parameter of the light-weighted design are within
acceptable range for safe high speed maneuvers.

   Overall the complete light weight vehicle achieved a total weight savings of 22 percent
(332kg) relative to the baseline vehicle (1480 kg).  The study has been peer reviewed by three
technical experts from the industry, academia and a DOE national lab.  The project team
addressed the peer review comments in the report and also composed a response to peer review
comment document. The final report, CAE model and cost model are published in docket
NHTSA-2010-0131 and can also be found on NHTSA's website.FF The peer review comments
with responses to peer review comments can also be found at the same docket and website.

8.3.2.2 Second Honda Accord Study

   After the LWV design was complete, IIHS added the Small Overlap (SOL) crash test to its
program.  The test replicates what happens when the front corner of a vehicle strikes another
vehicle or an object like a tree or a utility pole. In the test, 25 percent of a vehicle's front end on
the driver side strikes a 5-foot-tall rigid barrier at 40 mph. Small overlap crashes accounted for
nearly 25 percent of the frontal crashes involving serious  or fatal injury to front seat occupants.
In many vehicles the impact at a 25 percent overlap misses the primary structures designed to
manage crash energy in a frontal impact. That increases the risk of severe damage to or collapse
of the occupant compartment structure.  Also, vehicles tend to rotate and slide sideways during
DD LS-DYNA is a software developed by Livermore Software Technologies Corporation used widely by industry
  and researchers to perform highly non-linear transient finite element analysis.
EE MSC/ADAMS: Macneal-Schwendler Corporation/Automatic Dynamic Analysis of Mechanical Systems.
FF Final report, CAE model and cost model for NHTSA's light weighting study can be found at NHTSA's website:
  http://www.nhtsa.gov/fuel-economy.

-------
                                                     Assessment of Vehicle Safety Effects
this type of collision, and that can move the driver's head outboard, away from the protection of
the front airbag.

   Additionally, Honda provided comments to the agency on the findings located here
                                                                              In 2013,
NHTSA awarded a subsequent contract to Electricore to modify the initial LWV design to: 1)
Update the original LWV design to address Honda's comments (LWV1.1); and 2) Update the
LWV design model to correlate to the IIHS Small Overlap (SOL) crash test results (LWV1.2).

   The Electricore team created a detailed finite element model of the MY2011 Baseline Honda
Accord. The team then re-designed the original LWV version  1.0 to version 1.1 to address the
comments from Honda, including improving the vehicle's torsional stiffness and the
performance on IIHS offset barrier, side crash and rear impact.

   In addressing Honda's comments, the weight of the body structure of the LWV 1.1 is
increased by 11.5 kg and the cost is reduced by $13.08 from the original LWV 1.0 design.  In
addition, some of Honda's  recommendations for NVH and drivability were also accepted.  The
total weight and cost of the LWV 1.1 increased by 21.75 kg and $18.13, respectively.

   The LWV1.1 was then upgraded to address the IIHS SOL test (LWV1.2).  To address the
IIHS SOL test (LWV 1.2) the weight of the vehicle is increased by 6.90 kg and the cost by
$26.88.  The new LWV 1.2 design was modeled and assessed for the performance of
crashworthiness in seven crash safety tests such as frontal NCAP test, lateral NCAP moving
deformable test, lateral NCAP pole test, IIHS roof crush test, IIHS lateral moving deformable
test, IIHS moderate frontal offset test and IIHS small overlap front test. The new design
achieved a "good," rating in all crash tests which are comparable to the safety rating of the
MY2013 Accord. When the new design was applied to each of the light vehicle sub-classes,
which span sub-compact cars to large SUV/light trucks, the project mass saving potential
decreased  from a range of 17.7 percent to 19.3 percent (18.2 percent on average) for LWV 1.0 to
a range of 15.8 percent to 17.5 percent (16.3 percent on average) for LWV 1.2.

   In summary, the study demonstrated that the mass of a current production vehicle could be
reduced and yet achieves a "good" rating in all crash tests, including the new IIHS Small
Overlap (SOL) crash test.

8.3.2.3 NHTSA Silverado Study and Light-Duty Fleet Analysis

   In September 2013, NHTSA awarded a contract to automotive design and engineering
company EDAG, Inc., to conduct vehicle weight reduction and cost study of a full size pick-up
truck, specifically, the 2014 Chevrolet Silverado. The goal was to determine the maximum
feasible weight reduction while maintaining the same vehicle functionalities, such as
performance,  safety, and crash rating, as the baseline vehicle. The light weighted version of the
full size pick-up truck (LWT) uses technologies, materials, and manufacturing processes
projected to be available in model year 2025-2030 and capable of high volume production.

   The EDAG team performed a comprehensive teardown/benchmarking of the baseline vehicle
for engineering analysis that included manufacturing technology assessment, material utilization
and complete vehicle geometry scanning. The geometry and material test data from the baseline
vehicle tear down was used to build detailed finite element analysis (FEA) simulation models
suitable crash worthiness using Livermore Software (LS-DYNA) simulation program. Before

-------
                                                     Assessment of Vehicle Safety Effects
the vehicle teardown, torsional stiffness tests, bending stiffness tests, and normal modes of
vibration tests were performed on the baseline vehicle so that these results can be compared with
the light-weighted design. The FEA LS-DYNA models based on the tear-down information and
necessary material properties, such as the stress-strain curve, were based on test results and
information from other available databases or Computer Aided Engineering (CAE) models.  An
FEA LS-DYNA model was created and correlated to the baseline vehicle crash results which
include FMVSS, New Car Assessment Program NCAP and Insurance Institute for Highway
Safety (IIHS) tests. All of the modeled tests were comparable to the actual crash tests performed
on the 2014 Silverado. For load cases that did not have real vehicle test data of which to
correlate to, the results are compared with similar reference vehicles, such as, the 2015 Ford F-
150.

   The project team then used computer modeling and optimization techniques to design the
light-weighted pickup truck and optimized the vehicle structure. The recommended materials,
manufacturing processes, and assembly methods are at present used, some to a lesser degree than
others. These technologies can be fully developed within the normal product design cycle using
the current design and development methods.  The researchers then developed a comprehensive
direct manufacturing incremental  cost estimate for the LWT concept vehicle, including both
detailed direct manufacturing and indirect cost estimates for tooling and equipment investment.

   From the various technologies  that were reviewed for future mass saving potential, four
different vehicle build scenarios were developed. Ranging from a vehicle mass saving of about
11 percent to  23 percent, the light weighting vehicle build  options are as follows:

       1) For an all Advanced High Strength Steel (AHSS) intensive LWT design, including
          cab, pickup box, closures, chassis frame, seat frames and instrument panel beam
          structures.
       2) Design with AHSS chassis frame structure and aluminum cab, pickup box, closures,
          and multi-material seats.
       3) An aluminum intensive solution, using aluminum for body structure, closures, chassis
          frames and magnesium for seats.
       4) An advanced carbon fiber and multi-material Solution, using carbon fiber reinforced
          composite body structure, CFRP/magnesium/aluminum closures, aluminum chassis
          frames and magnesium/composite seat structures.
   From the options above, the design with AHSS chassis  frame structure and aluminum cab,
pickup box and multi-material seats and closures (Option 2), was selected as most likely to be
implemented  for production years 2025 to 2030.  The selected technology options were included
in the detail design and comprehensive Computer-Aided Engineering (CAE) performance
assessment of the complete LWT  design.  The recommended design for LWT achieved a vehicle
mass saving of over 17 percent (428 kg) relative to the baseline weight (2,432 kg).  To maintain
the same vehicle performance as the baseline vehicle, the size of the engine is proportionally
reduced from the baseline 5.3L (355 HP) to 5.0L (335HP)  for the LWT. Without the mass
reduction allowance for the powertrain, the mass saving for the LWT 'glider'  is about 21 percent
(379 kg).

   The report details engineering  analyses and documentation showing how the functionalities
for the light-weighted vehicle are  maintained or improved.  These functionalities include safety,
fuel economy, vehicle utility/performance (e.g. towing, acceleration, etc.), Noise Vibration and

-------
                                                     Assessment of Vehicle Safety Effects
Harshness (NVH), vehicle dynamics (e.g. vehicle weight distribution, rollover stability, etc.),
manufacturability, aesthetics, ergonomics, durability and serviceability.  Appropriate CAE tools
as used by OEMs for this vehicle class were used when comparing baseline vehicle
functionalities to the light-weighted design, such as for safety, NVH, powertrain performance,
towing, durability, etc. Mass reduction technologies assessed for the lightweight truck (LWT)
were applied to other light-duty passenger vehicles and light-duty trucks to estimate the mass
savings while maintaining vehicle size, performance and functionality. This assessment was
conducted for the following light-duty vehicle classes:

       •   Subcompact passenger cars
       •   Compact passenger cars
       •   Midsize passenger cars
       •   Large passenger cars
       •   Minivans
       •   Small CUV/SUV/light-duty trucks
       •   Midsize CUV/SUV/light-duty trucks
       •   Large CUV/SUV/light-duty trucks
   The chosen mass reduction technologies are feasible within the time frame of model years
2017-2025 and would be available across the  passenger car and light-truck vehicle fleet.  In
addition to the introduction of weight saving technologies, consideration was also given to the
capability of suppliers to deliver these mass saving measures in sufficient volumes to support this
initiative.

   All of the weight reduction technologies developed for the LWT program using the 2014
Chevrolet Silverado 1500 as the baseline vehicle can readily be introduced to all of the selected
vehicles within  each of the vehicle subclasses, subcompact to large SUV/light truck, to achieve
weight savings from 15 percent to 18 percent  over next two design cycles for model years 2020
and 2025. Further, there is a significant weight improvement when downsizing the powertrain;
this shows the importance of matching the powertrain to the vehicle weight when undergoing a
weight reduction program as this impacts other sub-systems within the vehicle.

   As demonstrated through detailed design and computer simulation of LWT, these estimated
weight reductions can be achieved. It is important to use the latest weight saving optimization
tools such body structure CAE optimization for material gage-grade-geometry selection. Taking
full advantage of mass compounding and resizing all sub-systems is also critical to achieve the
most mass efficient design.  The pick-up truck lightweighting study and fleet analysis is
currently undergoing peer-review and not publicly available, but is expected to be available in
2016.

8.3.2.4EPA Midsize CUV "Low Development" Study

   EPA, along with ICCT, funded a contract with FEV, with subcontractors EDAG (CAE
modeling) and Munro & Associates, Inc. (component technology research) to study the
feasibility, safety and cost of 20 percent mass reduction on a 2017-2020 production ready mid-
size CUV (crossover utility vehicle) specifically, a Toyota Venza while trying to achieve the
same or lower cost. The EPA report is entitled "Light-Duty Vehicle Mass-Reduction and Cost
Analysis - Midsize Crossover Utility Vehicle."37 This study is a Phase 2 study of the low

-------
                                                     Assessment of Vehicle Safety Effects
development design in the 2010 Lotus Engineering study "An Assessment of Mass Reduction
Opportunities for a 2017-2020 Model Year Vehicle Program"38, herein described as "Phase 1."

   The original 2009/2010 Phase 1 effort by Lotus Engineering was funded by Energy
Foundation and ICCT to generate a technical paper which would identify potential mass
reduction opportunities for a selected vehicle representing the crossover utility segment, a 2009
Toyota Venza. Lotus examined mass reduction for two scenarios - a low development (20
percent mass savings and 2017 production with technology readiness of 2014) and high
development (40 percent mass savings and 2020 production with technology readiness of 2017).
Lotus disassembled a 2009 Toyota Venza and created a bill of materials (BOM) with all
components. Lotus then investigated emerging/current technologies and opportunities for mass
reduction. The report included the BOM for full vehicle, systems, sub-systems and components
as well as recommendations for next steps. The potential mass reduction for the low
development design includes material changes to portions of the body in white (underfloor and
body, roof, body side, etc.),  seats, console, trim, brakes, etc. The Phase 1 project achieved 19
percent (without the powertrain), 246 kg, at 99 percent of original cost at full phase-in after peer
review comments taken into consideration.GG,HH  This was calculated to be -$0.45/kg utilizing
information from Lotus.

   The peer reviewed Lotus Phase 1 study created a good foundation for the next step of
analyses of CAE modeling for safety evaluations and in-depth costing (these steps were not
within the scope of the Phase 1 study) as noted by the peer reviewer recommendations.39

   Similar to Lotus Phase 1  study, the EPA Phase 2 study "low development" begins with
vehicle tear down and BOM development. FEV and its subcontractors tore down a MY2010
Toyota Venza in order to create a BOM as well as understand the production methods for each
component.  Approximately 140 coupons from the BIW were analyzed in order to understand the
full material composition of the baseline vehicle.  A baseline CAE model was created based on
the findings of the vehicle teardown and analysis.  The model's results for static bending, static
torsion, and modal  frequency simulations (for evaluating NVH) were obtained and compared to
actual results from  a Toyota Venza vehicle. After confirming that the results were within
acceptable limits, this model was then modified to create light-weighted vehicle models.  EDAG
reviewed the Lotus Phase 1 low development BIW ideas and found redesign was needed to
achieve the full set of acceptable NVH characteristics. EDAG utilized a commercially available
computerized optimization tool called FEEDS MDO to build the optimization model. The
model consisted of 484 design variables, 7 load cases (2 NVH + 5 crash), and 1 cost evaluation.
The outcome of EDAG's lightweight design optimization included the optimized vehicle
assembly and incorporated the following while maintaining the original BIW design: optimized
gauge and material grades for body structure parts, laser welded assembly at shock towers,
rocker, roof rail, and rear structure subassemblies, aluminum material for front bumper, hood,
and tailgate parts, TRBs on B-pillar, A-pillar, roof rail, and seat cross member parts, design
change on front rail side members. EDAG achieved 13 percent mass reduction in the BIW
including closure.   If aluminum doors were included then an additional decrease of 28kg could
be achieved for a total of 18 percent mass reduction from the body structure.  All other systems
GG The original powertrain was changed to a hybrid configuration.
HH Cost estimates were given in percentages - no actual cost analysis was presented for it was outside the scope of
  the study, though costs were estimated by the agency based on the report.

-------
                                                    Assessment of Vehicle Safety Effects
within the vehicle were examined for mass reduction, including the powertrain (engine,
transmission, fuel tank, exhaust, etc.). FEV and Munro incorporated the Lotus Phase 1 low
development concepts into their own idea matrix. Each component and sub-system chosen for
mass reduction was scaled to the dimensions of the baseline vehicle, trying to maximize the
amount of mass reduction with cost effective technologies and techniques that are considered
feasible and manufactureable in high volumes in MY2017. FEV included a full discussion of the
chosen mass reduction options for each component and subsystem.

   Safety performance of the baseline and light-weighted designs (Lotus Phase 1 low
development and the final EPA Phase 2 design) were evaluated by EDAG through their
constructed detailed CAD/CAE vehicle models.  Five federal safety crash tests were performed,
including FMVSS flat frontal crash, side impact, rear impact and roof crush (using IIHS
resistance requirements) as well as  Euro NCAP/IIHS offset frontal crash. Criteria including the
crash pulse, intrusion and visual crash information were evaluated to compare the results of the
light weighted models to the results of the baseline model. The light weighted vehicle achieved
equivalent safety performance in all tests to the baseline model with no damage to the fuel tank.
In addition, CAE was used to evaluate the BIW vibration modes in torsion, lateral bending, rear
end match boxing, and rear end vertical bending, and also to evaluate the BIW stiffness in
bending and torsion.

   The Phase 2 study 2010 Toyota  Venza lightweight vehicle achieved, with powertrain, a total
weight savings of 18 percent (312 kg) relative to the baseline vehicle (1710 kg) at -$0.43/kg, and
the cost figure is near zero at 20 percent.  The study report and models have been peer reviewed
by four technical experts from a material association, academia, DOE, and a National
Laboratory.  The peer review comments for this study were generally complimentary, and
concurred with the ideas and methodology of the study. A few of the comments required further
investigation, which were completed for the final report. The project team addressed the peer
review comments in the report and  also composed a response to peer review comment document.
Changes to the BIW CAE models resulted in minimal differences. The final report is published
in EPA's docket EPA-HQ-OAR-2010-0799 and the CAE LS DYNA model files and overview
cost model files are found on EPA's website
http://www.epa.gov/otaq/climate/publications.htm#vehicletechnologies.  The peer review
comments with responses to peer review comments can also be found at the same docket and
website.

8.3.2.5 CARB Phase 2 Midsize CUV "High Development" Study

   The California Air Resources Board (CARB) funded a study with Lotus Engineering to
further develop the high development design from Lotus' 2010 Toyota Venza work ("Phase 1").
The CARB-sponsored Lotus "Phase 2" study provides the updated design, crash simulation
results, detailed costing, and analysis of the manufacturing feasibility of the BIW and closures.
Based on the safety validation work, Lotus strengthened the design with a more aluminum-
intensive BIW (with less magnesium). In addition to the increased use of advanced materials,
the new design by Lotus included a number of instances in which multiple parts were integrated,
resulting in a reduction in the number of manufactured parts in the lightweight BIW. The Phase
2 study reports that the number of parts in the BIW was reduced from 419 to 169. The BIW was
analyzed for torsional stiffness and crash test safety with Computer-Aided Engineering
(CAE). The new design's torsional stiffness was 32.9 kNm/deg, which is higher than the

-------
                                                    Assessment of Vehicle Safety Effects
baseline vehicle and comparable to more performance-oriented models.  The research supported
the conclusion that the lightweight vehicle design could pass standard FMVSS 208 frontal
impact, FMVSS No. 210 seatbelt anchorages, FMVSS child restraint anchorage, FMVSS No.
214 side impact and side pole, FMVSS 216 roof crush (with Sxcurb weight), FMVSS 301 rear
impact, IMS low speed front, and IIHS low speed rear. Crash tests simulated in CAE showed
results that were listed as acceptable for all crash tests analyzed. No comparisons or conclusions
were made if the vehicle performed better or worse than the baseline Venza. For FMVSS 208
frontal impact, Lotus based its CAE crash test analyses on vehicle crash acceleration data rather
than occupant injury as is done in the actual vehicle crash. The report from the study stated that
accelerations were within acceptable levels compared to current production vehicle acceleration
results and it should be possible to tune the occupant restraint system to  handle the specific
acceleration pulses of the Phase 2 high development vehicle. FMVSS No. 210 seatbelt
anchorages are concerned with seatbelt retention and certain dimensional constraints for the
relationship between the seatbelts and the seats.  Overall both the front and rear seatbelt
anchorages met the requirements specified in the standard. FMVSS No. 214 side impact show
the energy is effectively managed.  Since dummy injury criteria was not used in the CAE
modeling, a maximum intrusion tolerance level of 300mm was instituted which is the typical
distance between the door panel and most outboard seating positions.  For example, the Phase 2
design was measured at 1 15mm for the crabbed barrier test. The  side pole test resulted in
120mm intrusion for the 5th percentile female and intrusion was measured at 190mm for the
50th percentile male. The report stated FMVSS 216 roof crush simulation shows the Phase 2
high development vehicle will meet roof crush performance requirements under the specified
load case of 3 times the vehicle weight. For the FMVSS rear impact, results show plastic strain
in the fuel tank/system components to be less than 3.5 percent, which is  less than the 10 percent
strain allowed in the test. The pressure change in the fuel tank is  less than 2 percent  so risk of
tank splitting is minimal. The IIHS low speed front and rear show no body structural issues,
however styling adjustments should be made to improve the rear bumper low speed performance.

   The Lotus design achieved a 37 percent (141 kg) mass reduction in the body structure, a 38
percent (484kg) mass reduction in the vehicle excluding the powertrain, and a 32 percent (537
kg) mass reduction in the entire vehicle including the powertrain. The report was peer reviewed
by a cross section of experts and the comments were addressed by Lotus in the peer review
documents. The comments requiring modification were incorporated into the final document.
The documents can be found on EPA's website
8.3.2. 6 EPA Light Duty Truck Study

   The U.S. EPA contracted with FEV North America to perform this study utilizing the
methodology developed in the Midsize CUV light -weighting effort (2012) and the study was
completed in 2015. The results of this work went through a detailed and independent peer
reviewed as well as through the SAE paper publication process.  Feedback was received by
OEM's and others independent of the official peer review process.

   For this study a 201 1 Silverado 1500 was purchased and torn down. The components were
placed into 19 different systems. The components were evaluated for mass reduction potential
given research into alternative materials and designs. The alternatives were evaluated for the
best cost and mass reduction and then compared to each other. CAE analyses for NVH and

-------
                                                    Assessment of Vehicle Safety Effects
safety was completed for the baseline and the light-weighted aluminum intensive vehicle.  A
high strength steel structure with aluminum closures was the first choice of a solution for this
project; however, this was not fully completed for the decision was made by the project team to
change course and pursue the aluminum structure solution due to the expected introduction of the
aluminum intensive F150 into the marketplace. Durability analyses on both the baseline and
light-weighted vehicle designs were performed through data gathered by instrumenting a
Silverado 1500 light duty pickup truck and operating it over various road conditions. Included in
the durability analyses are durability evaluations on the light weighted vehicle frame, door and
other components in CAE space.  The crash and durability CAE analyses allowed for gauge and
grade determinations for specific vehicle components. Load path redesign of the light duty truck
structure (cabin and box structure and vehicle frame) was not a part of this project.

   Most mass reduction was achieved in the cabin and box structure and the closures, which
were converted from steel to aluminum. The suspension system is the second highest system for
mass reduction and includes composite fiber leaf springs. A 50kg and $150 allowance was
considered to mitigate NVH.  Secondary mass savings achieved were based on the amount of
total primary mass reduction achieved.  In this study the engine was able to be downsized 7
percent due to the mass reduction in the vehicle design and still maintain the current towing and
hauling capacities.  The other systems that were reduced in size, while considering truck
performance characteristics, included the transmission, bumpers, suspension, brake, frame and
mounting systems, exhaust, and fuel systems.

8.4    How have the Agencies Estimated Safety Effects for the Draft TAR?

8.4.1   What was the Agencies' Methodology for Estimating Safety Effects?

   As explained above, the agencies consider the latest 2016 preliminary  statistical analysis of
historical crash data by NHTSA/Volpe to  represent the current best estimates of the potential
relationship between mass reduction and fatality increases in the future fleet. This section
discusses how the agencies used the NHTSA/Volpe's 2016 preliminary analysis to calculate
specific estimates of safety effects in the Draft TAR, based on the analysis of how much mass
reduction manufacturers might use to meet the CAFE and GHG standards.

   The CAFE/GHG standards do not mandate mass reduction, nor require that mass reduction
occur in any specific manner. However, mass reduction is one of the technology applications
available to the manufacturers and a degree of mass reduction is used by both agencies' models
to determine the capabilities of manufacturers and to predict both cost and fuel
consumption/emissions impacts of more stringent CAFE/GHG standards. To estimate the
amount of mass reduction to apply in the rulemaking analysis, the agencies considered fleet
safety effects for mass reduction.  As shown in Table 8.3 and Table 8.4, both the Kahane 2012
final report and the NHTSA/Volpe 2016 preliminary  report  show that applying mass reduction to
CUVs, minivans, and light duty trucks will generally decrease societal fatalities, while applying
mass reduction to passenger cars will increase fatalities.  The CAFE model uses coefficients
from the 2016 preliminary report along with the mass reduction level applied to each vehicle
model to project societal fatality effects in each model year. NHTSA used the CAFE model and
conducted iterative modeling runs varying the maximum amount of mass reduction applied to
each subclass in order to identify a combination that achieved a high level of overall fleet mass
reduction while not adversely affecting overall fleet safety.  These maximum levels of mass

-------
                                                      Assessment of Vehicle Safety Effects
reduction for each subclass were then used in the CAFE model for the Draft TAR analysis. The
agencies believe that mass reduction of up to 20 percent is feasible on light trucks, CUVs and
minivans. Thus, the amount of mass reduction selected is based on our assumptions about how
much is technologically feasible without compromising safety. While we are confident that
manufacturers will build safe vehicles and meet (or surpass) all applicable federal safety
standards, we cannot predict with certainty that they will choose to reduce mass in exactly the
ways that the agencies have analyzed in response to the standards. In the event that
manufacturers ultimately choose to reduce mass and/or footprint in ways not analyzed or
anticipated by  the agencies, the safety effects of the rulemaking may likely differ from the
agencies' estimates.

   In the 2012 final rule analysis, NHTSA utilized the 2012 Kahane study relationships between
weight and safety, expressed as percent changes in fatalities per 100-pound mass reduction while
holding footprint constant. However, several identifiable safety trends already were occurring,
or expected to  occur at the time of 2012 FRM, which were not accounted for in the study.  For
example, the two important new safety standards that were  discussed above for electronic
stability control and side curtain airbags, have already been issued and began phasing in after
MY2008. Also in 2012, the shifts in market shares in 2012 from pickups and SUVs to cars and
CUVs were growing due to high gasoline prices, but if the gasoline prices fell, then the demand
for SUVs, CUVs or LDT could rise and consequent growth in vehicle miles travelled if the
economy does not stagnate. And improvements in driver (and passenger) behavior, such as
higher safety belt use rates, may continue.  All of these will tend to reduce the absolute number
of fatalities in the future.  The agencies estimated the overall change in fatalities by calendar year
after adjusting for ESC, Side Impact Protection, and other Federal safety standards and
behavioral changes projected through this time period.

   To estimate the amount of mass reduction to apply in the analysis, the agencies considered
fleet safety effects for mass reduction. As previously discussed the agencies believe that mass
reduction of up to 20 percent is feasible on light trucks, CUVs and minivans, n but that less mass
reduction should be implemented on other vehicle types to avoid increases in societal fatalities.
To find  a safety-neutral compliance path for use in the  agencies' Draft TAR analysis, NHTSA
uses the fatality coefficients derived in the NHTSA/Volpe 2016 preliminary report with mass
reduction levels presented in Table 8-11.  Maximum mass reduction level are 7.5 and 10 percent
for small and medium cars, respectively. Light trucks, CUVs, and minivans achieve mass
reduction levels up to 20 percent.
      Table 8.11 Mass Reduction Levels to Achieve Safety Neutral Results in the Draft TAR Analysis
Mass
Reduction
Level
MR1
MR2
Passenger Car
SmallCar
5%
7.5%
MedCar
5%
7.5%
SmallSUV
5%
7.5%
Light Truck
SmallSUV
5%
7.5%
MedSUV
5%
7.5%
Pickup
5%
7.5%
CUV/Minivan
SmallSUV
5%
7.5%
MedSUV
5%
7.5%
11 When applying mass reduction, NHTSA capped the maximum amount of mass reduction to 20 percent for any
  individual vehicle class. The 20 percent cap is the maximum amount of mass reduction the agencies believe to be
  feasible inMYs 2017-2025 time frame.

-------
                                                       Assessment of Vehicle Safety Effects
MRS
MR4
MRS
.
-
-
10%
-
-
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
Notes:
*MR1-MR5: different levels of mass reduction used in CAFE model

   For the CAFE model, these percentages apply to a vehicle's total weight, including the
powertrain. Table 8.12 shows the amount of mass reduction in pounds for these percentage mass
reduction levels for average vehicle weight in each subclass.

 Table 8.12 Examples of Mass Reduction (in Pounds) for Different Vehicle Subclasses Using the Percentage
                      Information as Defined for the CAFE Draft TAR Analysis
Mass Reduction
(Ibs)
Average Vehicle
Weight (sales-
weighted)
MR1: 5%
MR2: 7.5%
MRS: 10%
MR4: 15%
MRS: 20%
Passenger Car
Small
Car
2,908
145
218
-
-
-
Med
Car
3,576
179
268
358
-
-
Small
SUV
3,490
175
262
349
524
698
Light Truck
Small
SUV
3,693
185
277
369
554
739
Med
SUV
4,633
232
347
463
695
927
Pickup
5,053
253
379
505
758
1,011
CUV/Minivan
Small
SUV
3,621
181
272
362
543
724
Med
SUV
4,348
217
326
435
652
870
   These maximum amounts of mass reduction discussed above were applied in the technology
input files for the CAFE model. NHTSA divides vehicles into classes for purposes of applying
technology in the CAFE model in a way that differs from the Kahane study which divides
vehicles into classes for purposes of determining safety coefficients.  These differences require
that the "safety class" coefficients be applied to the appropriate vehicles in the CAFE
"technology subclasses."  For the reader's reference, for purposes of this Draft TAR, the safety
classes and the technology subclasses relate" as shown in 3.
11 This is not to say that all vehicles within a technology subclass will necessarily fall within a single safety class - as
  the chart shows, some technology subclasses are divided among safety classes.

-------
                                                       Assessment of Vehicle Safety Effects
        Table 8.13 Mapping between Safety Classes and Technology Classes in the CAFE Analysis
Safety Class
PC (Passenger Car)
LT (Light Truck)
CM (CUV and Minivan)
Technology Class
Small Car
Medium Car
Small SUV
Small SUV
Medium SUV
Pickup
Small SUV
Medium SUV
                   Note:*CM = CUV and MiniVan
   Table 8.144 shows CAFE model results for societal safety for each model year based on the
application of the above mass reduction limits.KK These are the estimated increases or decreases
in fatalities over the lifetime of the model year fleet. A positive number means that fatalities are
projected to increase, a negative number (indicated by parentheses) means that fatalities are
projected to decrease. The results are significantly affected by the mass reduction limitations
used in the CAFE model, which allow more mass reduction in light trucks, CUVs,  and minivans
than in other vehicles.  As the negative coefficients  only appear for light trucks, CUVs, and
minivans, a statistically significant improvement in  safety can only occur if more weight is taken
out of these vehicles than out of passenger cars.  Combining passenger car and light truck safety
estimates for the Draft TAR analysis results in a decrease in fatalities over the lifetime of the
nine model years of MY2017-2025 of 24 fewer fatalities with the 2015 baseline. Broken up into
passenger car and light truck categories, there is an increase of 464 fatalities in passenger cars
and a decrease of 488 fatalities in light trucks with the 2015 baseline.
  Table 8.14  NHTSA Calculated Mass-Safety-Related Fatality Impacts of the Draft TAR Analysis over the
               Lifetime of the Vehicles Produced in each Model Year Using 2015 Baseline
Regulatory Class
Passenger Cars
Light Trucks
Total
MY 2017
1
0
1
MY 2018
9
1
12
MY 2019
11
(46)
(35)
MY 2020
21
(48)
(28)
MY 2021
58
(44)
13
MY 2022
70
(52)
17
MY 2023
84
(108)
(24)
MY 2024
98
(104)
(6)
MY 2025
114
(125)
(11)
Total
465
(525)
(61)
   Using the same coefficients from the 2016 NHTSA/Volpe study, EPA used the OMEGA core
model to estimate the impact of weight reduction on net fatalities per mile driven by the fleet.
This is done using the weight reductions applied by OMEGA and applying to those weight
reductions the safety metrics shown in Table 8.15. The "Change per 100 Ibs" column, presented
earlier in Chapter 8 (Table 8.4) shows the change in the number of fatalities as a percentage for
KK NHTSA has changed the definitions of a passenger car and light truck for fuel economy purposes between the
  time of the Kahane 2003 analysis and the NPRM (as well as the final rule). About 1.4 million 2 wheel drive
  SUVs have been redefined as passenger cars instead of light trucks.  The Kahane 2011 and 2012 analyses and the
  2016 NHTSA/Volpe study continue to use the definitions used in the Kahane 2003 analysis.

-------
                                                     Assessment of Vehicle Safety Effects
each 100 pounds of weight removed from vehicles described by the "Safety Class Description"
column.  The "FMVSS Adjustment" factor is also applied to calculate the impact on fatalities per
billion miles of vehicle travel. All of the inputs presented in Table 8.15 are consistent with inputs
used in the CAFE modeling supporting NHTSA's analysis.40

                      Table 8.15 Metrics Used in the OMEGA Safety Analysis
Safety Class Description
PC below 3197
PC above 3197
LT below 4947
LT above 4947
CUE Minivan
Change per 100 Ibs
1.49%
0.51%
-0.10%
-0.72%
-0.99%
Base per billion miles
13.59
11.15
14.35
16.06
9.00
FMVSS Adjustment
0.904
0.904
0.904
0.904
0.904
   Using these metrics, EPA calculated the impact of mass reduction on net vehicle-related
fatalities, as shown in Table 8.16, which shows the results of EPA's safety analysis over the
lifetimes of MY2021 to 2025 vehicles (EPA explains in Chapter 12 why MY2021 vehicles are
included even though this Draft TAR is considering the MY2022 to 2025  standards). A positive
number would mean that fatalities are projected to increase; a negative number means that
fatalities are projected to decrease.  As shown, the EPA analysis projects considerable fatality
decreases in the reference and control cases. Those decreases should be seen as being relative to
the current fleet moving forward in time without mass reductions in response to new standards
(i.e., relative to the projected MY2021 through 2025 baseline fleet). The reference case standards
reduce fatalities relative to the projected baseline fleet (a fleet that continues to meet the 2014
standards in place for the year upon which our baseline fleet is generated) due to mass reduction
done to move the fleet from the 2014 standards to the 2021 standards (the reference case
standards). In the reference case, those 2021 standards continue indefinitely for subsequent
model year vehicles. The control case (i.e., the 2022 through 2025 standards) then result in
further mass reduction beyond the reference case level. This further mass reduction further
reduces fatalities relative to both the baseline and reference cases.  On net, the EPA analysis
shows small net fatality decreases over the lifetimes of MY2021 through 2025 vehicles.
           Table 8.16 EPA's Net Fatality Impacts over the Lifetimes of MY2021-2025 Vehicles
Case
AEO 2015 reference fuel price case using ICMs
AEO 2015 high fuel price case using ICMs
AEO 2015 low fuel price case using ICMs
AEO 2015 reference fuel price case using RPEs
Fatality Impacts
Reference Case
-800
-448
-994
-923
Fatality Impacts
Control Case
-874
-484
-1063
-929
Net Fatality Impacts
-74
-36
-69
-6
8.4.2   Why Might the Real-World Safety Effects be Less Than or Greater Than What the
Agencies Have Calculated?

   As discussed above, the ways in which future technological advances could potentially
mitigate the safety effects estimated for this Draft TAR include the following: lightweight
vehicles could be designed to be both stronger in materials without becoming more intrusive in
crash force; restraint systems could be improved to deal with higher crash pulses in lighter
vehicles; crash avoidance technologies could reduce the number of overall crashes; roofs could

-------
                                                     Assessment of Vehicle Safety Effects
be strengthened to improve safety in rollovers. As also stated above, however, while we are
confident that manufacturers will strive to build safe vehicles, it will be difficult for both the
agencies and the industry to know with certainty ahead of time how crash trends will change in
the future fleet as light-weighted vehicles become more prevalent.  Going forward, we will
continue to monitor the crash data as well as changes in vehicle mass and conduct analyses to
understand the interaction of vehicle mass and size on  safety.

   Additionally, we note that the total amount of mass reduction used in the agencies' analysis
was chosen based on our assumptions about how much is technologically feasible without
compromising safety.  Again, while we are  confident that manufacturers are motivated to build
safe vehicles, we cannot predict with certainty that they will choose to reduce mass in exactly the
ways or amounts that the agencies have analyzed in response to the standards.  In the event that
manufacturers ultimately choose to reduce mass and/or footprint in ways not analyzed by the
agencies, the safety effects may likely differ from the agencies' estimates.

   The agencies note that the standard is flat for vehicles smaller than 41 square feet and that
downsizing in this category could help achieve overall compliance, if the vehicles are desirable
to consumers.  The agencies note that 4.4 percent of MY2015 passenger cars were below 41
square feet, and due to the overall lower level of utility of these vehicles, and the engineering
challenges involved in ensuring that these vehicles meet all applicable federal motor vehicle
safety standards (FMVSS), we  do not expect a significant increase in the use of mass reduction
in this segment of the market.

   The agencies acknowledge that the final  rule did not prohibit manufacturers from redesigning
vehicles to  change wheelbase and/or track width (footprint).  However, as NHTSA explained in
promulgating the MY2008-2011 light truck CAFE standards and the MY2011 passenger car and
light truck CAFE standards, and as the agencies jointly explained in promulgating the MYs
2012-2016  CAFE and GHG standards, we believe that such engineering changes are significant
enough to be unattractive as a measure to undertake solely to reduce compliance burdens.

   Similarly, the agencies acknowledge that a manufacturer could, without actually
reengineering specific vehicles to increase footprint, shift production toward those that perform
well with respect to their footprint-based targets. However, NHTSA and EPA have previously
explained, because such production shifts could run counter to market demands, they could also
be competitively unattractive.

8.4.3   What Are the Agencies' Plans Going Forward?

   The agencies continue to closely monitor the visible effects of CAFE/GHG standards on
vehicle safety as these standards are implemented, and will conduct a full analysis of safety
impacts as part of further steps  in EPA's midterm evaluation and NHTSA's future rulemaking to
establish final MYs 2022-2025 standards.

   NHTSA will closely monitor the safety data, the trends in vehicle weight and size,  the trends
in vehicle mass reduction, as well as the trend for the active and passive vehicle safety during the
period between the release of this Draft TAR and the future rulemaking to establish final CAFE
standards for MYs  2022-2025.  Consistent with confidentiality and other requirements, NHTSA
intends to make these data publicly available when they are compiled. NHTSA will also make
appropriate updates to the statistical study of historical data on the  effects on mass and size

-------
                                                    Assessment of Vehicle Safety Effects
societal safety on an ongoing basis.  At the same time, working closely with EPA and DOE,
NHTSA will continue to assess its analytical methods for assessing the effects of vehicle mass
and size on societal safety and make appropriate updates, including a final version of the 2016
NHTSA/Volpe preliminary report.

-------
                                                             Assessment of Vehicle Safety Effects
References
1 National Research Council, "Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards,"
National Academy Press, Washington, DC (2002), Finding 2, p. 3, Available at
http://www.nap.edu/openbook.php?isbn=0309076013 (last accessed Aug. 2, 2012).
2 Kahane, CJ. (2012, August). Relationships Between Fatality Risk, Mass, and Footprint in Model year 2000-2007
Passenger Cars and LTVs - Final Report. (Report No. DOT HS 811 665). Washington, DC: National Highway
Traffic Safety Administration.
3 DOT HS 812 232, DOT HS 812 248, and paper on  "Effect of Frontal Crash Pulse Variations on Occupant
Injuries" by Steve Mark of Honda R&D  Americas, Paper No. 400 in ESV Conference 18.
4 Khazzoom, J. Daniel. 1994. "Fuel Efficiency and Automobile Safety: Single-Vehicle Highway Fatalities for
Passenger Cars." The Energy Journal, Vol. 15, No. 4  (1994), pp. 49-101.
5Noland, Robert B. 2004. "Motor Vehicle Fuel Efficiency and Traffic Fatalities." The Energy Journal, Vol. 25, No.
4 (2004), pp. 1-22.
6 Ahmad. S. andD.L. Greene. 2005. "The Effect of Fuel Economy on Automobile Safety: A Reexamination."
Transportation Research Record No. 1941, pp. 1-7, Washington, DC, January.
7Evans, Leonard. 2001. "Causal Influence of Car Mass and Size on Driver Fatality Risk." American Journal of
Public Health 91: 1076-81.
8 Anderson M, Auffhammer M. 2014. Pounds that kill: The external costs of vehicle weight. Review of Economic
Studies 82:535-571.
9 White, Michelle J. 2004. "The 'Arms Race' on American Roads: The Effect of Sport Utility Vehicles and Pickup
Trucks on Traffic Safety." The Journal of Law and Economics 47: 333-355.
10 Gayer, Ted. 2004. "The Fatality Risks of Sport-Utility Vehicles, Vans, and Pickups Relative to Cars." The Journal
of Risk and Uncertainty 28 (2): 103-133.
1J Anderson M. 2008. Safety for whom? The effect of light trucks on traffic fatalities. Journal of Health Economics
27: 973-989.
12 Li, Shanjun. 2012. "Traffic Safety and Vehicle Choice: Quantifying the Effects of the 'Arms Race' on American
Roads." Journal of Applied Econometrics 27: 34-62.
13 JacobsenMR. 2013b. Fuel economy and safety: The influences of vehicle class and driver behavior. American
Economic Journal: Applied Economics 5.
1475 Fed. Reg. 25324 (May 7, 2010); the discussion of planned statistical analyses is on pp. 25395-25396.
15 Brewer, John. An Assessment of the Implications of "Smart Design" on Motor Vehicle Safety. 2011. Docket No.
NHTSA-2010-0131.
16National Academy of Sciences. Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles. (613 pp, 15MB, 2015). Committee on the Assessment of Technologies for Improving Fuel Economy
of Light-Duty Vehicles, Phase 2; Board on Energy and Environmental Systems; Division on Engineering and
Physical Sciences; National Research Council. The National Academies Press, Washington, D.C.
17Kahane, C. J. (2003). Vehicle Weight, Fatality Risk and Crash Compatibility of Model Year 1991-99 Passenger
Cars and Light Trucks, NHTSA Technical Report. DOT HS 809 662. Washington, DC: National Highway Traffic
Safety Administration,  http://www-nrd.nhtsa.dot.gov/Pubs/809662.PDF.
18 Wenzel, T.  2012a. Assessment of NHTSA's Report "Relationships Between Fatality Risk, Mass, and Footprint in
Model Year 2000-2007 Passenger Cars and LTVs."  Final report prepared for the Office of Energy Efficiency and
Renewable Energy, US Department of Energy. Lawrence Berkeley National Laboratory. August. LBNL-5698E.
http://energy.lbl.gov/ea/teepa/pdf/lbnl-5698e.pdf.
19 Wenzel, T.  2012b. An Analysis of the Relationship between Casualty Risk per Crash and Vehicle Mass and
Footprint for Model Year 2000-2007 Light-Duty Vehicles. Final report prepared for the Office of Energy Efficiency
and Renewable Energy, US Department  of Energy. Lawrence Berkeley National Laboratory. August. LBNL-
5697E.
http ://energy. Ibl. gov/ea/teepa/pdf/lbnl-5697e .pdf.
20 EPA. 2012. Peer Review of LBNL Statistical Analysis of the Effect of Vehicle Mass & Footprint Reduction on
Safety (LBNL Phase 1  and 2 Reports). Final report prepared by Systems Research and Application Corporation for
the Office of Transportation and Air Quality, US Environmental Protection Agency. August. EPA-420-R-12-020.
21 Wenzel, Tom. 2013. Assessment ofDRI's Two-Stage Logistic Regression Model Used to Simultaneously Estimate
the Relationship between  Vehicle Mass or Size Reduction and U.S. Fatality Risk, Crashworthiness/Compatibility,
                                                     8-64

-------
                                                             Assessment of Vehicle Safety Effects
and Crash Avoidance.  Draft report prepared for the Office of Energy Efficiency and Renewable Energy, US
Department of Energy; Lawrence Berkeley National Laboratory; January.
22 Wenzel, Tom. 2014.  Sensitivity of Light-Duty Vehicle Crash Frequency per Vehicle Mile of Travel to Additional
Vehicle and Driver Variables. Draft report prepared for the Office of Energy Efficiency and Renewable Energy, US
Department of Energy; Lawrence Berkeley National Laboratory; February.
23 Wenzel, Tom.  2015. Effect of Accounting for Crash Severity on the Relationship between Mass Reduction and
Crash Frequency and Risk per Crash. Draft report prepared for the Office of Energy Efficiency and Renewable
Energy, US Department of Energy; Lawrence Berkeley National Laboratory; March.
24 National Academy of Sciences. Cost, Effectiveness and Deployment of Fuel Economy Technologies for Light-
Duty Vehicles; Table 6.22.
25Puckett, S.M. and Kindelberger, J.C. (2016, June). Relationships between Fatality Risk, Mass, and Footprint in
Model Year 2003-2010 Passenger Cars and LTVs - Preliminary Report. (Docket No. NHTSA-2016-0068).
Washington, DC: National Highway Traffic Safety Administration.
2675 Fed. Reg. 25324 (May 7, 2010); the discussion of planned statistical analyses is on pp. 25395-25396.
27 Wenzel, T., (2016, June) Assessment of NHTSA's Report "Relationships Between Fatality Risk, Mass, and
Footprint in Model Year 2003-2010 Passenger Cars and LTVs" Preliminary report (LBNL-1005177) prepared for
the Office of Energy Efficiency and Renewable Energy, US Department of Energy.
28 Allison, P.O.. Logistic Regression Using SAS, Theory and Application.  SAS Institute Inc., Gary NC, 1999.
29Menard, S.  Applied Logistic Regression Analysis, Second Edition. Sage Publications, Thousand Oaks CA, 2002.
30 O'Brien, R.M.  "A Caution Regarding Rules of Thumb for Variance Inflation Factors," Quality and Quantity, (41)
673-690, 2007.
31 Samaha, R. R., Prasad, P., Marzougui, D., Cui, C., Digges, K., Summers, S., Patel S., Zhao, L., & Barsan-Anelli,
A. (2014, August). Methodology for evaluating fleet protection of new vehicle designs: Application to lightweight
vehicle designs. (Report No. DOT HS 812 051 A). Washington, DC: National Highway Traffic Safety
Administration.
32 Samaha, R. R., Prasad, P., Marzougui, D., Cui, C., Digges, K., Summers, S., Patel, S., Zhao, L., & Barsan-Anelli,
A. (2014, August). Methodology for evaluating fleet protection of new vehicle designs: Application to lightweight
vehicle designs, appendices. (Report No. DOT HS 812 05 IB). Washington, DC: National Highway Traffic Safety
Administration.
33 Singh, H., Kan, C-D., Marzougui, D., & Quong, S. (2016, February). Update to future midsize lightweight vehicle
findings in response to  manufacturer review and IIHS  small-overlap testing (Report No. DOT HS 812 237).
Washington, DC: National Highway Traffic Safety Administration.
34 U.S. Environmental Protection Agency (2012, August). Light-Duty Vehicle Mass Reduction and Cost Analysis —
Midsize Crossover Utility Vehicle (Report No. EPA-420-R-12-026).
35 NHTSA Laws and Regulations,
http://www.nhtsa.gov/Laws+&+Regulations/Vehicles?ruleSortBy=fmvss&ruleOrder=asc.
36 IIHS HLDI, http://www.iihs.org/iihs/ratings/ratings-info/frontal-crash-tests.
37 FEV, '"''Light-Duty Vehicle Mass-Reduction and Cost Analysis - Midsize Crossover Utility Vehicle" July 2012,
EPA Docket: EPA-HQ-OAR-2010-0799.
38 Systems Research and Application Corporation, "Peer Review of Demonstrating the Safety and Crashworthiness
of a 2020 Model-Year, Mass-Reduced Crossover Vehicle (Lotus Phase 2 Report)," February 2012, EPA docket:
EPA-HQ-OAR-2010-0799.
39 RTI International, "Peer Review of Lotus Engineering Vehicle Mass Reduction Study" EPA-HQ-OAR-2010-
0799-0710, November  2010.
40 "Preliminary Safety Analysis Metrics," note from NHTSA/Volpe to Todd Sherwood, EPA, via email, dated
January 13, 2016.
                                                     8-65

-------
                                            Assessment of Alternative Fuel Infrastructure
Table of Contents

Chapter 9:   Assessment of Alternative Fuel Infrastructure	9-1
  9.1    Overview	9-1
  9.2    Electric Vehicle Infrastructure	9-2
     9.2.1   Classification of Electric Vehicle Supply Equipment (EVSE)	9-3
       9.2.1.1  Level 1 EVSE	9-3
       9.2.1.2  Level 2 EVSE	9-4
       9.2.1.3  Direct Current (DC) Fast Charge	9-5
     9.2.2   Where People Charge	9-7
     9.2.3   Installation Costs and Equipment Costs	9-11
       9.2.3.1  Installation Costs (Residential and Non-Residential)	9-12
       9.2.3.2  Installation Costs Trends	9-13
       9.2.3.3  EVSE Equipment Costs	9-13
       9.2.3.4  Equipment Costs Trends	9-15
     9.2.4   Status of National PEV Infrastructure	9-15
       9.2.4.1  Number of Connectors and Stations	9-15
       9.2.4.2  Trends, Growth	9-17
       9.2.4.3  Networks and Corridors	9-19
         9.2.4.3.1  West Coast Electric Highway (Baja California to British Columbia)	9-19
         9.2.4.3.2  Northeast Electric Vehicle Network (D.C. to Northern New England)... 9-20
         9.2.4.3.3  Tesla Super Charging Network (Coast to Coast)	9-20
         9.2.4.3.4  FAST Act -  Nationwide Alternative Fuel Corridors	9-20
       9.2.4.4  Challenges and Opportunities with PEV Infrastructure	9-20
         9.2.4.4.1  Challenge - Multi-Unit Development (MuD)	9-20
         9.2.4.4.2  Challenge - Increasing Battery Capacity	9-21
         9.2.4.43  Challenge and Opportunity - Inductive Charging	9-21
         9.2.4.4.4  Opportunity - Vehicle Grid Integration (VGI)	9-22
         9.2.4.4.5  Opportunity - Utility Demand Response	9-22
       9.2.4.5  Further Analysis and Developments	9-22
       9.2.4.6  Status of Public PEV Infrastructure Network	9-23
       9.2.4.7  Summary of PEV Infrastructure	9-25
  9.3    Hydrogen Infrastructure Overview	9-25
     9.3.1   Hydrogen Network Development and Status	9-27
     9.3.2   Retail Experience	9-30
     9.3.3   Hydrogen Fueling Station Capacity	9-32
     9.3.4   Hydrogen Fueling Station Costs	9-32
     9.3.5   Paradigms for Developing Networks	9-35
     9.3.6   Challenges and Opportunities for Hydrogen Fueling Stations	9-39
  9.4    Fueling Infrastructure for Other Alternative Fuel Vehicles	9-41
  9.5    Summary of Alternative Fuel Infrastructure	9-41

Table of Figures
Figure 9.1 Charging Schematics for Electric Vehicles	9-3
Figure 9.2 J1772 Connector and Cord Sets for Level 1 EVSEs	9-4
Figure 9.3 Commercial and Residential Level 2 EVSEs	9-5
Figure 9.4 DC Fast Charge Connectors	9-6

-------
                                                  Assessment of Alternative Fuel Infrastructure
Figure 9.5 SAE Charging Configurations and Ratings Terminology	9-6
Figure 9.6 Charging Pyramid	9-8
Figure 9.7 Key Findings of the EV Project by INL	9-9
Figure 9.8 Key Findings of the UC Davis White Paper on EV Charging	9-10
Figure 9.9 Key Findings of NREL's California Statewide PEV Infrastructure Assessment	9-11
Figure 9.10 Important Messages from the 2013 EPRI and 2015 DOE Reports	9-12
Figure 9.11 Average Residential Level 2 Installation Costs by Metro Area	9-13
Figure 9.12 Range of Level 2 Equipment Costs by Type	9-14
Figure 9.13 Projected Global EVSE Annual Sales by Region: 2016-2025	9-15
Figure 9.14 Comparison of EVSE Connector Types	9-17
Figure 9.15 Annual Growth of Level 2 Connectors	9-18
Figure 9.16 Annual Growth of DC Fast Connectors	9-19
Figure 9.17 Hydrogen Production Methods in California	9-27
Figure 9.18 Locations of California's Funded and Operational Network of 50 Hydrogen Fueling Stations	9-29
Figure 9.19 Projections for Cost Reductions in Hydrogen Fueling Infrastructure	9-34
Figure 9.20 Sample Financial Evaluation of a 180 Kg/Day Delivered Gaseous Hydrogen Fueling Station Based on
            Experience in California	9-35
Figure 9.21 Optimization of Coverage in STREET	9-36
Figure 9.22 Multiple-Station Coverage Estimation in CHIT	9-37
Figure 9.23 Nationwide Identification and Timing of Urban Areas for FCEV Markets in SERA	9-38
Figure 9.24 Fleet-Based Planning for Infrastructure Networks in the Northeast States Produced by NEESC	9-38


Table of Tables

Table 9.1 Vehicle Range Added at Various Charging Levels	9-5
Table 9.2 EVSE Unit Cost and Installation Cost Range	9-14
Table 9.3 Number of Non-Residential Connectors (June 6, 2016)	9-17
Table 9.4 Representative Hydrogen Fueling Station Costs	9-33

-------
                                           Assessment of Alternative Fuel Infrastructure
Chapter 9: Assessment of Alternative Fuel Infrastructure

9.1    Overview

   As part of the midterm evaluation, one of the relevant factors to be examined included "actual
and projected availability of public and private charging infrastructure for electric vehicles, and
fueling infrastructure for alternative fueled vehicles."1 In September 2010, EPA, NHTSA, and
CARB issued a joint interim technical assessment report (TAR, or 2010 TAR) on light-duty
vehicle GHG emission standards and Corporate Average Fuel Economy (CAFE) standards for
model years 2017-2025, which supported the final rulemaking issued in 2012. The 2010 TAR
included a discussion of infrastructure for plug-in electric vehicles (PEVs) and hydrogen fueled
fuel cell electric vehicles (FCEVs).  These analyses recognized PEVs and FCEVs, among others,
as technologies that  could potentially be used to meet future CAFE and GHG standards.  In the
2012 final rule, EPA and NHTSA projected that only a few percent of PEVs, and no FCEVs,
would be needed to meet the MY2025 standards;  the agencies' show similar projections with this
Draft TAR analysis as discussed in Chapters 12 and 13.  Since then, electric drive vehicles have
entered the market with significant growth in the number of models offered and have proven to
reduce or eliminate GHG emissions and improve  fuel economy compared to conventional
technologies. In addition, electric drive vehicles have the potential to derive some or all of their
fuel from sustainable pathways with up to 100 percent renewable fuel sources. With zero
tailpipe emissions, and with nearly half of Americans living in the regions where PEVs produce
lower GHG emissions than even the most fuel-efficient gasoline hybrids on the market today
(greater than 50 mpg)2, electric drive vehicles hold the promise to dramatically transform the
future vehicle fleet into one with a lower carbon footprint and petroleum consumption.

   Though the agencies are projecting in this Draft TAR that only a very small fraction of the
fleet will need to be  PEVs to meet the MY2025 standards, alternative fuel vehicles such as
battery electric vehicles (BEVs) and Plug-in Hybrid Electric Vehicles (PHEVs) (collectively
called PEVs), and FCEVs are an essential part of any future vehicle fleet intended to meet long
term climate and air quality goals.  In additional, other alternative fuels such as ethanol (E85)
and compressed natural gas (CNG) have the potential to contribute to GHG emission reductions.
This chapter is intended to provide an overview of the status, costs, and trends in PEV charging
infrastructure and  hydrogen infrastructure today, as well as examine the challenges being
addressed to scale up the infrastructure as advanced vehicle sales grow in response to market
demand and for compliance with the federal standards.

   Electric vehicle charging infrastructure is different from other alternative fuel infrastructure.
PEVs rely on access to the existing electric grid and distribution network.  At a minimum, most
PEVs can charge at low power using the charging equipment supplied with the vehicle; all they
need is access to a standard household electrical outlet with a dedicated circuit.  Since the 2012
FRM, the U.S. Department of Energy (U.S. DOE) has supported efforts to study how and where
PEV drivers charge their vehicles.  This research reveals that,  currently, the majority of charging
is taking place at home.3   Further, public and workplace charging network infrastructure has
greatly expanded,  offering higher power charging in a greater number of locations. This rapid
expansion of PEV infrastructure is continuing to alter the paradigm of charging behavior and
PEV use patterns.  This dynamic paradigm coupled with a rapidly expanding PEV infrastructure
landscape and evolving battery/vehicle technology will impact how additional PEV
infrastructure is planned and developed; it may actually lessen the need for, or change the power
                                              9-1

-------
                                            Assessment of Alternative Fuel Infrastructure
requirements of, future public infrastructure. As discussed more fully in section 9.2, PEV
charging infrastructure expansion may transform how PEVs are viewed and ultimately change
their usage patterns. However, charging infrastructure growth will adjust as vehicle needs
change.

   With regard to hydrogen FCEVs, a robust network of hydrogen stations, comparable to
conventional gasoline stations, is required to facilitate wide-spread commercialization. Although
California may be the first state to plan, fund, and develop a hydrogen station network, other
regions, such as the Northeast states, have commenced hydrogen infrastructure planning and
development.

   This chapter will examine the status of hydrogen fueling infrastructure in the United States
with a focus on progress in California and the Northeast states.  Section 9.3 will draw from
California's  work in planning, funding, and development of a statewide hydrogen station
network and apply  the lessons learned from these efforts toward a national hydrogen
infrastructure.  With current public and private investments in California, the hydrogen network
is currently sufficient for FCEVs to launch in California and establish an example for how other
regions can further develop their markets around the country. While the agencies do not expect
FCEVs to be needed to meet the 2025 national program standards, the agencies recognize the
importance of these vehicles in meeting longer term climate goals.

   This chapter also discusses the status and trends in fueling infrastructure for compressed
natural gas (CNG) vehicles and E85 (Flex-Fuel) vehicles.

9.2    Electric Vehicle Infrastructure

   PEVs store electrical energy in on-board batteries that supply power to electric motors for
vehicle propulsion.  Today's PEVs have on-board chargers, which are systems that monitor,
regulate, and convert AC power from an external source to DC power for on-board storage. The
electricity supplied to these on-board chargers can be managed by off-board Electric Vehicle
Supply Equipment (EVSE) devices which include connectors with well-insulated power cables,
energy management systems, and telemetry systems. EVSEs are often called "chargers" even
though they  do no actual charging. (Figures 9. la and 9. Ib) details the components of an EVSE
and related vehicle and utility equipment  associated with various types of charging described
later in the document.
                                              9-2

-------
                                            Assessment of Alternative Fuel Infrastructure



Utility 120-V AC or
240-V AC
Utility
480-V
3-Phase
1
            Control Device
                                    EVSE
                                                                            EVSE
    Charger

           Cord
DC Fast
Charging Connector
                                                DC Fast
                                                Charging Inlet
                                               Eattery
   AC Level 1 & Level 2 Charging Schematic           DC Fast Charging Schematic

                    Figure 9. la                               Figure 9. Ib
                      Figure 9.1  Charging Schematics for Electric Vehicles4


9.2.1   Classification of Electric Vehicle Supply Equipment (EVSE)

   EVSE devices are typically classified as Level 1, Level 2, or DC Fast Charge. Each of these
types of EVSE is described in further detail below.

9.2.1.1 Level 1 EVSE

   The lowest power, and most common, EVSE is often referred to as Level 1 or a "Level 1
charger" or "Level 1  cord set." A Level 1 cord set provides AC power at 120 volts, and 12 amps
from a standard 3-prong (NEMA 5-15) household electrical plug/receptacle. Most household
garages have a standard 3-prong electrical receptacle on a 15 amp circuit so no additional
electrical work or expense is required.  Although there is no additional expense in this scenario,
the power transfer under Level 1 charging is ultimately limited by the available circuit amperage.

   Most, if not all, OEMs provide a Level 1 cord set at no additional charge with each sale or
lease of a PEV. Since the cost of the Level 1 cord set is factored into the price of the vehicle,
there is no additional out-of-pocket expense to the consumer opting to use this option to charge
their vehicle.

   The hardware at the end of the cord set that physically attaches to the vehicle is called a
connector and is designed to a common architecture standard specified by  SAE J1772.  This
ensures operational and dimensional interoperability between vehicle OEMs and electrical
equipment suppliers.  The J1772 connector utilizes 5 pins to deliver up to 240V at 80 amps of
AC power to the vehicle. The  J1772 connector is used in both Level  1 and Level  2 charging.  In a
Level 1 cord-set, one end terminates in a J1772 connector while the other end terminates in a
standard household 3-prong electrical plug (see Figure 9.2 below).
                                              9-3

-------
                                           Assessment of Alternative Fuel Infrastructure
           From PR Log                 From Plug-In America       From PEV Collaborative

                  Figure 9.2 J1772 Connector and Cord Sets for Level 1 EVSEs
   As mentioned, Level 1 EVSEs provide a low level of power, typically 120V AC at 12 amps,
to the vehicle. At maximum power, a Level 1 EVSE will fully charge a 2015MY Nissan Leaf in
17 hours or a 2015MY Chevrolet Volt in approximately 8 hours. The most common application
of Level 1 charging is residential over-night or in the workplace where a driver may park for 8 to
9 hours a day. Due to the relatively slow charge rate of only 2-5 miles of range per hour
charging, Level  1 EVSEs may be most appropriate for PHEVs with smaller battery packs or for
BEVs at locations with long dwell times. As battery size and vehicle range continues to grow
with new PEV product offerings, the practicality of Level 1 may decrease.

9.2.1.2 Level 2 EVSE

   For higher power charging, a Level 2 EVSE provides AC power up to 240V at up to 80 amps.
Level 2 EVSEs also use the aforementioned SAE J1772 connector. A Level 2 EVSE can be
either hard-wired to a dedicated building circuit or plugged into a 240V wall receptacle similar to
that used for an electric dryer, range, or recreational vehicle (RV) electrical receptacle.  A Level
2 EVSE is not standard with the purchase of most PEVs. In addition, many household garages
do not have the required wiring to support a Level 2 EVSE. Therefore, additional costs are
associated with installing Level 2 charging; these cost are discussed in section 9.2.3.

   The advantage of a Level 2 EVSE over a Level 1  EVSE is the higher power output.  This
allows most PEVs to charge in a fraction of the time required using Level 1 EVSEs. For
example, a Level 2 EVSE can charge a 2015MY Nissan Leaf equipped with a 6.6kW on-board
charger in approximately 4  hours.  Since a Level 2 EVSE can deliver more power to a PEVs on-
board charger, they are most appropriate for PEVs with larger batteries, or in locations where the
vehicle may have a shorter  dwell time, such as parking lots, shopping centers, churches, libraries,
civic buildings, college campuses, etc. Figure 9.3 below depicts several commercial and
residential Level 2 EVSEs.
                                             9-4

-------
                                              Assessment of Alternative Fuel Infrastructure
          From Clipper Creek              From AeroViroment             From Charge Point

                       Figure 9.3  Commercial and Residential Level 2 EVSEs
9.2.1.3 Direct Current (DC) Fast Charge

   Direct Current (DC) Fast Charge is a fast, high power charging system that uses high voltage,
3-phase Alternating Current (AC) grid electricity and converts it to DC power for direct storage
in vehicle batteries. Unlike Level 1 and Level 2 charging, the conversion of AC power to DC
power occurs off-board in the charging equipment.  This additional conversion equipment
combined with the very high input power (3-phase at 480V or higher) makes DC Fast Charge
systems significantly higher in cost to install, operate, and maintain.  As a result, nearly all DC
Fast Chargers are located in public, workplace, or commercial settings.

   Table 9.1 details the various charging levels, the supply power requirements, and the
additional ranges per unit of time and power.
                     Table 9.1  Vehicle Range Added at Various Charging Levels5
                  Charging Level
                   AC Level 1
Vehicle Range Added per
Charging Time and Power
   4 mi/hour (a1. 1.4kW

   6 mi/hour   1.9kW
    Supply Power
     120VACY20A
   (12-16A continuousl
                   AC Level 2
   10 mi/hour (a: 3.4kW

   20 mi/hour @ 6.6kW

  60 mi/hour @ 19.2 kW
  208/240VAC/20-100A
   (16-SOA continuous)
                 DC Fast Charging
 24 mi/20minutes @24kW

 50 mi/20minutes @50kW

 90 mi/20minutes @90kW
   208/480VAC 3-phase
f input current proportional to
     output power;
     -20-400.4 AC)
                                                 9-5

-------
                                                     Assessment of Alternative Fuel Infrastructure
   DC fast chargers can have different types of connectors (to connect to the vehicle itself);
currently there is no universal standard. Generally, DC fast charge connectors fall into one of
three types: SAE Combo Connector, CHAdeMO,A or Tesla Superchargers and examples of each
are provided in Figure 9.4 below.
                                      »
                                            -*
                       Tesla Connector               SAE Combo

                                  Figure 9.4 DC Fast Charge Connectors
    CHAdeMO
   Figure 9.5 below details the SAE J1772 connector, the SAE Combo connector (DC Fast
charge) and the charging times associated with each. For example, using a Level 2 EVSE, a
BEV with a 25 kWh battery pack and a 6.6 kW on-board charger, can charge from a 20 percent
state of charge (SOC) to a 100 percent SOC in approximately 3.5 hours.  Using DC  fast charge,
this same vehicle can complete the same charge in approximated 1.2 hours.  Given the shorter
charge times associated with DC fast charging, this type of infrastructure is  well suited for
interregional corridors or along interstate routes.
                           AC level 1
                           (SAE J1772'")

                           PEV includes on-board charger
                           120V, 1.4 kW@ 12 amp
                           120V, 1.9 kW@16amp
                           Est. charge time:
                           PHEV: 7hrs (SOC" - 0% to full)
                           BEV:17hrs(SOC-20%tofull)
                           AC level 2
                           (SAEJ1772™)

                           PEV includes on-board charger (see below for different
                           types!
                           240V, up to 19.2 kW (80 4)
                           Est. charge time for 3.3 kW on-board charger
                               PEV: 3 hrs (SOC--0% to full)
                               BEV: 7 hrs (SOC- 20% to full)
                           Est. charge time for 7 kW on-board charger
                               PEV: 1.5 hrs (SOC*-0% to full)
                               BEV: 3.5 hrs (SOC-20% to full]
                           Est. charge time for 20 kW on board charger
                               PEV: 22 min. (SOC* - 0% to full)
                               BEV: 1.2 hrs (SOC-20% to full)
DC Level 1
(SAEJ1772-)

EVSE includes an off-board
charger
200-500 V DC, up to 40 kW (80 A)
Est. charge time (20 kW off-board
charger):
    PHEV: 22 min. (SOC' -
    0% to 80%)
    BEV: 1.2 hrs. (SOC - 20%
    to 100%)
DC Level 2
(SAEJ1772™)

EVSE includes an off-board
charger
200 500 V DC up to 100 kW (200
A)
Est. charge time (45 kW off-board
charger):
    PHEV:10min.(SOC-0%
    to 80%)
    BEV: 20 min. (SOC-20%
    to 80%)
            Voltages are nominal configuration voltages, not coupler ratings
            Rated Power is at nominal configuration operating voltage and coupler rated current
            Ideal charge times assume 90% efficient chargers, ISOWto 12V loads and no balancing of Traction Battery Pack
            Notes:
            1) BEV (25 kWh usable pack size) charging always starts at 20% SOC, faster than a 1C rate (total capacity charged in one hour) will also stop at 80% SOC instead of
            100%
            2) PHEV can start from 0% SOC since the hybrid mode is available.                                        ver 100312
                    Figure 9.5 SAE Charging Configurations and Ratings Terminology
A CHAdeMO is an abbreviation of the phrase "CHArge de MOve," which is equivalent to the translation of Japanese
  phrase "move using charge" or "move by charge."
                                                        9-6

-------
                                            Assessment of Alternative Fuel Infrastructure
9.2.2   Where People Charge

   In the most general terms, charging of a PEV occurs in one of two places: at home or away
from home.  Away from home charging can be further subdivided into workplace charging or
non-work (public) charging.  Both home and the workplace are well suited for Level 1 charging
since an individual usually spends several contiguous hours at both locations. Some public
installations, like airport parking, can be accommodated with Level 1 EVSEs. Movie theaters,
shopping centers, hospitals, churches, or other publicly accessible locations are better suited for
Level 2 EVSEs since an individual usually has a shorter dwell time at these public charging
locations.  DC fast charging sites could be well placed along routes serving inter-regional or
inter-state travel such as roadside rest areas.  DC fast charge locations are much less common
than Level 1 or Level 2 charging sites.  As detailed in section 9.2.1.3, DC fast chargers deliver
high, direct current power to  a PEV and are most appropriate where vehicles have a short dwell
time and need a large amount of power.

   Many studies have been, and continue to be, conducted on the charging patterns and
behaviors of PEV drivers.  The results from these various studies can be summarized using a
construct called the "charging pyramid."  Argonne National Laboratory developed one such
"charging pyramid" (Figure 9.6) which graphically depicts the interconnected relationships
between charger type, location, costs, and frequency of charge events. The majority of charging
events occur at home, at lower costs, and over longer periods of time. However, as power
transfer rates increase, charging time decreases, but costs increase leading to fewer charging
events at that level.  As the charging pyramid depicts, the majority of charge events occur at low
cost Level 1, followed by more expensive Level 2. The fewest charging events occur at
relatively high cost DC fast chargers.
                       Cars charge most often where they are parked most often.
                                   Source: Argonne National Laboratory
                                              9-7

-------
                                            Assessment of Alternative Fuel Infrastructure
                                Figure 9.6 Charging Pyramid
   One study regarding charging behavior was conducted with the EV Project by the Idaho
National Laboratory (INL).  In 2009, the U.S. DOE funded the EV Project which was an
infrastructure deployment and analysis project where one of the goals was to evaluate the
effectiveness of PEV infrastructure. The ultimate goal of the EV Project was to utilize lessons
learned from the early deployment of infrastructure and vehicles and enable the efficient
deployment of subsequent PEVs and infrastructure across the United States.

   The EV Project included an analysis of the charging patterns of over 4,000 Nissan Leaf
drivers studied from October 2012 through December 2013. Study participants were given a
Level 2 EVSE for home charging, and their vehicles were outfitted with tracking devices.
Although the participants were early adopters and had access to Level  2 charging, the key
finding of this study can be interpreted for the larger PEV population.  Figure  9.7 shows the key
findings:
                                              9-8

-------
                                                  Assessment of Alternative Fuel Infrastructure
            Leaf drivers relied on home charging for the bulk of their charging. Of all charging
            events, 84 percent were performed at drivers' home locations. Over 80 percent of those
            home charges were performed overnight, and about 20 percent of home charges were
            performed between trips during the day.

            The remaining 16 percent of charging events were performed away from home. The
            vast majority of these were daytime Level 1 or Level 2 charges.

            Overall, usage of DCFC (DC fast charging) by drivers of vehicles in this study, all
            having access to a Level 2 charging unit at home and some having workplace charging
            access, was low. DC fast charging (all away from home) represented only about 1
            percent of all charging events and charging energy consumed. Ignoring charges by
            vehicles that never charged away from home, DC fast chargers were used for 6 percent
            of all away-from-home charging events. However, some drivers used DC fast chargers
            more than others and may have relied on fast charging to meet their need for driving
            range.

            Although all vehicles in this  study had access to home charging, some vehicles rarely
            charged at home. Instead, they relied on frequent away-from-home charging during the
            day. This demonstrates the viability of publicly accessible and/or workplace charging
            infrastructure for drivers of electric vehicles without access to home charging.
                                • Home Overnight U or L2

                                • Home Daytime LI or L2

                                • Away Overnight LI or L2

                                • Away Overnight DCFC*

                                • Away Daytime LI or L2

                                 Away Daytime DCFC


                                ' «1% of all charge events
       Figure 1. Percent of charging events performed by location,
       power level, and time of day.
                                                     40%
8 30%

a
•3 20'-,
  10%
  o%
                                                          Percent of Charges Away from Home
                           Figure 9.7 Key Findings of the EV Project by INL6
   In addition to INL's work on PEV charging, in 2013, the Institute of Transportation Studies at
the University of California, Davis (ITS-Davis) published a white paper titled, California
Statewide Charging Assessment Model for Plug-in Electric Vehicles: Learning from Statewide
Travel Survey1 This research focused on how different infrastructure types/locations can enable
more BEV driving. (See Figure 9.8).
                                                     9-9

-------
                                                Assessment of Alternative Fuel Infrastructure
                 If all statewide vehicles were 80 mile range BEVs and began the day with a full
                 charge, 71 percent of miles (95 percent of home-based tours) are possible with home
                 charging alone. Travel that requires some charging accounts for a corresponding 29
                 percent of miles (5 percent of tours). See chart below.

                 Workplace charging can enable about 7 percent more electric vehicle miles traveled
                 (eVMT), public Level 2 at stops greater than 1.5 hours could provide an additional 4
                 percent of eVMT, and DC fast charging could provide an additional 12 percent of
                 eVMT.

                 Scenarios show that for a 30 mile range PHEV, 61 percent of miles could be
                 completed with home charging alone.4
                   Percent of Total CA Daily VMT Enabled by Number of Charging
                          Events/Types with entire fleet of 80 mile BEVs
                                                                      Unserved

                                                                     • 3 Charge Event or More

                                                                      2 Charge Events or Fewer

                                                                      Home Charging Miles
                             100            200            300
                                  Number of DC Fast Charge Locations
           Source: Institute of Transportation Studies - DC Davis
                 Figure 9.8 Key Findings of the UC Davis White Paper on EV Charging
   Building upon the body of knowledge developed by INL, UC Davis and others, the California
Energy Commission (CEC) sponsored the National Renewable Energy Laboratory (NREL) to
conduct a PEV Infrastructure analysis for California. This analysis was developed with the goal
of facilitating charging infrastructure for 1.5 million ZEVs on California roadways by 2025 as
envisioned by  California Governor Jerry Brown's Executive Order B-16-2012 in March 2012.
Key findings from NREL's assessment are described in Figure 9.9 below.
                                                  9-10

-------
                                              Assessment of Alternative Fuel Infrastructure
              •   Entities should identify their objectives for installing EVSE before trying to
                  determine EVSE numbers, types (such as, Level 1, Level 2, or fast charge), and
                  locations.

              •   Near-term PEV charging will occur primarily at home, so this is the greatest
                  opportunity for charging infrastructure support for the next few years. Other
                  outstanding near-term infrastructure opportunities include workplaces and
                  multiunit dwellings where management has indicated support for infrastructure
                  and surveys indicate likely PEV adoption; garaged fleet locations that have or will
                  have significant numbers of PEVs; and crowded airport and commuter parking
                  locations, provided certain conditions are met.

              •   In many cases, there should be a reasonable belief that installed EVSE will be
                  used by significant numbers of PEVs; however, there are compelling reasons to
                  consider installing EVSE infrastructure besides expected short-term use. Some of
                  these reasons address safety and convenience concerns, as well as building
                  consumer confidence in PEVs and associated infrastructure.

         Figure 9.9 Key Findings of NREL's California Statewide PEV Infrastructure Assessment

   Ultimately, uncertainty regarding "where people charge" will be managed with the growth of
various charging infrastructure investments and pricing policies.  At this time, there does not
appear to be a clear trend or convergence for where non-home based charging will occur.
However, the following factors will likely continue to influence where charging occurs:

       •   PEV vehicle technology (especially driving range and rate of charging) influencing
           the need and convenience of daily, nightly, or travel corridor charging
       •   Employers increasingly providing workplace charging8
       •   Many public chargers currently operating for free may eventually implement fees to
           charge, (again, no clear trend has yet been established but a wide  range of fees and
           non-fee structures are being explored depending on the site host business model)
       •   Electric utilities are beginning to make direct investment in the PEV infrastructure
           (see section 9.2.4.5) and may distribute the costs over a large ratepayer population
       •   DC Fast charge networks are growing rapidly and may affect usage of Level 2 EVSEs
9.2.3  Installation Costs and Equipment Costs

   One factor driving PEV adoption rates is the cost savings related to fuel. Electricity is cheaper
than gasoline on a per-mile basis; refueling a PEV may  require additional equipment and
installation costs.  This section will explore costs related to capital equipment and installation for
PEV refueling.

   As referenced in section 9.2.2, the majority of PEV drivers predominantly charge at home.
Approximately 85 percent of charging events occur at home and much of that is at Level 1.
Since Level 1 cord-sets typically are included with PEVs, and many homes have a 120V power
outlet in close proximity to the PEV, a large portion of PEV drivers incur no additional expense
related to EVSE purchase or EVSE installation costs.
                                                9-11

-------
                                             Assessment of Alternative Fuel Infrastructure
9.2.3.1 Installation Costs (Residential and Non-Residential)
   In November 2015, the U.S. DOE released a report titled, Costs Associated With Non-
Residential Electric Vehicle Supply Equipment. This report provides the most recent
compilations of EVSE costs and factors influencing cost trends. This report was a synthesis of
various studies on the subject in addition to data collected from EVSE owners, electric utilities,
manufacturers, and installers.  One study included in this synthesis was a 2013 report from the
Electric Power Research Institute (EPRI) titled Electric Vehicle Supply Equipment Installed Cost
Analysis.

   The 2015 U.S. DOE report identified several cost drivers associated with the installation of
Level 2 EVSEs. These drivers include:

       •   Trenching or boring to install electrical conduit from the transformer to the electrical
           panel or from the electrical panel to the EVSE;
       •   Upgrading the electrical panel to create dedicated circuits for each EVSE;
       •   Upgrading the electrical service to provide sufficient electrical capacity for the site;
       •   Locating EVSE on parking levels above or below the level with electrical service;
       •   Meeting accessibility requirements such as ensuring the parking spaces are level.
   Figure 9.10 shows some important messages from the reports:
         It is important to work with the electric utility early in the process to minimize costs,
               optimize the electrical design, and eliminate scheduling bottlenecks.

        • Level 2 commercial sites that required special work such as trenching or boring
         were about 25 percent more costly than those that did not need special work.

        • Fundamental EVSE Electrical Needs:

           1. A dedicated circuit for each EVSE unit on the electrical panel.

           2. Sufficient electrical capacity from the utility connection to the electrical
             panel.

           3. Sufficient electrical capacity at the panel.

        • Assuming $100 per foot to trench through concrete, lay the conduit, and refill,
         it would cost $5,000 to trench 50 feet.

        • Upgrading the electrical service for future EVSE loads and installing conduit to
         future EVSE locations during the initial EVSE installation can result in
         significant future cost savings.

             Figure 9.10 Important Messages from the 2013 EPRI and 2015 DOE Reports
   The 2015 U.S. DOE report identifies labor costs associated with non-residential EVSE
installation as a variable but ultimately based on the contractor's hourly rate and the time it takes
                                               9-12

-------
                                            Assessment of Alternative Fuel Infrastructure
to perform the work. These costs are affected by the contractor's experience and typical labor
rates in the geographic location.

  Residential installation costs for Level 2 EVSEs can vary significantly by geographic region.
This may be attributed to varying labor rates and material costs across regions, as well as the
condition and age of existing housing stock. For example, the EPRI report suggests that between
10 and 20 percent of the installations studied required electrical upgrades.9 These upgrades and
associated costs are less necessary in newer construction where higher capacity electrical panels
are more common.  Additionally, installation costs are lowest when a home has an existing 240V
receptacle on a dedicated circuit. Figure 9.11 from the EPRI report illustrates the geographic
installation costs for Level 2 EVSEs in 12 regions across the United States.
               o
               u
               o

                                   - - - - Weighted Average
              Figure 9.11 Average Residential Level 2 Installation Costs by Metro Area10
9.2.3.2 Installation Costs Trends

   EVSE installation costs have been trending downward since 2009. As mentioned, many of
the installations included in the EPRI study and the EV Project were part of demonstration
programs that required prevailing wages to be paid. These programs are phasing out, and in a
competitive market it is expected that labor rates will decrease 15-25 percent.  Additionally,
with the expected increase in the number of EVSE installations, the resulting competition for
these projects and associated large scale material procurements should help continue the
downward trend in installation costs.11

9.2.3.3 EVSE Equipment Costs

   The aforementioned 2015 U.S. DOE report includes recent EVSE equipment costs and factors
influencing cost trajectories.
                                              9-13

-------
                                            Assessment of Alternative Fuel Infrastructure
   Cost drivers for EVSEs include charging level and amperage, number of charging ports or
connectors, mounting option, advanced features such as network communication, point of sale
capability, access control features (radio frequency identification (RFID)), and intended use
(home vs commercial). As a result EVSE costs can vary greatly depending upon the
manufacturer and the cost drivers included with a specific EVSE installation.  In the November
2015 U.S. DOE report, the costs for EVSE non-residential equipment were estimated using a
variety of sources. The findings summarized in Table 9.2 and Figure 9.12 show similar cost
estimates for the equipment itself and, for example, represent a range of approximately $400-
$6,500 for Level 2 EVSE equipment and an additional $3,000, on average, for installation of the
equipment.
                      Table 9.2 EVSE Unit Cost and Installation Cost Range12
EVSE Type
Level 1
Level 2
DCFC
EVSE Unit* Cost Range
(single port)
$300-$1,500
$400-$6,500
$10,000-140,000
Average Installation Cost
(per unit)
not available
-$3,000
EV Project (INL 2015b)
-$21,000
EV Project (INL 201 15d)
Installation Cost Range (per unit)
$0-$3,000**
Source: Industry Interviews
$600-$12,700
EV Project (INL 2015b)
$4,000-151,000
EV Project (INL 2015d)
and (OUC 2014)
'EVSE unit costs are based on units commercially available in 20/5.
"The $0 installation cost assumes the site host is offering an outlet for PEV users to plug in their Level 1 EVSE cordsets and that the

      $500-51,000
                           $1,200-51,700
                                                $1,700-$2,700
                                                                          $3,000-$6,000
                      Figure 9.12 Range of Level 2 Equipment Costs by Type
Source: Costs Associated With Non-Residential Electric Vehicle Supply Equipment. U.S. DOE, November 2015.
Image from Kristina Rivenbark, New West Technologies
                                              9-14

-------
                                            Assessment of Alternative Fuel Infrastructure
9.2.3.4 Equipment Costs Trends

   From 2012, the first full year of PEV sales, the global market for PEVs has grown from
approximately 30,000 vehicles to nearly 500,000 in 2015, an impressive compound annual
growth rate of 102 percent.13 This expansion in PEV sales has led to solid growth in the EVSE
market. Navigant Research expects the global market for EVSE to grow from around 425,000
units in 2016 to 2.5 million in 2025.  These include sales of all EVSE units—residential and
commercial and Level 1, Level 2, DC fast charging, and wireless charging. While the EVSE
market will continue to grow as long as the PEV market grows, it is growing at a slightly higher
rate than PEVs.14

   Figure 9-13 below illustrates that global sales of commercial and residential EVSEs are
projected to grow to approximately 2.5 million units annually by 2025.15
          3,000,000
          2,500,000
          2,000,000
i North America
 Western Europe
i Eastern Europe
i Asia Pacific
 Latin America
i Middle East & Africa
          1,500,000
          1,000,000
           500,000
                    2016   2017    2018   2019    2020   2021    2022   2023    2024   2025
                                                         (Source: Navigant Research)
               Figure 9.13 Projected Global EVSE Annual Sales by Region: 2016-2025
   The cost of commercial and residential EVSE has declined in recent years through technology
development and through economies of scale. A Level 2 residential EVSE, formerly priced
between $900 - $1,000 in 2013, is currently priced in the $500-$600 range for basic units, and is
expected to fall below $500 in the near term. As robust as the residential EVSE market forecasts
are, the growth in the commercial EVSE market is expected to be even stronger.  The same
market forces that are applying downward price pressure on residential EVSE will also apply to
commercial EVSE.

9.2.4   Status  of National PEV Infrastructure

9.2.4.1 Number of Connectors and Stations
                                              9-15

-------
                                            Assessment of Alternative Fuel Infrastructure
   When analyzing PEV infrastructure, it is important to distinguish between the number of
connectors, the number of stations, and the number of vehicles that may charge at a station
simultaneously. As mentioned in section 9.2.1.1, a connector is defined as the hardware that
physically attaches to a vehicle. A "station" is a physical location that contains at least one
EVSE with at least one connector on a dedicated electrical circuit. However, an EVSE may have
multiple connectors and may be able to charge multiple vehicles simultaneously. A typical
station contains multiple EVSEs, with multiple connectors, on multiple circuits. The physical
layout of a parking facility or the on-site power management systems may limit the number of
vehicles that charge simultaneously.

   Another important distinction when referring to PEV infrastructure is the identification of a
station as either "private" or "public." Consistent with the most common usage, this report refers
to a public station as one that is publicly accessible while a private station designation refers to
one that does not allow access to the general public (e.g., located behind a gate or other method
that limits access). Common examples of private stations include workplace or company fleet
vehicle charging locations restricted to employee access. Public stations include those that are
located in places like parking garages and shopping centers. For this report, Tesla supercharger
DC fast charge stations are considered public stations even though usage is  currently limited to
Tesla vehicle owners.

   The Alternative Fuels Data Center (AFDC), managed by NREL, has  compiled a
comprehensive database on Alternative Fuel Stations.  The AFDC database includes extensive
information of PEV infrastructure including number of stations, number of  connectors, locations
of stations, connector types, and power level of EVSEs. Further information on public and
private stations is  included.  The value from a singular, national database is of such importance
that California law requires station operators to report a station's location and other attributes
directly to NREL for inclusion in this database.16 The database shows that currently there are
over 12,000 public and private PEV charging stations across the United  States with over 38,000
connectors.17 Table 9.3 and Figure 9.14 break down these numbers into further detail.
                                              9-16

-------
                                            Assessment of Alternative Fuel Infrastructure
                  Table 9.3 Number of Non-Residential Connectors (June 6,2016)
Publically Accessible Connectors and Stations
Level 1
California Connectors 647
National Connectors1 2,977
National Stations2 1,546
Level 2 DC Fast Charge Total
8,186 880 9,713
26,859 3,738 33,574
12,176 1,760 13,649
Privately Accessible Connectors and Stations3
| Level 1
California Connectors I 416
National Connectors1 702
National Stations2 145
Level 2 | DC Fast Charge | Total
1,582 | 18 | 2,016
4,633 32 5,367
2,408 23 2,455
Total (Public and Private) Connectors and Stations
National Connectors 3,679
National Stations 1,691
National numbers include California numbers
2A station may include multiple charging types, therefore station
3Does not include home charging
As of 6/6/20 16
Source: Alt Fuels Data Center (US DOE)









31,492 3,770 38,941
14,584 1,783 16,104

total is not a direct summation of types.




Number of Connectors by Type (National)
Includes Public and Private EVSEs
1.0%
3.9% 	 ^ ^




Source: AFDC - November 09, 2015


• J1772 (Level 2 & Level 1)
• CHAdeMO (DC Fast Charge)
1
• SAE Combo (DC Fast Charge)
• Tesla (Level 1, 2 & DC Fast Charge)










                       Figure 9.14 Comparison of EVSE Connector Types
9.2.4.2 Trends, Growth
                                              9-17

-------
                                              Assessment of Alternative Fuel Infrastructure
   The U.S. DOE's AFDC maintains detailed records of public and private charging stations and
connectors dating back to the 1990s.  Figure 9.15 and Figure 9.16, created using this database,
clearly show that since the 2010 TAR, PEV infrastructure has increased substantially.  In 2010,
there were approximately 206 public and private Level 2 charging stations and 347 Level 2
connectors. As of May 2016, there are over 14,000 public and private Level 2 charging stations
and nearly 31,000 Level 2 connectors. That represents nearly a 70 fold increase in the number of
connectors and stations in the past 5 years.

   Of the 14,000 Level 2 charging stations, nearly 12,000 are public stations while the remaining
stations are private.  As noted above, public and private in the context of EV infrastructure refers
to the type of access to the station, not ownership. With regards to ownership of the stations,
approximately 56 percent of Level 2 and DC fast charge stations are currently owned, operated
or networked by one of the four largest private entities in the EV infrastructure market.18
                     Level 2 Connectors by Year (Public and Private)*
         I
         z
             30000
             25000
             20000
             15000
             10000
              5000
                      2010       2011       2012       2013       2014       2015
                                              Year
          * A station may contain multiple connectors, therefore total station count is less than the values presented.

                                                                    Source: NREL, October 2015
                        Figure 9.15  Annual Growth of Level 2 Connectors19
                                                9-18

-------
                                            Assessment of Alternative Fuel Infrastructure
                 DC Fast Charge Connectors by year & type (Public & Private)
           E?
           re
           6
           t;
           Cn
           L_
           U
           G
           •s
           a!
           .a
           3
           Z
               1000
               800
               600
                                                            TESLA

                                                           • SAECOMBO

                                                           • CHAdeMO
                      2010


              i:NREL, October 2015
                              2011
                         2012      2013
                             Year
                                                        2014
                                                                 2015
                        Figure 9.16 Annual Growth of DC Fast Connectors
9.2.4.3 Networks and Corridors

   The data in section 9.2.4.1 and section 9.2.4.2 detail an initial assessment of national PEV
infrastructure. The current PEV infrastructure landscape is robust, and the trends indicate it will
continue with strong growth.

   Equally important to the number of charging stations and connectors is the geographic
location of the stations. Compared to traditional technologies, most current PEVs have a limited
electric range making a strategic network of charging stations critical for interregional or
interstate travel.  As detailed below, several strategic charging networks or corridors are planned,
under development, or are operational.  For a map of current charging stations nationwide, see
the AFDC database.20
9.2.4.3.1
West Coast Electric Highway (Baia California to British Columbia)
   California, Oregon, and Washington are partnering with the Canadian province of British
Columbia to construct the "West Coast Electric Highway," an extensive network of DC fast
charging stations located every 25 to 50 miles along Interstate 5 and other major roadways in the
Pacific Northwest. The goal is to provide a seamless consumer experience for PEV drivers
traveling from Baja California to British Columbia (BC to BC) and all points in between.
Recently, the CEC awarded $8.87 million to four companies to install DC fast charging stations
on nine corridor segments to fill the gaps between the Oregon border and Baja California.  The
CEC also released a second competitive $9.97 million Grant Funding Opportunity to construct
DC fast charge stations on additional interregional corridors in California.21
                                              9-19

-------
                                            Assessment of Alternative Fuel Infrastructure
9.2.4.3.2      Northeast Electric Vehicle Network (D.C. to Northern New England)

   A coalition of 12 Mid-Atlantic and New England states and the District of Columbia have
joined forces to implement the Northeast Electric Vehicle Network. This network will pave the
way for the deployment of an anticipated 200,000 electric vehicles (EVs) and facilitate PEV
travel from D.C. to Maine. Already, more than 1,700 charging stations are publicly available in
this region.22

9.2.4.3.3      Tesla Super Char sins Network (Coast to Coast)

   Tesla Motors has constructed the most extensive network of DC fast charging stations in the
nation.  With over 500 stations and nearly 2,000 connectors, Tesla's proprietary network
provides coast-to-coast mobility to Tesla drivers.23 Although this charging station network is
limited to Tesla vehicles, it provides a model for OEM-based charging networks.

9.2.4.3.4      FAST Act - Nationwide Alternative Fuel Corridors

   In December 2015, President Obama signed the Fixing America's Surface Transportation
(FAST) Act. This bill not only authorized funding for traditional surface transportation projects,
but section 1413 of the bill requires the U.S. Department of Transportation (DOT) to designate
corridors to improve mobility of passenger and commercial vehicles that employ electric,
hydrogen fuel  cell, propane, and natural gas fueling technologies across the U.S. by December
2016. Although the bill does not provide direct funding for alternative fuel infrastructure, the
U.S. DOT can support these corridors through technical assistance, analytical support, peer
review,  marketing and branding. In addition, this bill amended the Congestion Mitigation and
Air Quality Improvement (CMAQ) Program to give priority to designated EV and CNG
corridors.  This bill facilitates the planning activities required in the construction and
implementation of nationwide PEV corridors.

9.2.4.4 Challenges and Opportunities with PEV Infrastructure

   The PEV infrastructure  environment, in its current state, has been in development and
refinement for nearly a decade, and many of the initial challenges have been met: technical
standards, communication protocols, signage and design guidelines have all been adopted. In
addition to its  "Workplace  Charging  Challenge," which aims to achieve a tenfold increase in the
number of U.S. employers  offering workplace charging by 2018, the U.S. DOE, through its
Clean Cities coalition, has awarded $8.5 million to projects in 24 States and the District of
Columbia.  The CEC has funded $40 million for over 7,700 charging stations in California as
well as PEV Community Readiness grants for $5.7 million to help local communities prepare for
PEVs and charging infrastructure.24

   As a result of meeting these initial milestones, consumer acceptance and private capital
market involvement have followed. However, challenges and opportunities surrounding PEV
infrastructure exist and the following paragraphs detail some of the more prominent issues.

9.2.4.4.1      Challenge - Multi-Unit Development (MuD)

   Electric utilities estimate that over 80 percent of all current PEV charging  occurs at home,
usually in a garage with access to electrical power.25 However, nationwide, approximately 36
percent  of households reside in rental housing with 60 percent of those households living in
Multi-unit Dwellings (MuDs). Most MuDs do not provide EVSE or access to electrical power in
                                              9-20

-------
                                            Assessment of Alternative Fuel Infrastructure
proximity to parking.26 In order to expand the PEV market, access to EVSE in MuDs is
important; however, many challenges still exist and stakeholders are beginning to address them.
These include:

       •  Physical Facilities: Age, existing electrical infrastructure, and physical layout of
          parking within a MuD all present unique challenges in installing and operating PEV
          infrastructure.
       •  Diversity: MuDs are comprised of a variety of structures from modern, urban high-
          rise buildings to sprawling, midrise suburban apartment complexes to low-density
          townhome condominiums. Given this physical diversity, there is no universal
          solution or standardized cost for providing EVSE access in MuDs.
       •  Economics: Costs associated with installing, maintaining, and operating EVSE needs
          to be accounted for; however, equitable distribution of these costs among building
          occupants, PEV drivers, and the building owner remains a challenge.
9.2.4.4.2     Challenge - Increasing Battery Capacity

   Vehicle battery costs are declining while energy density is increasing.27  Currently, most
BEVs sold today have a range under 100 miles; the most common BEV on the  road today, the
Nissan Leaf, has a range of 84-107 miles depending upon model year.28  Tesla vehicles are  the
primary exception, offering a range in excess of 200 miles but at a much higher price. However,
several automakers, including General Motors and Tesla, have announced plans to deploy
affordable BEVs with larger battery packs and ranges over 200 miles at a price near $30,000
after federal incentives. These developments hold the potential to alter the need for, and use of,
public charging infrastructure in ways unknown.  For example, larger battery packs will take
longer to charge which may increase the demand for DC fast charging and decrease the demand
for Level 1  and  Level 2 public charging.  However, it is also likely that longer range PEVs will
charge less often which may also impact public charging infrastructure.  These uncertainties
require on-going analysis of the PEV market and charging behavior.

9.2.4.4.3     Challenge and Opportunity - Inductive Charging

   The current PEV charging standards and protocols involve connected, conductive charging.
PEV batteries are charged by physically attaching the vehicle to a power source via the EVSE.
Currently, this physical connection is essential to almost all PEV charging.

   However, some automakers, third party vendors, and charging providers have begun to
develop wireless, inductive charging. Inductive charging uses an electromagnetic field to transfer
energy between the vehicle and the power source where no physical connection is required. This
has the potential to revolutionize charging and charging infrastructure by literally "cutting the
cord."  Inductive charging technology can facilitate charging in non-traditional locations such as
stop lights,  along curbs, or even along routes while the vehicle is in motion. Although, current
inductive charging systems may have lower efficiency, the technology is developing and the
convenience may be worth slightly higher charge rates to many users. In addition, it is likely
that the ease and convenience of inductive charging will draw drivers of conventional vehicles
into PEVs.  How these wireless inductive charging systems are designed, developed, installed,
and utilized by drivers presents uncertainty and an opportunity in the PEV infrastructure
landscape.
                                              9-21

-------
                                            Assessment of Alternative Fuel Infrastructure
9.2.4.4.4      Opportunity - Vehicle Grid Integration (VGI)

   PEVs store a large amount of energy in their on-board batteries.  Current EVSE and charging
specifications and protocols are intended to facilitate the one-way power transfer from the
electrical grid to the vehicle. However, new protocols and standards are being developed and
tested to facilitate the two-way transfer of energy from the vehicle back to the grid; this is
referred to as Vehicle to Grid Integration (VGI). VGI holds the potential to assist electric
utilities in meeting their peak power demands by tapping a new source of power storage - a large
PEV fleet. Many programs across the nation are in place to study VGI including programs in
California, Delaware, and at the U.S. Department of Defense. The CEC, in coordination with the
California Independent System Operator developed  a Vehicle Integration Roadmap29 in 2014 to
outline a way to develop solutions that enable PEVs to provide grid services while still meeting
consumer driving needs.30

9.2.4.4.5      Opportunity - Utility Demand Response

   In broad terms, electrical power on the grid comes from central electric generation facilities.
This electricity is purchased by an electric utility and resold to its customers. Although most
utility bills make the cost of electricity appear relatively uniform, the actual cost to procure
electricity from a generator can vary greatly. Prices can spike (or fall) quickly and with little
notice. Factors that affect the price of electrical power include temperature, weather, time of
day, demand for power, availability of operational power plants, and many others.

   PEVs charge when they are parked, and most vehicles, including PEVs, are parked 96 percent
of the time.31  Therefore, a PEV doesn't need to be charging at all times when it is parked.  This
fact, coupled with emerging technologies that allow an electric utility to communicate with
advanced EVSEs and control the power transfer, gives utilities a unique opportunity. Utilities
could effectively manage PEV power demands in the broader context of regional grid operation,
power generation and supply, local transformer capacity, and price fluctuations. The next
generation of networked EVSEs provides a valuable opportunity for utilities to operate more
efficiently and effectively.

9.2.4.5 Further Analysis and Developments

   Commercial OEM-built PEVs have been around  for nearly two decades while more recent,
modern advanced battery technology PEVs have been on the market for approximately five
years.  Over that time, vehicle technology has changed dramatically and is still continuing to
evolve.  With regards to the technology adoption curves for PEVs, the market is currently
transitioning from the "innovators" (a.k.a. first adopters) phase to the "early adopters" (a.k.a. fast
followers) phase.  As a  result of this transition and technology advancement, charging behavior
has changed and is continuing to  evolve. Further study of charging patterns and behavior,
optimal charging network configuration, and public  charging infrastructure sufficiency, are
warranted and currently being investigated by many stakeholders. The following is a partial list
of additional analysis and implementation efforts in  the area of PEV infrastructure which should
yield results that will enhance the current level of understanding in this topic and enable even
more efficient investment in public charging infrastructure:

       •  As  mentioned in section 9.2.2, NREL conducted a statewide PEV infrastructure
          analysis for  the CEC.  The CEC has recently contracted with NREL to use this
                                              9-22

-------
                                            Assessment of Alternative Fuel Infrastructure
          analysis as a basis to create an actionable plan that will prioritize specific charging
          locations and guide regional PEV infrastructure planning and other stakeholder
          actions in California.  The recommendations from these studies can be utilized by
          other states interested in promoting ZEVs in their jurisdiction. The CEC also funded
          12 PEV planning regions which will each develop charging infrastructure plans along
          with other critical actions to prepare for increasing numbers of PEVs. The lessons
          learned from these planning activities can be used by local agencies in other states.

       •  The California Public Utilities Commission (CPUC), the entity that regulates Investor
          Owned Utilities (IOU) and sets rate tariffs in California, has approved Phase 1 pilot
          projects by two lOUs and is reviewing a proposal by a third IOU. Combined, the two
          approved pilot projects aim to install up to 5,000 public charge stations or related
          infrastructure. When these proposals come to fruition, not only will the large number
          of new charging stations transform the current PEV infrastructure landscape, but the
          introduction of electric utilities into the infrastructure marketplace could be
          transformative. The U.S. DOE EV Everywhere program is working with other states
          to encourage similar actions and several  states have already commenced action.
          Examples are included below:

       •  The State of Oregon has introduced SB 1547 (Beyer), which allows their PUC to
          direct electric  companies to file applications for programs to accelerate transportation
          electrification, including customer rebates for electric vehicle charging  and related
          infrastructure.

       •  The New York Power Authority (NYPA), and others, are collaborating in an initiative
          called ChargeNY which aims to reach 3,000 PEV charging stations to support an
          expected 30,000 - 40,000 PEVs on the road in New York by 2018

       •  In March 2016, Utah lawmakers enacted SB 115 (Snow), the Sustainable
          Transportation and Energy Plan (STEP). STEP establishes a five-year pilot program,
          under which regulators will authorize the State's power company, Rocky Mountain
          Power, to spend up to $2 million per year on electric vehicle infrastructure.
   California enacted SB  350 (de Leon) which directs the CPUC to guide the lOUs' investments
in the widespread transportation electrification including the deployment of charging
infrastructure. This law is significant for several reasons: it will allow lOUs to ultimately
commence "phase 2" electrification programs if they are determined to meet specific
requirements, thereby potentially greatly expanding infrastructure for PEVs and other mobile
sources in California.  In addition, SB 350 defines how ratepayers benefit from transportation
electrification (reduced emissions, reduced impacts  to public health and the environment,
increased use of alternative fuels, renewable energy integration, and economic benefits), and
therefore can participate, through utility rates, in the funding of electrification programs.

9.2.4.6 Status of Public PEV Infrastructure Network

   The question  of infrastructure sufficiency is an important topic in regards to facilitating the
expansion of the PEV market to assist in meeting federal GHG and CAFE standards.
Specifically, how does the current infrastructure landscape and trajectory meet the needs of
                                              9-23

-------
                                           Assessment of Alternative Fuel Infrastructure
current and projected vehicle fleets and, within that fleet, what role will PEVs play in meeting
federal rules?

   When addressing this question of infrastructure sufficiency in the context of PEV adoption, it
is important to distinguish between BEVs, which inherently rely on charging infrastructure to
operate and PHEVs, which can run exclusively on gasoline and only require charging
infrastructure to operate electrically. Intuitively,  it is less likely that PHEV adoption rates are as
dependent upon robust EV infrastructure as BEVs. Given this important distinction, the question
of infrastructure sufficiency will be addressed for BEVs by examining a snapshot of current BEV
numbers in relationship to the EV landscape and trends, and comparing that relationship to work
performed by NREL for the CEC. Although the majority of PEV charging  occurs at home, data
related to the availably of home charging infrastructure (e.g., 110V outlets in home garages) is
extremely limited.  Therefore, the analysis of EV infrastructure sufficiency  is focused on public
and workplace charging.

   A recent CEC contract with NREL looked at the question of sufficiency. NREL analyzed two
potential charging scenarios -a "home dominant" charging scenario and a "high public access"
charging scenario.  Based upon these two scenarios and the composition of California's current
and projected BEV and PHEV fleet, NREL calculated that the minimum ratio of non-home
based charge points (both workplace and public) to PEVs is 0.14 per PEV in the home dominate
scenario and 0.24 per PEV in the high public access scenario.

   Applying these ratios on a national scale, infrastructure development at its current pace
appears to be  sufficient in meeting today's charging demands of BEVs. As of April 2016, a
cumulative total of over 227,000 BEV and nearly 214,000 PHEV sales were recorded
nationwide.32 Studies have shown that, on average, over 80 percent of all charging events occur
at home.  Using the home dominant NREL ratio  of 0.14 charge points per BEV, the nation would
need approximately 31,700 charge points for the current BEV fleet.  At the  end of May 2016
there were over 38,000 public and private charge points33 (i.e. connectors) nationwide.
Therefore, the existing charging network appears sufficient for the existing  BEV fleet. However,
if the PHEV fleet were added to the existing BEV fleet, the combined fleet  of 441,000 vehicles
would, under NREL's methodology, requires approximately 61,000 charge points nationwide.
While the existing workplace and public charging network falls short of that number, the existing
and forecasted sales of PHEVs demonstrate that  public infrastructure is less critical for PHEV
adoption.

   Currently, PEV  sales are a small percentage of overall light duty vehicle sales and public
charging infrastructure is sufficient to meet the current demand of BEVs in  a home dominant
charging scenario.  However, as PEV technology becomes more broadly accepted and less
expensive, and as automakers increase PEV production, infrastructure will need to continue to
keep pace with demand.  Although this development is not a guarantee, there is evidence to
suggest it will sufficiently expand. Some private electric utilities are eager to enter the PEV
infrastructure market with large investments which has the potential to significantly increase the
number of charge points. In addition, with today's relatively small PEV fleet, private companies
have established business models  to compete in the PEV infrastructure market.  As the PEV fleet
grows, those business models should become even more viable.  Using current technology, the
current number of public and private charge points may need to be expanded by nearly a factor
of 10 to provide sufficient charging capacity (as  defined by the home dominant NREL ratio of
                                             9-24

-------
                                            Assessment of Alternative Fuel Infrastructure
0.14) for the combined number of BEVs and PHEVs projected by 2025 in this Draft TAR.
However,  as section 9.2.4.2 details, there has been a nearly 70 fold increase in the number of
connectors and stations in the past five years. And, this includes PHEVs which, as noted above,
are far less likely to be as dependent on charging infrastructure. Lastly, developments such as
longer range BEVs, high power charging, and inductive charging will alter the current charging
paradigm which may lessen the ratio of public chargers per PEV, thereby decreasing the
projected charger network needs.

   The current national charging infrastructure network continues to grow with investment in
infrastructure by government, corporations, private capital markets, and electric utilities.  There
are infrastructure challenges as noted earlier (e.g., multi-unit dwellings), but they are
systematically being addressed, and infrastructure is progressing sufficiently to support the scale
of the electric vehicle market projected in this Draft TAR to be necessary to comply with the
national GHG standards.

9.2.4.7 Summary of PEV Infrastructure

   With over 16,000 (14,550 Level 2 and over 1,700 DC fast charger) public and private electric
charging stations with a total of over 38,000 connectors,34 the national PEV infrastructure
network is off to a robust start and continued strong growth is forecasted.  Although there are
remaining challenges, the initial challenges with technical specifications, communication
protocols,  and operability standards have largely been addressed. Over $250 million of private
capital has entered the infrastructure market,35 supported by emerging business cases for
charging networks. New challenges are being addressed and, as referenced herein, tremendous
opportunities in PEV infrastructure are on the horizon. Given the overall strength of the PEV
infrastructure landscape (as detailed in section 9.2.4.6), infrastructure is progressing sufficiently
to  support vehicles with PEV technology to be used in meeting the 2022-2025 national program
GHG and  CAFE standards.  However, PEV infrastructure needs are expected to be greater in
states with ZEV regulations than in states where only federal GHG and CAFE standards are
applicable.

9.3    Hydrogen Infrastructure Overview

   Hydrogen (typically in the form of a compressed gas) is the  primary fuel source for the Fuel
Cell  Electric Vehicle (FCEV). Hydrogen is abundant as a constituent of readily-available natural
resources, though it does not naturally occur in its elemental  form. In spite of this challenge,
many methods exist or are in development for its extraction from various resources, including
renewable energy sources.  The success of the FCEV as a commercial product will rely on the
development of a fueling infrastructure network that can provide that hydrogen with a retail
experience meeting the expectations of today's gasoline-fueled vehicle drivers. Significant
progress has been made towards this goal in recent years, with a network of 51  stations currently
under construction in California (a growth of 41 stations in addition to the 10 reported in the
2010 TAR36) to support the initial market. FCEVs are another  vehicle technology option that
makes use of an all-electric drivetrain, providing zero tailpipe emissions. In contrast to the plug-
in  electric vehicles discussed previously, FCEVs provide power to their electric motors by
generating the necessary electricity onboard (as opposed to receiving electricity from an external
source, through a plug).  The FCEV accomplishes this through  the electrochemical conversion of
hydrogen and air into electricity, water, and a small amount of heat.
                                              9-25

-------
                                            Assessment of Alternative Fuel Infrastructure
   Hydrogen fueling stations are designed to provide hydrogen to FCEV drivers in accordance
with design specifications of the FCEV onboard hydrogen storage tanks. Designs have evolved
over the past decade, but the prevailing on-board storage form across the industry has largely
converged on gaseous hydrogen compressed to a pressure of 70 MegaPascals (MPa). At this
pressure, hydrogen can be stored onboard in sufficient quantities to provide drivers with driving
range equivalent to typical gasoline-fueled vehicles without significant concessions in other
vehicle features in order to accommodate the storage tanks. As such, hydrogen fueling stations
are designed to dispense hydrogen at this high pressure, using fueling protocols that allow the
station to provide a complete, safe, reliable,  and accurately metered fill in a time on par with
current gasoline stations, typically around three minutes for light duty vehicles.

   Figure 9.17 provides a glimpse of the diversity in hydrogen production processes currently in
use, based on the developments in California, where many fueling stations  are in operation or
currently in development.37 The figure shows shares of production pathways for hydrogen
provided to all stations proposed in the most recent round of California's hydrogen fueling
station grant program and for California's operating and planned network, including stations
awarded in that program. The full mix shown in the 2015 network includes stations from the
research and demonstration era of hydrogen infrastructure development, which are expected to
continue to provide limited service for some time.  The differences in the shares between the full
network mix and the grant applications may be indicative of changing emphases in technology
development.  For example, electrolyzers make up a much greater portion of the 2014
applications than the full network, potentially indicating a trend for increasing participation of
this technology than was utilized in the demonstration-era stations. Similar variations in the mix
of hydrogen production technologies may be expected to continue over time as the respective
technologies develop and push the hydrogen industry to the most appropriate and cost-effective
solutions.

   The diversity of hydrogen production shown in Figure 9.17 is indicative of the latest state of
production and delivery technology and innovation in the hydrogen industry. This figure,  based
on counts of stations, shows the shares of hydrogen production methods in applications to  the
most recent round of California's grant program and the funded hydrogen fueling network in the
state.38 Stations deployed in earlier years of the network development also had smaller daily
fueling capacities on average than the newer stations. More recently, hydrogen fueling station
developers have proposed and built stations  relying on a wider array of hydrogen production
methods, with stations ranging in size from 100 kg/day up to 350 kg/day.39 Concurrently,
stations have been designed to meet more rigorous technical specifications that facilitate a retail
experience. It is expected that the hydrogen stations currently being built in California will serve
as the first examples of true retail stations with designs that can be largely reproduced or easily
modified for future expansion and establishment of regional fueling networks in other parts of
the country.
                                              9-26

-------
                                              Assessment of Alternative Fuel Infrastructure
               100.0%
               80.0%
               60.0%
               40.0%
               20.0%
                0.0%
                      California 2014 Grant
                         Applications
                                       California 2015 Network
m Pipeline


H Delivered Liquid-to-Gaseous
 NaCI03 By-product
Q Delivered Gaseous 100%
 Pipeline Biogas
D Gaseous Delivered 100% Ag
 Waste Gasifier
D Liquid Delivery SMR of NG and
 biogas
• Delivered Liquid NaCI03 By-
 Product
• Delivered Liquid SMR of NG


• Electrolyzer


• Delivered Gaseous SMR of NG


• On-SiteSMR
                      Figure 9.17 Hydrogen Production Methods in California

   Hydrogen dispensed from fueling stations to FCEVs is provided in gaseous form, but a
variety of solutions exist for storage of larger volumes at the station.  Gaseous hydrogen may be
stored in large cylinders installed at the station, often at various pressures up to or exceeding 70
MPa. Other stations store the hydrogen as a liquid, gasifying the hydrogen prior to dispensing to
a vehicle. Additionally, hydrogen may be delivered to the station from a central production
facility in either gaseous or liquid form or it may be produced on-site from methods like Steam
Methane Reformation (SMR), electrolysis (electrically-driven separation of water into hydrogen
and oxygen), or tri-generation. Tri-generation is a process utilizing a stationary fuel  cell and an
opportunity fuel like a wastewater treatment facility's digester gas to generate electricity, heat,
and hydrogen for vehicle  fueling. When hydrogen is delivered from  central production  facilities,
it may originate from a number of processes including SMR,  electrolysis, by-product from
industrial or chemical processes, biogas and biomass conversion, and other technologies
currently under development. Finally, hydrogen may be delivered via a direct pipeline link from
a major production facility. In California, this has been demonstrated at the Torrance station,
where an existing supply  line between a hydrogen production facility and an oil refinery was
accessed to divert a stream of hydrogen to the vehicle fueling station. In the future, the  source of
hydrogen provided via pipelines  could continue to serve a variety of end uses, but it is also likely
that some of the source hydrogen will be produced at central facilities specifically with the intent
of fueling FCEVs.

9.3.1   Hydrogen Network Development and Status

   FCEVs are currently envisioned to be introduced to the public fleet across the nation in a
series of releases that will coincide with development of fueling infrastructure.  In the past, the
regions where these first releases are likely to be concentrated have been referred to as network
"clusters." As FCEV and fueling infrastructure markets progress, these clusters will  be
connected by stations along major long-distance travel corridors, and smaller secondary clusters
will be established as the  demand and capability to fuel FCEVs spreads beyond the initial cluster
                                                9-27

-------
                                            Assessment of Alternative Fuel Infrastructure
areas. This strategy has begun to be exhibited in California, where the California Fuel Cell
Partnership (CaFCP) explicitly detailed such a strategy, focusing on five early adopter clusters
(two in the northern and three in the southern portion of the state), with various smaller clusters,
connector, and destination stations developing around the state.40 This strategy has been adopted
in past State funding programs, and newer analyses and programs continue to identify the need
for stations in some of the very same regions identified by the cluster paradigm.

   Similar development strategies are likely to be carried out in other areas of the nation where
there will be a high early adopter market demand for FCEVs.  In most cases, these high demand
areas will be in or near major urban areas, with other clusters developing as the demand spreads
outward from these focal regions.  Thus, the network of nationwide stations will likely develop
in smaller regions, established primarily to  support the daily needs of the first adopter FCEV
market. Connector stations will then link these major clusters and establish travel corridors for
further development. As these first clusters grow and spread to become interconnected with a
widening market for FCEVs, they will become more regional in scale and provide service
coverage to increasing portions of the nation's population. During this development, these
networks will be connected by long-distance connector stations, allowing for inter-regional and
nationwide travel via FCEV with ample opportunity for fueling.

  Figure 9.17 shows the current status of development for the hydrogen fueling network in
California.41 An early semblance of the clustering paradigm is visible in the stations located in
Los Angeles, Orange County, and around the San Francisco Bay in the northern part of the state.
The station shown in Coalinga will  serve as a connector enabling travel between the clusters in
the northern and southern halves of the states. Meanwhile, destination stations will be in place in
areas like Truckee and Santa Barbara to support vacation travel for FCEV drivers. The
California Air Resources Board (ARB) estimated that the 51  stations in operation or
development in 2015 (50 are shown in the map; a recent relocation has resulted in a  station
project converting into an upgrade for a legacy station) will be able to provide sufficient fueling
capability  for approximately  12,000 -  15,000 FCEVs.42 Assuming no decrease in State funding,
ARB also  estimated that a total of 86 stations could be built by 2021 and 100 by 2023.  In
December 2015, a more nuanced projection accounting for potential reductions in station costs
projected that 100 stations could be built by 2020, as long as the FCEV introduction rate was at
least as fast as the ARB estimate.43  If the introduction of FCEVs were to be delayed for 4 years,
then station rollout would correspondingly  decelerate, and 100 stations would not be built until
2024.44  However, in all analyses thus far performed by the State of California, the demand for
hydrogen to fuel FCEVs is projected to exceed the dispensing capacity if station deployment is
limited only to the AB 8-related funds  currently in use. The State's estimate for the timing of
insufficient dispensing capacity  depends on the assumed scenario and ranges from 2019 to 2026.
In response, State agencies have initiated dialog on addressing this potential  shortfall by working
with stakeholders to demonstrate the market opportunity and increase the magnitude of private
investment in the state's hydrogen infrastructure network.
                                              9-28

-------
                                                Assessment of Alternative Fuel Infrastructure
                       Emeryville
                                    San
                                    Ramon
                  South San
                  Francisco
 Hayward
                    Woodside
                             Palo
                             Alto
                                    Fremont
                                Mountain
                                      »San Jose
                           Altos
                               •
            III
                      Rohnert
                       Park
                                                            Woodland u N°rtn  . La Canada-Flintridge
                                                            Hills     Hollywood  Burbank*
        West
  Santa LA #2'
 Monica •
        Culver
         City
                                                West
                                                Hollywood
                                              I Beverly
                                              Hills
                                                                                       Pasadena
                                                                                       CSULA
                                             LAX

                                              Lawndale
                                                Jorrance Long

                                           Palos  Harbor  .Beach
                                          Verdes CitV
   Truckee
West
Sacramento
                                                                  Diamond
                                                                    Bar
                                            Chino
                                                                                Ontario
                                                                                        Riverside
                                                                    Anaheim

                                                                    Orange
Costa           Lake
Mesa ,UC lrvine .Forest
    • Newport
               San Juan
               Capistrano
                                 Coalinga
                                    Santa  Santa
                                   Barbara ciarita1
                                   Thousand
                                       Oaks
                                           Encinitas
                                                   San
                                                 %Diego
 Figure 9.18  Locations of California's Funded and Operational Network of 50 Hydrogen Fueling Stations45

   It is the goal of the State's financial support to set in place enough stations that customers
have sufficiently convenient access to stations  and sufficient confidence in availability of
hydrogen fueling locations to decide to purchase or lease a FCEV instead of a traditional
combustion-powered vehicle.  Given the early  state of both the FCEV and infrastructure markets,
the financial incentive is meant to increase the  financial viability of the earliest stations, when the
                                                  9-29

-------
                                            Assessment of Alternative Fuel Infrastructure
risk is greatest and the fueling throughput lowest. This will help keep more stations open longer,
allowing auto manufacturers time to develop the FCEV technology and introduce vehicles that
meet all of a retail customer's expectations, as is currently beginning to happen.  Eventually, as
both the FCEV and fueling station industries mature, the stations will be financially self-
sufficient and more attractive to private investment; at this point, the market opportunities are
expected to dictate the rate of growth of both markets and the State will be able to reduce its
financial participation.

   Following the sales trajectories of Figure 5-46 in Chapter 5, approximately 125,000 FCEVs
may be expected on California's roads by 2025, with approximately 37,000 new vehicle sales in
that year.  Effectively, the annual rate of installation of new fueling capacity by 2025 needs to be
greater than the full amount of capacity included in the first 100+ stations expected to be funded
by the State by through AB 8. This should signal a significant market opportunity to station
developers and private financers with an interest in hydrogen and FCEVs.  However, individual
assessments of risk and market opportunity will play a predominant role in determining how
rapidly the need will be met, as State funding is not expected to play as significant of a role in
2025 as it does now.  Further discussion of the costs and financial evaluations of deploying
hydrogen fueling infrastructure is included in section 9.3.4.

   Outside of California, the Alternative Fuels Data Center, maintained by the United States
Department of Energy (U.S. DOE), indicates that two other stations are currently operational
(one in Connecticut and one in South Carolina) and one station is in development in
Massachusetts.46 Additionally, U.S. DOE opened a 70 MPa station in Golden, Colorado for use
in research studies47 coinciding the event with National Hydrogen and Fuel Cell Day on October
08,2015.

   In addition to these stations, a number of activities across the nation are currently or soon will
be underway to establish and increase coverage provided by hydrogen fueling stations in the
expected early adopter markets. Connecticut is currently seeking applications for grant funding
of up to two stations in the Hartford area.48 Air Liquide and Toyota have announced a
partnership to establish a dozen fueling stations in the northeast states.49 Finally, H2USA (a
public-private  partnership established by U.S. DOE to address the challenges of establishing the
FCEV market  in the USA) is developing a plan for fueling station development across the
northeast, emphasizing fleet vehicles as the first market, with the intent of expanding into a retail
consumer-centric network model.50

9.3.2   Retail Experience

   Until very recently, many of the existing and funded hydrogen fueling stations have been
largely demonstration and/or research stations. These stations have been critical in providing
insights for station design, construction, and operation while still providing essential fueling
service to pre-commercial FCEV drivers. However, as fully commercial launches of FCEV
models have now begun (e.g., the Hyundai Tucson Fuel Cell and Toyota Mirai) and more are
planned for the near future (e.g., the Honda Clarity Fuel Cell expected in 2016),  the stations will
need to provide fueling service to a wider, more retail-oriented user base.  Over the past few
years, and often times directly as a result of experiences gained at the earlier demonstration
stations, new protocols and standards have been developed that will ensure future FCEV drivers
have consistent, reliable, retail-like experiences when filling their vehicles. Hydrogen quality
standards (SAE J271951) and dispenser fueling protocols (SAE J260152) are examples of recent
                                              9-30

-------
                                            Assessment of Alternative Fuel Infrastructure
advancements that will shape industry-wide development and implementation of fueling station
equipment and ultimately provide consistent and reliable fueling experiences to FCEV drivers.

   Additionally, a number of state and national efforts have and will provide tools that can
ensure stations adhere to these and other standards.  In California, the Hydrogen Field Standard
(FTPS), developed by NREL under contract with the California Division of Measurement
Standards and funded by the California Energy Commission, has allowed for the certification of
dispensers' metering accuracy. HFS was developed based on a need for a discrete method to
verify that dispensers could measure and dispense hydrogen accurately. The California
Department of Food and Agriculture's (CDF A) Division of Measurement Standards (DMS) has
jurisdiction over the retail  sale of motor vehicle fuels and has adopted by reference the methods
for sale and accuracy standards contained in National Institute of Standards  and Technology
(NIST) Handbooks 4453 and 130.54 NIST set the national hydrogen accuracy standard at 1.5
percent acceptance and 2 percent for in-use or maintenance tolerance.

   Workshops and early field testing indicated the 1.5 percent/2 percent NIST standards were
technologically infeasible with existing metering technology, so CDFA adopted temporary  tiered
accuracy classes of 3 percent, 5 percent, and 10 percent55.  This approach allowed the near-term
retail sale of hydrogen to consumers and provided time for industry to improve dispenser
metering methods.  In the past year, several  dispensers have been tested and certified using  the
HFS, including the world's first dispenser certified to be accurate enough to sell hydrogen to the
consumer by the kilogram at the station located on the Los Angeles campus of the California
State University.  Future station designs that incorporate type-certified dispensers will require
less-intensive accuracy testing during the commissioning process.

   The Hydrogen Fueling Infrastructure Research and Station Technology (H2FIRST) project
has developed the Hydrogen Station Equipment Performance (HyStEP) device.  The device is
designed to carry out various certification tests outlined in the CSA Hydrogen Gas Vehicle
(HGV) 4.3, 2015.56 These tests will be able to certify that a station's dispenser is capable of
providing safe, fast, and repeatable fills according to the protocols defined in SAE J2601. The
device has been validated in a research setting at NREL in Colorado and at retail stations in
California; it is now being used to perform validation testing of the operational stations in
California's fueling network. There, it will be used to test stations currently in service and
newly-constructed stations as they are completed. The device is trailer-mounted and has been
purposely designed with the intent of traveling not only within the state of California, but across
the nation as stations and networks are developed in other regions.

   With these devices, and others currently under consideration or development, state and
national stakeholders are gaining the capability to provide increasing confidence to  consumers
that their fueling experience will be safe, reliable, and consistent.  At the same time, industry
stakeholders have recently placed considerable effort into precisely defining additional features
to enhance the customer experience and allow a station to be considered "retail." For example,
many demonstration stations were placed behind card-key locks and thus  not freely accessible to
any public driver in the vicinity.  Additionally, given that a legal means was not yet in place to
sell hydrogen directly to consumers, stations did not have a Point-of-Sale system and customer
payment was managed through access agreements as opposed to the on-demand purchase
enabled by cash, debit, and credit card sales typical  of today's gasoline stations.  With the
                                              9-31

-------
                                            Assessment of Alternative Fuel Infrastructure
deployment of commercial vehicles, vehicle manufacturers expect that truly retail public stations
will not limit the customer base of their service.

9.3.3   Hydrogen Fueling Station Capacity

   Given the  limited number of FCEVs currently on the road and the demonstration nature of
many of the stations built to date, most hydrogen fueling installations have been designed with a
smaller capacity than is anticipated to become the norm in the future.  In 2014, ARB compared
the composition of the existing and funded hydrogen fueling network in California to the state's
gasoline fueling network in terms  of capacity. ARB reported that the state's gasoline fueling
infrastructure was comprised of very different types of fueling stations; the top 1  percent of
stations (in terms  of volume of gasoline sold) were typically seven times as large as the average
station in the  state.  In addition, over 50 percent of the gasoline was sold by only  the top 21
percent of stations.57 Thus, the  gasoline fueling infrastructure contains a large number of
comparatively small stations and a small number of very large stations.

   Thus far, the hydrogen  infrastructure development has not been as heterogeneous.  This is
partially due to the early development  stage; all these stations have served a similar
demonstration and pre-commercial market purpose. In the case of gasoline stations, the progress
of development has led to  station designs that are more tailored to different roles  within the
network (such as connector, destination, etc.). Over half of the hydrogen fueling stations built
and planned have been designed with a capacity in the range of 150 to 200 kg/day, with the
largest stations designed for 350 kg/day. The average for the state currently stands at 180
kg/day, also the most common design capacity in the state. Thus far, station capacities are
mainly a function of the hydrogen source; the composition of California's network is: 31 180
kg/day gaseous delivery (or combination of on-site  production and gaseous delivery)  stations, 7
350 kg/day liquid delivery stations, and 8 on-site electrolysis stations ranging from  100 to 130
kg/day.

   From its comparison to gasoline infrastructure, ARB concluded that hydrogen fueling stations
would not only need to grow larger in design capacity,  but also become more  diversified and
specially-designed for various network roles.  While today's gasoline stations provide on average
24 times the fueling capacity of a hydrogen station on an energy basis, the largest 1 percent of
gasoline stations can provide 80 times  as much energy per day as the largest hydrogen station
designs. As a result, hydrogen stations with the highest capacities in the network will need to
show the greatest  growth in order to provide the same magnitude of service as the largest
gasoline stations.  This growth in capacity will likely be a  smooth transition over time, requiring
careful balancing  of the financial constraints of greater capital investment, potential for greater
revenue due to greater throughput, and coordination with the timing of FCEV rollouts and sales.

9.3.4   Hydrogen Fueling Station Costs

   Hydrogen  fueling infrastructure is currently in a  period of transition from research and
demonstration to full retail and commercial market  development. This transition period has
meant that hydrogen fueling stations built in prior years have largely been hand-constructed,
individually designed stations.   Conversely, newer stations currently being constructed in
California and other parts  of the nation (and the world) are becoming increasingly standardized
in their design. Given this  transitional period, there  is currently a degree of uncertainty in the
likely costs to build and operate a  hydrogen fueling station in a fully-developed, retail-service
                                              9-32

-------
                                            Assessment of Alternative Fuel Infrastructure
FCEV and hydrogen fueling market.  However, as part of its infrastructure development
program, California recently released the first of its annual reports that evaluate the costs and
timing of building the currently-funded hydrogen fueling stations and the expectations for
stations to be funded up to the goal of at least 100 stations.  The 2015 Joint Agency Staff Report
on Assembly Bill 8: Assessment of Time and Cost Needed to Attain 100 Hydrogen Refueling
Stations in California58 discussed the costs and construction timelines observed over the course
of California's experience with installing hydrogen fueling stations.  In particular, the report
assessed the costs for three station types that are representative of the majority of the stations
currently in California's planned network and are expected to continue to play major roles in the
ongoing network development.

  Based on grant funding applications and follow-up interviews with awardees,  estimates of
costs for the currently funded station network in California were developed for the Joint
Assessment.59 Representative values for three common station designs are summarized in Table
9.4, based on the information provided in the report. Note that some of the values reported are
estimates generated for stations still in construction, and some underlying cost values may be
based on one or a few stations and all stations are being developed in the early years of network
development. As in Figure 9.19, these costs  may decline over the coming years.  As the State
continues to co-fund stations and learn more  about the development costs, estimates and trends
will likely become more precise and predictable.
                     Table 9.4 Representative Hydrogen Fueling Station Costs60
Hydrogen Source
Delivered Gaseous
Delivered Liquid
On-Site Electrolysis
Capacity
(kg/day)
180
350
100
Total Capital Costs
($ Million)
2.01
2.80
3.21
Equipment Costs
($ Million)
1.60
1.93
2.38
Construction Costs
($ Million)
0.28
0.60
0.46
Other Costs*
($ Million)
0.13
0.27
0.37
*OtherCosts include Engineering and Design, Permitting, Commissioning, and Project Management and Overhead


   Based on information available from station grant funding applications, invoices, and follow-
up meetings with station developers, the 2015 cost assessment estimated the current costs of
development for each of these stations.  It is important to note that any such estimation is only
representative; many variable costs included in the overall estimate may significantly alter the
assessment for an individual station.  These costs are "all-in" capital costs, including
engineering, permitting, equipment procurement, construction, commissioning, and other factors.
These costs do not include operations and maintenance costs, which would include the cost to
procure hydrogen, rent, variable electrical and potentially natural gas energy costs, and others.

   These costs are representative of today's technology, the relatively small number of stations
in development (compared to expectations for the future), and the still-developing supply chain
for manufacture of the equipment. In future years, as the rollout of FCEVs progress, larger
numbers of stations (and likely of larger rated capacity) will be needed and it is expected that
continued development of the equipment technology and the material  supply chains will enable
decreasing capital costs  on an individual station basis, as shown in Figure 9.19. Economies of
scale suggest larger reductions are possible for larger capacity stations. Note that although
shown in Figure 9.19, no retail station currently funded or in operation in the United States has a
capacity above 400 kg/day; stations with the larger capacities are expected to become more
                                              9-33

-------
                                              Assessment of Alternative Fuel Infrastructure
favorable and common as the volumes of vehicles on the road significantly increase in later
years.
          $5.0
             2014
2016
2018
2020
2022
2024
2026
2028
      D LH2 Truck (Awards)

      O GH2 Truck (Awards)

      A Electrolysis (Awards)
             •LH2 Truck HSCC 350 kg/day

           -O-GH2 Truck HSCC 180 kg/day

           -•^-Electrolysis HSCC100 kg/day
                              	LH2 Truck Future 300 kg/day (HRSAM)

                              	GH2 Truck 200 kg/day (HRSAM)

                              -0-Large HSCC 600 kg/day
            Figure 9.19 Projections for Cost Reductions in Hydrogen Fueling Infrastructure61
   Operations costs also play a major role in the overall financial viability of hydrogen fueling
stations. Especially during the early period of the FCEV market launch, the operations costs can
actually be the dominating concern for a station's viability. Small numbers of vehicles translate
to low utilization of the station and restricted hydrogen sales revenue, which provides the means
of paying for variable operating costs and amortized capital costs. In addition to the uncertainty
in the near-term demand for hydrogen fueling, there is also uncertainty in the price to procure
hydrogen and the eventual market price that can be charged for hydrogen.  Currently, hydrogen
is most often sold to other industries in much larger quantities and at much lower pressures than
are needed at today's FCEV fueling stations.  This "merchant hydrogen" price is thus not likely
representative of the price that hydrogen fueling station operators will need to pay.  In the 2015
Assessment,62 estimates of delivered hydrogen cost to the stations varied from $8.91/kg in 2015
to $7.64/kg in 2025; retail sale price to the consumer was estimated to decrease from $14/kg in
                                                9-34

-------
                                                Assessment of Alternative Fuel Infrastructure
2015 to $11.1 I/kg in 2025, all based on prior work and information from current stations
developers.  A sample of the assessment and major financial indicators for the gaseous delivery
station is included in Figure 9.20.  In the figure, the cost of delivered hydrogen is clearly a major
portion of the overall costs to the station (note that cash flows are shown in levelized terms).
Also note the importance of the capital and production incentives (as modeled through the AB 8
program); the analysis found that with current technology and vehicle  deployment projections,
these incentives play a major role in the financial viability of the early station network.
Station specification
Station type
Station capacity (kg/day)
Vintage year
SERA Scenario
Station cost (all-in)
Station capital incentives
O&M incentive (for 3 years)

Delivered gas
180
2018
Expected
$ 1,561,000
$ 1,249,000
$ 100,000
     Financial performance benchmark
       IRR
       Profitability index
       NPV after-tax, nominal @ 10% discount
       First year of positive E8ITD
       Real levelized retail margin ($/kg)
       Price of H2 to yield 10% IRR ($/kg)

     Cumulative investor cash flow, (Thousands)

29.6%
count S

$
) $
1.7
116,000
4
3.49
9.52
    400
    300
    200
    100

    100
    200
    300
s s s
a m -J
S 2 »
Real levelized cash flows ($/kgH2)
 Sales revenue        _^^_^_
 Capital incentive       __ 1.52
 Monetized tax deductions _ 0.99
 Inflow of equity        0.42
 Production incentives     0.34
 LCFS credit           0.28
 Incurrence of debt       0.13
Total cash inflow

 Delivered hydrogen
 Maintenance expense    ^0 3.14
 Equipment cost        I*-90
 Rent             | 0.71
 Dividend payments     | 0.58
 Credit card fees       I 0.25
 Sales tax           | 0.22
 Cost of electricity      | 0.21
 Interest expense       I 0.13
 Repayment of debt     I 0.09
 Property insurance      I 0.06
 Selling & administrative   I 0.05
 Licensing & permitting    | 0.02
 Reserve for cash on hand  I 0.01
 Installation expenditure
 Taxes payable
 Labor expense
 Cost of natural gas
Total cash outflow       ^^^^^^1
                                                                                            13.6
       ssssssssssssssssssssssssss
                Financing contributions
                 • Equity investment $390K
                 • Issued debt $104K
                 • Capital incentive $1,249K
                  Operating incentives S300K
   Figure 9.20 Sample Financial Evaluation of a 180 Kg/Day Delivered Gaseous Hydrogen Fueling Station
                                 Based on Experience in California63
9.3.5   Paradigms for Developing Networks

   While there is broad acceptance of some form of the cluster-connector-destination style of
fueling station placement and planning, significant and varied work has been targeted towards
the specific implementation of the strategy and translating the general concept into a plan that
can be implemented by state and local agencies.  One of the earliest examples is STREET
(Spatially and Temporally Resolved Energy and Environment Tool) developed by the Advanced
Power and Energy Program at the University of California, Irvine.64 This tool represented an
innovation in providing detailed spatial resolution in pinpointing ideal locations for hydrogen
fueling stations, based on projections of geographic distribution of the early adopter market. A
fundamental function of the STREET model is to determine the appropriate number and location
of stations to provide localized service coverage equivalent to the national average of coverage
provided by gasoline stations,  thereby providing the same measure of convenience to the driver.
The tool was instrumental in the development of a roadmap to meeting the CaFCP-defined
                                                   9-35

-------
                                           Assessment of Alternative Fuel Infrastructure
clusters and served the State of California for a number of years to help quantify the desirability
of proposed station locations. With STREET and stakeholder discussions led by the CaFCP, it
was determined that 68 stations would be necessary to launch the first adopter FCEV market.65

  More recently ARB has developed the California Hydrogen Infrastructure Tool (CHIT),
which shares some fundamental features with STREET and other models.  At its core, CHIT
identifies the areas with a large early market potential for FCEVs and compares this to an
estimation of the coverage provided by existing and funded stations.66 CHIT has been designed
as a tool that allows ARB and CEC to annually identify the areas with the greatest need for
additional station coverage, and emphasizes infrastructure planning rather than optimization. By
determining areas of greatest need for additional station coverage, CHIT provides a basis for
structuring State infrastructure funding programs while also allowing flexibility for station
developers to build proposals with more finely detailed information for specific sites that could
meet the identified early adopter market needs.  Thus, it fills a need for evaluation in grant
funding programs, as opposed to the optimization scheme that takes a central role in STREET.
Figure 9.2land Figure 9.22 show STREET and CHIT's coverage assessment outputs.
             STREET:
             Optimization of
             Station
             Placement
             along Major
             Road
             Intersections in
             Early Markets
                                   STREET: Optimized Configuration in Irvine, CA
0 Hydrogen refueling stations
^ Interstates & freeways
— Principal arterial roads
Travel Times;
• 2 mm
   mm
  4 min
                       Figure 9.21 Optimization of Coverage in STREET
                                             9-36

-------
                                            Assessment of Alternative Fuel Infrastructure

CHIT:
California-
Focused
Prioritization
based on
Assessment of
Gaps Between
Existing
Coverage and
Early Markets

CHIT: Coverage Estimation

Valley ^
Emeryville San
* ^ "3 R
jr -«^^ Q amon
**Francisco^ oHayward^
o
fvL- "~^V^Hl
Relative |^_ ^fc«*t_j!. ^
Coverage Fremont
• H'9h paid °
WoOdside (^Mountain
0 LosuView o53"^086
x Altos® Campbell
SaratogaOo
• __-^^^P^
***^
Low
                    Figure 9.22 Multiple-Station Coverage Estimation in CHIT
   On the national scale, the National Renewable Energy Laboratory has been developing the
Scenario Evaluation, Regionalization, and Analysis (SERA) tool to study likely scenarios of
hydrogen infrastructure development and deployment.67 The tool incorporates findings and
direct functionalities from a number of other hydrogen fueling station and vehicle choice models
from various DOE efforts in order to provide a full-spectrum analysis of potential nationwide
growth in FCEV adoption and complementary fueling station establishment.  Among other
factors, the model emphasizes the assessment of an Early Adopter Metric in determining the
order and magnitude of development of hydrogen fueling stations in Urbanized Areas (as defined
by the U.S. Census Bureau), and more recently incorporates an analysis to determine timing and
placement of connector stations between regional clusters once they reach critical size(s). The
consideration given to the model's various factors is flexible, allowing researchers to assess
scenarios that emphasize  proximity to early adopter markets, proximity to established fueling
infrastructure, the strength of incentive programs, or other fundamental considerations.  In
addition to the scheduling, siting, and capacity specification capabilities, the model also  provides
a means for assessing or optimizing the financial case for individual stations and the network.
Figure 9.23 shows projections of network size and phase-in date from analysis of a scenario with
successful launch of FCEVs nationwide.
                                              9-37

-------
                                            Assessment of Alternative Fuel Infrastructure
             SERA: National
             Urbanized Area
             Siting, Timing,
             and Sizing for
             FCEV Early
             Market
             Scenarios
                                    SERA: Networks in National Success ofFCEVs
                                  ntroductionMap -National Expansion
                      • • <  '     .
                   . .  » -.V>V$V  :.;.

                   -V
-------
                                            Assessment of Alternative Fuel Infrastructure
emphasize demographic information, known vehicle fleet locations, and transportation data to
plan station locations to meet primarily fleet-based needs.

   A number of other paradigms for building a robust hydrogen infrastructure network have also
provided valuable insights for planning and anticipating the regional and national need.
Modeling has been completed to demonstrate the benefit to network flexibility that could be
afforded by a larger number of mobile refuelers with the ability to act as dispatch able temporary
or semi-permanent stations.69'70 Other work emphasizes the importance of local demographics
and traffic patterns, especially near major highway access points, for fine-tuned optimization of
station placement.71 Origin-Destination studies further analyze traffic patterns, seeking to take
advantage of data available from case studies of local drivers' travel routes in order to find
station locations optimized by their proximity to the major travel routes, rather than to the
homes, of the early market adopters.72'73'74  Other examples of station placement planning have
been detailed in the literature, but an exhaustive review is beyond the scope of this report and the
concepts presented here have had demonstrable effects on considerations in current and past
fueling station planning.

9.3.6  Challenges and Opportunities for Hydrogen Fueling Stations

   While development of hydrogen fueling networks has been concentrated in California, future
expansion is anticipated in other early market areas of the nation. As previously mentioned, the
northeast states currently have multiple efforts underway including a grant program in
Connecticut,  development of a multi-state regional plan, and anticipated station development
through private partnerships (Toyota-Air Liquide). As these and other networks become
established and continue to grow, the development of local, statewide, and regional hydrogen
fueling networks is likely to evolve to meet the changing needs of the network.

   In particular, it is anticipated that the paradigm for locating new stations will shift from
providing maximum coverage to the early adopters to providing maximum fueling capacity for
broadening markets. A number of factors will motivate this shift. The first is that economies of
scale are expected to dictate that larger stations will provide more favorable business cases to the
station developers and operators. With the revenue gains afforded by larger throughput, station
operators will be able to capture shorter payback periods and will likely be able to provide
hydrogen at lower retail prices than with a  smaller station. Additionally, the increasing volume
of FCEV production and sales will necessitate greater capacities of hydrogen fueling in the
future. If the network cannot serve the projected growing numbers of vehicles, there will be a
risk that vehicle introduction rates will be curbed in order to avoid stressing the network and
diminishing the customer experience.  However, the timing and implementation of a transition to
larger capacity stations will need to be carefully gauged; larger capacity stations will individually
require greater capital  investment. The result would be fewer stations built with an equivalent
investment, potentially limiting the effectiveness of that investment to provide increased
coverage. This transition from coverage-focused to capacity-focused investment is not expected
to be abrupt; instead, a smooth transition that balances capacity and coverage appropriately will
likely lead to a more successful network.

   In addition to station capacity, it is expected that stations will continue to become more
technically capable. In California and Connecticut, performance requirements such as  back-to-
back fills capability have been specified. This ensures customers will not need to wait for station
equipment to be ready to fuel their vehicle when they arrive during the busiest, peak traffic times
                                              9-39

-------
                                            Assessment of Alternative Fuel Infrastructure
of the day.  The number of back-to-back fills and the speed of the fills are expected to be refined
over the coming years and each is likely to improve as technology progresses and increased
demand warrants more stringent performance standards.  Part of this development will also be
the move from single-hose designs that are currently the norm to stations with multiple
dispensers that can fuel multiple vehicles simultaneously. Projecting further out, stations will
also progress from the current model of co-location of individual hydrogen fueling islands on
gasoline station property to development of stations fully dedicated to hydrogen.

  Nearly all hydrogen fueling stations developed to date have received financial assistance via
some form of government grant funding.  This has been a necessary step in order to accelerate
the technological and market development of FCEVs and hydrogen fueling. The aim of the
government programs that supply these funds (like Assembly Bill 8 in California and EV
Connecticut) is to provide support to a developing industry and relieve some of the initial
investment risk. However, it is also the goal that the supported industry will become self-
sufficient and see real returns on investment without government assistance within a reasonable
timeframe.  As a benchmark, AB 8 in California has set 100 State-funded stations as an
evaluation point for determining whether the FCEV and hydrogen market is self-sufficient. As
fueling networks continue to develop, there will be an expectation that the business cases for new
stations will continue to improve through reduced risk, reduced costs,  and increased revenue
provided by a growing customer base.  Once early uncertainties and risks are overcome, new
stations will be able to be built with increasing proportions of private funds.

  The progress to-date in California and planned for the Northeast is working to ensure that
sufficient fueling infrastructure exists to support the needs of the early FCEV market. By
catalyzing this early fueling network development, government and private industry are making
the necessary developments to allow FCEVs to enter the retail commercial market and have
success in widespread consumer adoption. The current networks and planning target this
specific near-term need, but these developments are crucial for  establishing the FCEV market's
potential as a major aspect of achieving greenhouse gas emission reduction goals. The success
of these efforts will enable a national expansion of the FCEV market, fulfilling expectations of
the future role of the vehicles in the nation's fleet.

  For the stations that have been built to date, implementation  of renewable hydrogen sourcing
has posed a financial  challenge.  Recognizing the increased cost of generating hydrogen through
entirely renewable methods  (such as solar and/or wind-powered electrolysis, reformation of
biogas, and conversion or biomass), the California Energy Commission has previously provided
greater funding incentive for stations that demonstrate a 100 percent renewable fuel pathway. It
has long been a vision for the industry that renewable generation methods become less
expensive, enabling the economic viability of a hydrogen infrastructure network that will be
supplied by increasing volumes of renewably-sourced hydrogen.

  There exists a particularly notable potential to accelerate this industry development by
implementation of the power-to-gas paradigm. In this type of system, hydrogen plays a central
role as an energy carrier, providing energy storage for renewable electricity that would otherwise
be curtailed at times of low demand. The renewably-produced hydrogen can then be integrated
with local and regional natural gas pipeline systems potentially  for enrichment of the energy
content of the gas  or for long-distance transportation of hydrogen, and providing fuel to FCEVs.
This integrated approach is currently being researched by a number of organizations worldwide
                                              9-40

-------
                                            Assessment of Alternative Fuel Infrastructure
and efforts are underway to demonstrate the viability of the concept at real-world scales of
energy demand.  For example, there are approximately 30 power to gas projects in Germany in
various stages of planning and operation.75

9.4    Fueling Infrastructure for Other Alternative Fuel Vehicles

   As discussed in Chapters 12 and 13, and consistent with the agencies' projections in the 2012
final rule, the agencies project that OEMs will be able to comply with the standards without
large-scale development and commercialization of alternative fuel vehicle technologies.  While
the discussion above focused on infrastructure for electric and hydrogen-fueled vehicles, which
can achieve significant reduction in GHGs, it is also possible that vehicle manufacturers will
continue to market some light-duty vehicles using alternative fuels other than electricity and
hydrogen in the U.S. There are already a large number of flex fuel vehicles (FFVs), capable of
fueling on either gasoline or ethanol (E85), in the marketplace. In addition, there is existing
infrastructure capable of delivering blends of conventional fuels and biofuels; this existing
capacity is being enhanced by investment in additional capacity, including through investment
by USD A, matched  by state and private sector investment. It is also possible that there may
continue to be gradual growth in the numbers of natural gas vehicles, primarily compressed
natural gas (CNG) vehicles, into the foreseeable future if favorable market conditions continue.

   To the extent that some manufacturers produce alternative fueled  vehicles in the coming
years, sufficient fueling infrastructure will continue to be needed for purchasers of those
vehicles. For the two largest alternative fuel vehicle segments, CNG and E85, fueling
infrastructure has continued to grow to support vehicle fleet growth.  Numbers of CNG stations
have continued to rebound from a decline during the recent recession years, increasing each year
since 2009 and reaching an all-time high of over 1,600 stations currently, over 900 of which are
available to the public.  (The remainder of current CNG stations provide fuel to dedicated fleets
of vehicles, usually heavy-duty vehicles, and  are not available for fueling light-duty CNG
vehicles). The number of gasoline stations that provide E85 has increased from under 800
stations in 2006 to over 3,100 stations today, over 2,800 of which are available to the public.76
Also, the U.S. Department of Agriculture's Biofuels Infrastructure Partnership now underway
could increase the number of E85 stations.77

9.5    Summary  of Alternative Fuel Infrastructure

   In aggregate, the  status of alternative fuel infrastructure could be characterized as sufficient,
growing, or robust.  Moreover, the agencies' initial assessment for this Draft TAR is that the
MY2022-2025 standards can be met largely through continued advancements in gasoline vehicle
technologies, with the only alternative fuels needed to meet the MY2022-2025 standards being a
very small fraction of PEVs (see Chapters 12 and  13). As a result, infrastructure does not
present a barrier for  alternative fuel vehicles to be used in meeting the 2022-2025 national
program  GHG and CAFE standards. Of course, the agencies recognize that, apart from the
standards, auto manufacturers may decide to pursue alternative fueled vehicles for other reasons,
such as market demand.

   Although the majority  of PEV charging occurs at home and home-based charging is an option
for many PEV drivers, national PEV infrastructure in public and work locations is progressing
appropriately.  With over 12,000 public and private stations and over 38,000 connectors, public
charging needs are being addressed, additional public charge stations are opening weekly, and
                                              9-41

-------
                                             Assessment of Alternative Fuel Infrastructure
strong growth is forecasted.  With vehicle grid integration, inductive charging, and vehicle to
grid bi-direction power flow, tremendous opportunities in PEV infrastructure are on the horizon.
These opportunities coupled with a growing PEV market will further the commercial
infrastructure market and ultimately the availability of PEV infrastructure.

   The preceding section discusses existing infrastructure and trends for ethanol (E85) and
natural gas.
                                               9-42

-------
                                                   Assessment of Alternative Fuel Infrastructure
   References
1 77 FR 62784.
2 Cleaner Cars from Cradle to Grave. Union of Concerned Scientists, November 2015.
3 The EV Project Idaho National Laboratories, 2014.
4 Costs Associated with Non-Residential EVSE, U.S. DOE, November 2015.
5 Costs Associated with Non-Residential EVSE, U.S. DOE, November 2015.
6 What Kind of Charging Infrastructure Did Nissan Leaf Drivers in The EV Project Use and When Did They Use
It?, The EV Project, INL, 2014.
7 California Statewide Charging Assessment Model for Plug-in Electric Vehicles: Learning from Statewide Travel
Surveys, ITS UC Davis, 2013.
8 Work place Charging Challenge Mid Program Review Webinar. U.S. DOE, December 17, 2015.
9 Electric Vehicle Supply Equipment Installed Cost Analysis. EPRI2013.
10 Electric Vehicle Supply Equipment Installed Cost Analysis. EPRI2013.
11 Electric Vehicle Supply Equipment Installed Cost Analysis. EPRI 2013.
12 Costs Associated with Non-Residential EVSE, U.S. DOE, November 2015.
13 Electric Vehicle Market Forecasts Global Forecasts for Light Duty Hybrid, Plug-In Hybrid, and Battery Electric
Vehicle Sales and Vehicles in Use: 2015-2024, Navigant Research, Quarter 4, 2015.
14 Electric Vehicle Charging Services, Level 1, Level 2, DC Fast Charging, and Wireless Charging for Residential
and Commercial Applications: Global Market Analysis and Forecasts, Navigant Research, Quarter 2 2016.
15 Electric Vehicle Charging Services, Level 1, Level 2, DC Fast Charging, and Wireless Charging for Residential
and Commercial Applications: Global Market Analysis and Forecasts, Navigant Research, Quarter 2 2016.
16 California Senate Bill 454. (Corbett) 2013-2014.
17 Alternative Fuels Data Center. U.S. Department of Energy.
18 Combined,  Tesla Motors, ChargePoint, eVgo, and Blink own, operate or network 8054 of the 14,355 Level 2 and
DC Fastcharge stations. From AFDC accessed May 4, 2016
http://www.afdc.energy.gov/fuels/electricitv  locations.html.
19 For more information see AFDC database located at http://www.afdc.energv.gov/fuels/electricitv  locations.html.
20 For more information see AFDC database located at http://www.afdc.energv.gov/fuels/electricitv  locations.html.
21 GFO-15-601 - DC Fast Chargers for California's North-South Corridor. California Energy  Commission, 2015.
22 http://www.transportationandclimate.org/content/northeast-electric-vehicle-network.
23 Alternative Fuels Data Center. U.S. Department of Energy.
24 Electric Vehicle Community Readiness Projects. U.S. DOE Clean Cities - Financial Opportunities Web-site,
2015.
25 Plug-in Electric Vehicle Charging Infrastructure Guidelines for Multi-unit Dwellings. California PEV.
26 Quick Facts: Resident Demographics. National Multifamily Housing Council.
27 Rapidly falling costs of battery packs for electric vehicles. Bjorn Nykvist and Mans Nilsson. Nature Climate
Change, 2015.
28 www.fueleconomy.gov . U.S. Department of Energy, 2015.
29https://www.caiso.com/Documents/Vehicle-GridIntegrationRoadmap.pdf.
30 Cleaner Cars from Cradle to Grave. Union of Concerned Scientists, November 2015.
31 Spaced Out-Perspectives on Parking Policy. Bates and Leibling, RAC Foundation, 2012.
32 California PEV Collaborative Sales Dashboard, accessed May 15, 2015.
33 Alternative Fuels Data Center. U.S. Department of Energy.
34 Alternative Fuels Data Center. U.S. Department of Energy accessed May 31, 2016.
35 ChargePoint. EVgo electric-car charging networks get more investment. GreenCar Reports, May 5, 2016.
36 Interim Joint Technical Assessment Report: Light Duty Vehicle Greenhouse Gas Emission Standards and
Corporate Average Fuel Economy Standards for Model Years 2017-2025. U.S. Environmental Protection Agency,
National Highway Traffic Safety Administration, and California Air Resources Board, 2010.
37 Notice of Proposed Award: Alternative and Renewable Fuel and Vehicle Technology Program PON-13-607. 1
May 2014. http://www.energy.ca.gov/contracts/PON-13-607_NOPA.pdf.
38 Notice of Proposed Award: Alternative and Renewable Fuel and Vehicle Technology Program PON-13-607. 1
May 2014. http://www.energy.ca.gov/contracts/PON-13-607_NOPA.pdf.
                                                     9-43

-------
                                                   Assessment of Alternative Fuel Infrastructure
39 2015 Annual Evaluation of Fuel Cell Electric Vehicle Deployment and Hydrogen Fuel Station Network
Development. Sacramento: California Air Resources Board, 2015.
40 A California Road Map: The Commercialization of Hydrogen Fuel Cell Vehicles. Sacramento: California Fuel
Cell Partnership, 2012.
41 2015 Annual Evaluation of Fuel Cell Electric Vehicle Deployment and Hydrogen Fuel Station Network
Development. Sacramento: California Air Resources Board, 2015.
42 2015 Annual Evaluation of Fuel Cell Electric Vehicle Deployment and Hydrogen Fuel Station Network
Development. Sacramento: California Air Resources Board, 2015.
43 Assembly Bill 8 Report: Assessment of Time and Cost Needed to Attain 100 Hydrogen Refueling Stations in
California. Sacramento: California Energy Commission and California Air Resources Board, 2015.
44 Assembly Bill 8 Report: Assessment of Time and Cost Needed to Attain 100 Hydrogen Refueling Stations in
California. Sacramento: California Energy Commission and California Air Resources Board, 2015.
45 2015 Annual Evaluation of Fuel Cell Electric Vehicle Deployment and Hydrogen Fuel Station Network
Development. Sacramento: California Air Resources Board, 2015.
46 Alternative Fuels Data Center: Hydrogen Fueling Station Locations, 31 August 2015
http://www.afdc.energy.gov/fuels/hydrogen locations.html.
47 NREL Dedicates Advanced Hydrogen Fueling Station, 8 October 2015
http://www.nrel.gov/news/press/2015/20582.
48 Public Notice: Hydrogen Fueling Infrastructure Development Program Incentives. 21 July 2015
http://www.ct.gov/deep/lib/deep/air/electric vehicle/evct/2015-07-21  -
 Hydrogen Infrastructure Solicitation Notice.pdf.
49 Air Liquide plans network of new hydrogen filling stations in the United States, 17 November 2014
http://www.us.airliquide.com/en/air-liquide-plans-network-of-us-hvdrogen-filling-stations.html.
50 Station Maps, 15 June 2015. http://h2usa.org/station-maps .
51 SAE J2719 Sept. 2011 Hydrogen Fuel Quality for Fuel Cell Vehicles.
52 SAE J2601 July 2014 Fueling Protocols for Light Duty Gaseous Hydrogen Surface Vehicles.
53 NIST Handbook 44 2015 Specifications, Tolerances, and Other Technical Requirements for Weighing and
Measuring Devices.
54 NIST Handbook 130 2015 Uniform Laws and Regulations in the Areas of Legal Metrology and Engine Fuel
Quality.
55 H2-Device Cal.  Code Regs., tit. 4, § 4002.9.
56 CSA HGV 4.3 2012 Test Methods for Hydrogen Fueling Parameter Evaluation.
57 Annual Evaluation of Fuel Cell Electric Vehicle Deployment and Hydrogen Fuel Station Network Development.
Sacramento: California Air Resources Board,  2014.
58 Assembly Bill 8 Report: Assessment of Time and Cost Needed to Attain 100 Hydrogen Refueling Stations in
California. Sacramento: California Energy Commission and California Air Resources Board, 2015.
59 Assembly Bill 8 Report: Assessment of Time and Cost Needed to Attain 100 Hydrogen Refueling Stations in
California. Sacramento: California Energy Commission and California Air Resources Board, 2015.
60 Assembly Bill 8 Report: Assessment of Time and Cost Needed to Attain 100 Hydrogen Refueling Stations in
California. Sacramento: California Energy Commission and California Air Resources Board, 2015.
61 Assembly Bill 8 Report: Assessment of Time and Cost Needed to Attain 100 Hydrogen Refueling Stations in
California. Sacramento: California Energy Commission and California Air Resources Board, 2015.
62 Assembly Bill 8 Report: Assessment of Time and Cost Needed to Attain 100 Hydrogen Refueling Stations in
California. Sacramento: California Energy Commission and California Air Resources Board, 2015.
63 Assembly Bill 8 Report: Assessment of Time and Cost Needed to Attain 100 Hydrogen Refueling Stations in
California. Sacramento: California Energy Commission and California Air Resources Board, 2015.
64 Stephens-Romero, Shane D., Brown, TimM., Kang, Jae E., Recker, Wilfred W., and Samuelsen, G. Scott. 2010,
"Systematic planning to optimize investments in hydrogen infrastructure deployment," International Journal of
Hydrogen Energy, 35, 4652-4667.
65 A California Road Map: The Commercialization of Hydrogen Fuel Cell Vehicles. Sacramento: California Fuel
Cell Partnership, 2012.
66 2015 Annual Evaluation of Fuel Cell Electric Vehicle Deployment and Hydrogen Fuel Station Network
Development. Sacramento: California Air Resources Board, 2015.
67 Scenario Evaluation, Regionalization & Analysis (SERA). 17 January 2011.
http://en.openei.org/wiki/Scenario Evaluation. Regionalization & Analysis (SERA).
68 Station Maps, 15 June 2015. http://h2usa.org/station-maps.

                                                     9-44

-------
                                                   Assessment of Alternative Fuel Infrastructure
69 An Analysis of Near-Term Hydrogen Vehicle Rollout Scenarios for Southern California. Davis: Michael Nicholas
and Joan Ogden, 2010.
70 Roadmap for Hydrogen and Fuel Cell Vehicles in California: A Transition Strategy through 2017. Davis: Ogden,
Joan, Cunningham, Joshua, and Nicholas, Michael, 2010.
71 Michael A. Nicholas. 2010, "Driving Demand: What can gasoline refueling patterns tell us about planning an
alternative fuel network?," Journal of Transport Geography, 18, 738-749.
72 Michael A. Nicholas. 2010, "Driving Demand: What can gasoline refueling patterns tell us about planning an
alternative fuel network?," Journal of Transport Geography, 18, 738-749.
73 Kuby, M., Lines, L., Schultz, R., Xie, Z., Kim, J., and Lim, S. 2009, "Optimization of hydrogen stations in Florida
using the flow-refueling locations model," International Journal of Hydrogen Energy, 34(15), 6045-6064.
74 Scott Kelley and Michael Kuby. 2013, "On the way or around the corner? Observed refueling choices of
alternative-fuel drivers in Southern California," Journal of Transport Geography, 33, 258-267.
75 POWER TO GAS - Demonstration Projects and H2 Filling Stations in Germany by Germany Trade and Invest.
http://www.gtai.de/GT AI/Content/EN/Invest/_SharedDocs/Downloads/GTAI/Maps/RER/map-energy-storage-
power-to-gas.pdf?v=2. Accessed 2/26/16.
76 Data on fueling stations is from the Alternative Fuels Data Center (AFDC), either directly (www.afdc.
energy.gov/afdc/fuels/stations_counts.html) or from historical Transportation Energy Data Books
(www.osti.gov/bridge/basicsearch.jsp).
77 See http://www.fsa.usda.gov/programs-and-services/energy-programs/bip/.
                                                     9-45

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

Table of Contents


Chapter 10: Economic and Other Key Inputs Used in the Agencies' Analyses	10-1
   10.1   The On-Road Fuel Economy "Gap"	10-1
     10.1.1   The "Gap" Between Compliance and Real World Fuel Economy	10-1
     10.1.2   Real World Fuel Economy and CCh Projections	10-2
   10.2   Fuel Prices and the Value of Fuel Savings	10-4
   10.3   Vehicle Mileage Accumulation and Survival Rates	10-6
   10.4   Fuel Economy Rebound Effect	10-9
     10.4.1   Accounting for the Fuel Economy Rebound Effect	10-9
     10.4.2   Summary of Historical Literature on the LD V Rebound Effect	10-10
     10.4.3   Review of Recent Literature on LDV Rebound since the 2012 Final Rule	10-15
     10.4.4   Basis for Rebound Effect Used in the Draft TAR	10-19
   10.5   Energy Security Impacts	10-21
     10.5.1   Implications of Reduced Petroleum Use on U.S.  Imports	10-21
     10.5.2   Energy Security Implications	10-24
        10.5.2.1   Effect of Oil Use on the Long-Run Oil Price	10-25
        10.5.2.2   Macroeconomic Disruption Adjustment Costs	10-28
        10.5.2.3   Cost of Existing U.S. Energy Security Policies	10-33
        10.5.2.4   Military Security Cost Components of Energy Security	10-34
   10.6   Non-GHG Health and Environmental Impacts	10-35
     10.6.1   Economic Value of Reductions in Particulate Matter	10-36
   10.7   Greenhouse Gas Emission Impacts	10-41
   10.8   Benefits from Reduced Refueling Time	10-50
   10.9   Benefits and Costs from Additional Driving	10-53
     10.9.1   Travel Benefit	10-53
     10.9.2   Costs Associated with Crashes, Congestion and Noise	10-53
   10.10    Discounting Future Benefits and Costs	10-54
   10.11    Additional Costs of Vehicle Ownership	10-55
     10.11.1  Maintenance  & Repair Costs	10-55
     10.11.2  Sales Taxes	10-55
     10.11.3  Insurance Costs	10-56

Table of Figures
Figure 10.1 Comparing AEO 2015 and AEO 2016 Early Release Retail Fuel Price Projections	10-5
Figure 10.2 U.S. Expenditures on Crude Oil from 1970 through 2015	10-22
Figure 10.3 Projected and Historical U.S. Expenditures, and Expenditure Share, on Crude Oil	10-30
Figure 10.4 Path from GHG Emissions to Monetized Damages (Source: Marten et al., 2014)	10-49


Table of Tables
Table 10.1 EPA Projections for Fleet-wide CO2 Standards Compliance and On-road Performance for Cars	10-3
Table 10.2 EPA Projections for Fleet-wide CO2 Standards Compliance and On-road Performance for Trucks.... 10-3
Table 10.3 EPA Projections for Fleet-wide CO2 Standards Compliance and On-road Performance for the Fleet. 10-4
Table 10.4 Gasoline Prices for Selected Years in Various AEO 2015 Cases	10-4
Table 10.5 Updated Vehicle Survival Rates (from MOVES 2014a)	10-7
Table 10.6 2011 Mileage Schedule (fromMOVES 2014a)	10-8
Table 10.7 Estimates of the Rebound Effect Using U.S. Aggregate Time-Series Data on Vehicle Travel	10-10

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

Table 10.8  Estimates of the Rebound Effect Using U.S. State Level Data	10-11
Table 10.9  Estimates of the Rebound Effect Using U.S. Survey Data	10-11
Table 10.10 Projected Trends in U.S. Oil Exports/Imports, and U.S. Oil Import Reductions Resulting from the
             Program in Selected Years from 2022 to 2050 (Millions of barrels per day (MMBD)	10-24
Table 10.11 Energy Security Premiums in Selected Years from 2022 to 2050, (2013$/Barrel)*	10-25
Table 10.12 PM-Related Benefits-per-ton Values (thousands, 2012$)a	10-37
Table 10.13 Human Health and Welfare Effects of PM25	10-38
Table 10.14 Social Cost of CO2, 2015-2050 (in 2013$ per metric ton)*	10-45
Table 10.15 Social Cost of CH4 and Social Cost of N2O, 2015-2050 (in 2013$ per metric ton)	10-47
Table 10.16 Metrics Used in Calculating the Value of Refueling Time	10-51
Table 10.17 Metrics Used in Calculating the Value of Refueling Time by NHTS A	10-51
Table 10.18 Estimating the Value of Travel Time for Urban and Rural (Intercity) Travel ($/hour)	10-52
Table 10.19 Estimating the Value of Travel Time for Light-Duty Vehicles ($/hour)	10-52
Table 10.20 Metrics Used to Calculate the Costs Associated with Congestion, Crashes and Noise Linked to
             Rebound Miles Traveled	10-54
Table 10.21 Maintenance Event Costs & Intervals (2013$)	10-55

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

Chapter 10:Economic and Other Key Inputs Used in the Agencies' Analyses

10.1  The On-Road Fuel Economy  "Gap"

10.1.1 The "Gap" Between Compliance and Real World Fuel Economy

   Real world tailpipe CCh emissions are higher, and real world fuel economy levels are lower,
than the corresponding values from EPA standards compliance tests.  This is because laboratory
testing cannot reflect all of the factors that can affect real world operation, and, in particular, the
city and highway tests used for compliance  do not encompass the broad range of driver behavior
and climatic conditions experienced by typical U.S. drivers.A In  the rulemakings that established
the National Program standards through MY2025, EPA and NHTSA applied a 20 percent fleet-
wide fuel economy "gap," i.e., that average, fleet-wide real world fuel economy would be 20
percent lower than EPA compliance test values.6 This 20 percent value was based on data from
MY2004-2006.1 For example, a vehicle with  a fuel economy compliance test value of 30 mpg
would be projected to have a real world fuel economy of 30 multiplied by 0.8 (equivalent to a 20
percent reduction) or 24 mpg. The inverse of 0.8 is 1.25, and a vehicle with a CCh emissions
compliance test value of 300 grams/mile would be projected to have a real world CCh emissions
value of 300 multiplied by 1.25  or 375 grams/mile.

   More recent data suggests that the gap between 2-cycle compliance test and 5-cycle
methodology values may have increased very  slightly in the last decade.  For example, the use of
final MY2014 and projected MY2015 data suggest that the fuel economy gap between 2-cycle
data and 5-cycle data may now be approximately 21 percent.2  EPA believes that further analysis
is needed before incorporating such  small changes into calculations of the overall gap. In
addition, some analysis suggests that the gap between 2-cycle compliance tests and real world
fuel economy may be increasing in recent years, but the evidence is not conclusive.3  One factor
which has clearly changed and can be quantified is ethanol content in gasoline. When the 20
percent fuel economy gap was first projected in 2005-2006, ethanol accounted for a small
fraction of the gasoline pool. For this Draft TAR analysis, EPA adjusts for projected differences
in the energy content due to increased ethanol penetration of retail gasoline  relative to test fuel
for MY2022 and beyond.. Ethanol contains about 35 percent less energy than gasoline, on a
volumetric basis, and EPA projects that average in-use gasoline will contain about 3.5 percent
less energy in 2025 than it did in the 2005-2006 timeframe. Using  the "base" 20 percent fuel
economy gap between 2-cycle and 5-cycle data and the projected impact of the ethanol increase
in 2025 yields an effective gap of 23 percent (or a fuel  economy factor of 0.77), and this is the
A EPA has recognized that the "2-cycle" city and highway tests are not representative of real world fuel economy
  performance for over 30 years. From MY 1985 through MY2007, EPA based new vehicle window labels on the
  fuel economy compliance test values adjusted downward by 10% for the city test and by 22% for the highway
  test. Beginning in MY2008, EPA has based vehicle labels on a 5-cycle methodology that includes three additional
  tests (reflecting high speed/high acceleration, hot temperature/air conditioning, and cold temperature operation) as
  well as a 9.5% downward fuel economy adjustment for other factors not reflected in the 5-cycle protocol.
B Note that this is an average fleet-wide value, in reality the true fuel economy gap is data driven and will be lower
  for some vehicles and higher for other vehicles. In general, all things being equal, today's data suggests that the
  gap is generally smaller for lower-fuel economy vehicles and greater for higher-fuel economy vehicles.
                                              10-1

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

overall fuel economy gap that we use in this report. Multiplying 2-cycle fuel economy by 0.77
yields projected real world fuel economy.0

   The fuel economy gap is data driven, so any 2025 projection involves uncertainty. EPA
expects that, all other things being equal, as average fuel economy increases over time, the gap
would likely increase as well. On the other hand, it is also possible that powertrain designs will
be designed to be more robust in the future, which would impact the gap in the opposite
direction. EPA will continue to monitor the relevant data on this issue.

10.1.2 Real World Fuel Economy and CCh Projections

   Except when noted, CCh emissions and fuel economy values cited in this report represent
standards compliance values. As discussed above, real world tailpipe CCh emissions are higher,
and real world fuel economy levels are lower, than the corresponding values from EPA standards
compliance tests.

   This has led to widespread public confusion as there are two sets of fuel economy "books,"
one for fuel economy standards compliance (mandated by statute for cars) and one for the
vehicle label estimates that EPA provides to consumers to estimate real world fuel economy.
The projected real world fuel economy values shown below are the most meaningful fuel
economy values for citizens and reporters as they provide a good comparison with label values,
EPA Fuel Economy Trends report values, vehicle dashboard display values, and fuel economy
calculations performed by some drivers, and also correspond to real world fuel consumption and
CCh emissions.

   Table 10.1 through 10.3 show EPA's best projections of the real world CCh emissions and
fuel economy values associated with the projected CCh standards compliance emissions levels
presented throughout this report, as well as how "the numbers add  up," for cars, trucks, and the
combined car/truck fleet, respectively.  These values use as a starting point the projected
industry-wide CCh 2-cycle targets.  The first step is to "back out" the impact of the direct air
conditioner refrigerant credits, since reducing leakage and/or substituting lower-GHG
refrigerants will not increase real world fuel economy. Backing out these credits requires adding
the value of the air conditioner refrigerant credits to the target values, as doing so increases the
CCh value and decreases the  projected real world fuel economy level. The sum of the 2-cycle
target and the "backed out" air conditioner refrigerant credits is the "fuel economy-relevant
adjusted 2-cycle CCh emissions value," shown as the effective CCh value in the tables which can
also be expressed as an effective mpg by dividing it into 8887 (which represents the number of
grams of CCh that results from the combustion of a gallon of test gasoline). The second step is to
multiply the adjusted 2-cycle, or effective mpg value by 0.77, the fuel economy "gap"  factor
discussed above.  This step converts from the adjusted 2-cycle mpg to a real world, on-road mpg
value. On-road tailpipe CO2 emissions are projected by dividing the real world mpg value into
8488  (which represents the number of grams of CO2that results from the combustion of a gallon
of retail gasoline). Subtracting back the A/C leakage credit value provides an on-road CO2
equivalent (CO2 e) value as shown.
c The corresponding CCh "gap" is 1.24, i.e., multiplying 2-cycle tailpipe CCh by 1.24 yields projected real world
  CCh emissions. This 1.24 factor is actually less than the 1.25 factor used in the past because of the lower carbon
  content of ethanol.
                                              10-2

-------
 	Economic and Other Key Inputs Used in the Agencies' Analyses

 Table 10.1 EPA Projections for Fleet-wide CCh Standards Compliance and On-road Performance for Cars

MY
2021
2022
2023
2024
2025
2-Cycle
C02
Target
(g/mi)
171
165
159
153
147
C02
Target
As
MPG
51.9
53.9
56.0
58.2
60.3
A/C
Leakage
Credit
(g/mi)
13.8
13.8
13.8
13.8
13.8
A/C
Efficiency
Credit
(g/mi)
5.0
5.0
5.0
5.0
5.0
Off-
cycle
Credit
(g/mi)
0.6
0.7
0.9
1.0
1.1
Tailpipe
C02
(g/mi)
191
184
178
173
167
MPG
46.6
48.2
49.8
51.5
53.2
Adjustments to 2-Cycle
to Reflect Real World
Impacts
A/C
Efficiency
& Off-
cycle
Credits
(g/mi)
5.6
5.7
5.9
6.0
6.1
Effective
C02
(g/mi)
185
179
173
167
161
Effective
MPG
48.1
49.8
51.5
53.3
55.2
On-road
Gap
.773
.773
.773
.773
.773
On-
road
MPG
37.1
38.4
39.8
41.2
42.6
On-
road
Tailpipe
C02
(g/mi)
229
221
213
206
199
On-
road
CO2e
(g/mi)
215
207
200
192
186
Note: The on-road values reflect adjustments for both the historical 2-cycle-to-5-cycle gap as well as the projected
ethanol content in retail gasoline, and corresponding energy content. The on-road CC>2 is calculated by dividing
8488, the estimated CCh grams/gallon from combustion of a gallon of retail gasoline, by the on-road MPG. The on-
road CChe column subtracts from the on-road tailpipe €62 values the A/C leakage value to yield a value that reflects
overall real world CChe emissions performance.


Table 10.2 EPA Projections for Fleet-wide CCh Standards Compliance and On-road Performance for Trucks



MY





2021
2022
2023
2024
2025
2-Cycle


C02
Target
(g/mi)



242
232
223
214
206
C02
Target
As
MPG


36.7
38.3
39.9
41.6
43.2
A/C
Leakage
Credit
(g/mi)


17.2
17.2
17.2
17.2
17.2
A/C
Efficiency
Credit
(g/mi)


7.2
7.2
7.2
7.2
7.2
Off-
cycle
Credit
(g/mi)


2.3
2.6
2.9
3.2
3.5
Tailpipe
CO2
(g/mi)



269
259
250
241
233
MPG





33.1
34.3
35.6
36.8
38.1
Adjustments to 2-Cycle to
Reflect Real World
Impacts
A/C
Efficiency
& Off-
cycle
Credits
(g/mi)
9.5
9.8
9.9
10.4
10.7
Effective
C02
(g/mi)



259
250
240
231
223
Effective
MPG




34.3
35.6
37.0
38.5
39.9
On-road


Gap





.773
.773
.773
.773
.773
On-
road
MPG



26.5
27.5
28.6
29.7
30.8
On-road
Tailpipe
C02
(g/mi)


321
309
297
286
276
On-
road
CO2e
(g/mi)


304
292
280
269
258
Note: The on-road values reflect adjustments for both the historical 2-cycle-to-5-cycle gap as well as the projected
ethanol content in retail gasoline, and corresponding energy content. The on-road CO2 is calculated by dividing
8488, the estimated CCh grams/gallon from combustion of a gallon of retail gasoline, by the on-road MPG. The on-
road CChe column subtracts from the on-road tailpipe CO2 values the A/C leakage value to yield a value that reflects
overall real world CChe emissions performance.
                                                    10-3

-------
 	Economic and Other Key Inputs Used in the Agencies' Analyses

  Table 10.3 EPA Projections for Fleet-wide CCh Standards Compliance and On-road Performance for the
                                          Fleet


MY





2021
2022
2023
2024
2025
2-Cycle

C02
Target
(g/mi)



206
198
190
182
175
C02
Target
As
MPG


43.1
44.9
46.8
48.8
50.8
A/C
Leakage
Credit
(g/mi)


15.5
15.5
15.5
15.4
15.4
A/C
Efficiency
Credit
(g/mi)


6.1
6.1
6.1
6.1
6.0
Off-
cycle
Credit
(g/mi)


1.5
1.7
1.9
2.1
2.3
Tailpipe
C02
(g/mi)



229
221
213
206
199
MPG





38.8
40.2
41.7
43.2
44.7
Adjustments to 2-Cycle to
Reflect Real World Impacts
A/C
Efficiency
& Off-
cycle
Credits
(g/mi)
7.6
7.8
8.0
8.2
8.4
Effective
C02
(g/mi)



222
213
205
198
190
Effective
MPG




40.1
41.6
43.3
45.0
46.7
On-road

Gap





.773
.773
.773
.773
.773
On-
road
MPG



30.9
32.1
33.4
34.7
36.0
On-road
Tailpipe
C02
(g/mi)


274
264
254
245
236
On-
road
CO2e
(g/mi)


259
249
239
229
220
Note: The on-road values reflect adjustments for both the historical 2-cycle-to-5-cycle gap as well as the projected
ethanol content in retail gasoline, and corresponding energy content. The on-road CC>2 is calculated by dividing
8488, the estimated CCh grams/gallon from combustion of a gallon of retail gasoline, by the on-road MPG. The on-
road CChe column subtracts from the on-road tailpipe €62 values the A/C leakage value to yield a value that reflects
overall real world CO2e emissions performance.
   EPA projects the industry-wide real world fuel economy associated with the MY2025 GHG
standards to be 36 mpg.  This value provides a good comparison with average label and Fuel
Economy Trends values.
10.2  Fuel Prices and the Value of Fuel Savings

   Fuel prices and the projection of fuel prices remain critical in the analysis of GHG and fuel
economy standards. EPA has continued to use the methodology described in section 4.2 of the
2012 Joint Technical Support Document, with some notable updates.  EPA continued to rely on
the fuel price projections from the U.S. Energy Information Administration's (EIA) Annual
Energy Outlook (AEO) for this analysis, updated to the AEO 2015 Reference Case. The
Reference case  projection is a business-as-usual trend estimate, given known technology and
technological and demographic trends. EIA has published annual projections of energy prices
and consumption levels for the U.S. economy since 1982 in its Annual Energy Outlook reports.
These projections have been widely relied upon by federal agencies for use in regulatory analysis
and for other purposes. Since 1994, EIA's annual forecasts have been based upon the agency's
National Energy Modeling System (NEMS), which includes detailed representation of supply
pathways, sources of demand, and their interaction to determine prices for different forms of
energy. In addition to the AEO 2015 Reference Case as the central case, EPA has also included
the AEO 2015 low and high fuel price cases as sensitivities. A comparison of these cases is
presented below in Table 10.4.
               Table 10.4 Gasoline Prices for Selected Years in Various AEO 2015 Cases

AEO 2015 Reference Case
AEO 2015 "Low" Case
AEO 2015 "High" Case
2025
$ 2.95
$ 2.40
$ 4.56
2030
$ 3.20
$ 2.45
$ 5.05
2040
$ 3.90
$ 2.60
$ 6.33
                                              10-4

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

   The retail fuel price forecasts presented in AEO 2015 span the period from 2012 through
2040.  Measured in constant 2013 dollars, the AEO 2015 Reference Case projections of retail
gasoline prices during calendar year 2025 is $2.95 per gallon, rising gradually to $3.90 by the
year 2040 (these values include federal and state taxes). However, valuing fuel savings over the
full lifetimes of passenger cars and light trucks affected by the standards for MYs 2012-25
requires fuel price forecasts that extend through approximately 2060, approximately the last year
during which a significant number of MY2025 vehicles will remain in service. Due to the
difficulty in accurately projecting fuel prices over this long time span, EPA has assumed constant
fuel prices after the year 2040 for the Draft TAR analysis.

   The AEO 2016 Early Release (AEO 2016ER) was released in June 2016, as the agencies'
Draft TAR analyses were well underway. While there are some differences between the AEO
2015 and AEO 2016ER fuel price projections, especially in earlier years, the projection prices
are similar over the 2022 and beyond timeframe. Moreover, the AEO 2016ER fuel price
projections fall well within the range of the AEO 2015 low and high fuel price sensitivity cases
analyzed as part of Chapter 12.4.  The agencies plan to update their analyses based on the latest
available AEO projections for later steps of the midterm evaluation and  CAFE rulemaking
process.
          $6.50

          $6.00

          $5.50

          $5
	AE02016ER Retail Gasoline
	AE02015 Retail Gasoline "Central Case"
	AE02015 Retail Gasoline "Low Price Case"
	AE02015 Retail Gasoline "High Price Case"
          $2.
          $1.50
              2010
                       2015
                               2020
                                        2025
                                                 2030
                                                          2035
                                                                   2040
                                                                           2045
       Figure 10.1 Comparing AEO 2015 and AEO 2016 Early Release Retail Fuel Price Projections

   The value of fuel savings resulting from improved fuel economy and reduced GHG emissions
to buyers of light-duty vehicles is determined by the retail price of fuel, which includes federal,
state, and any local taxes imposed on fuel sales. Total taxes on gasoline, including federal, state,
and local levies, averaged $0.41 per gallon during 2013, while those levied on diesel averaged
$0.48. Because fuel taxes represent transfers of resources from fuel buyers to government
agencies, rather than real resources that are consumed in the process of supplying or using fuel,
their value must be deducted from retail fuel prices to determine the value of fuel savings
resulting from more stringent fuel efficiency and GHG standards to the U.S.  economy.  When
calculating the value of fuel saved by an individual driver, however, these taxes are included as
                                               10-5

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

part of the value of realized fuel savings.  Over the entire period spanned by the agencies'
analysis, this difference causes each gallon of fuel saved to be valued by about $0.39 (in constant
2013 dollars) more from the perspective of an individual vehicle buyer than from the overall
perspective of the U.S. economy.

10.3   Vehicle Mileage Accumulation and  Survival Rates

   EPA's analysis of benefits from fuel economy and GHG standards for passenger cars and
light trucks, including GHG reductions, oil reductions, and fuel savings, begin by estimating the
resulting changes in fuel use over the entire lifetimes of affected cars and light trucks.  The
change in total fuel consumption by vehicles produced during each of these model years is
calculated as the difference in their total lifetime fuel use over the entire lifetimes of these
vehicles as compared to a reference case.

   EPA's approach for this analysis remains largely the same as that found in the 2012 FRM
TSD, Chapter 4.2. Since the FRM, EPA has updated a few key inputs related to vehicle lifetime
survival rates and total vehicle miles traveled (VMT), as described in Table 10.5 and Table 10.6
below.  These updates were made in order to align this analysis with inputs developed in
conjunction with the EPA MOVES 2014a model4, which has integrated new activity and
population data sources from R.L. Polk, FHWA, and the EIA Annual Energy Outlook following
the release of the FRM.5 Additionally, the MOVES model is also already  used as part of other
EPA rulemaking analyses, allowing this analysis to take advantage of updates from those efforts.
Methodologies for the derivation of fuel savings and related benefits (including future year
projections,  VMT growth factor, and fuel cost per mile) from these inputs remain identical to
those found in the FRM TSD.
                                              10-6

-------
         Economic and Other Key Inputs Used in the Agencies' Analyses
Table 10.5 Updated Vehicle Survival Rates (from MOVES 2014a)
VEHICLE AGE
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
ESTIMATED SURVIVAL FRACTION (CARS)
1.000
0.997
0.994
0.991
0.984
0.974
0.961
0.942
0.920
0.893
0.862
0.826
0.788
0.718
0.613
0.510
0.415
0.332
0.261
0.203
0.157
0.120
0.092
0.070
0.053
0.040
0.030
0.023
0.013
0.010
0.007
0.002
ESTIMATED SURVIVAL FRACTION (LIGHT TRUCKS)
1.000
0.991
0.982
0.973
0.960
0.941
0.919
0.891
0.859
0.823
0.784
0.741
0.697
0.651
0.605
0.553
0.502
0.453
0.407
0.364
0.324
0.288
0.255
0.225
0.198
0.174
0.153
0.133
0.117
0.102
0.089
0.027
                            10-7

-------
     Economic and Other Key Inputs Used in the Agencies' Analyses
Table 10.6 2011 Mileage Schedule (from MOVES 2014a)
VEHICLE AGE
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
TOTAL
ESTI MATED VMT CARS
13,843
13,580
13,296
12,992
12,672
12,337
11,989
11,630
11,262
10,887
10,509
10,129
9,748
9,370
8,997
8,629
8,270
7,922
7,586
7,265
6,962
6,679
6,416
6,177
5,963
5,778
5,623
5,499
5,410
5,358
5,358
278,134
ESTIMATED VMT LIGHT TRUCKS
15,962
15,670
15,320
15,098
14,528
14,081
13,548
13,112
12,544
12,078
11,595
11,131
10,641
10,153
9,691
9,239
8,797
8,383
8,009
7,666
7,358
7,089
6,862
6,684
6,556
6,481
6,466
6,466
6,466
6,466
6,466
310,610
                        10-8

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

 10.4  Fuel Economy Rebound Effect

 10.4.1 Accounting for the Fuel Economy Rebound Effect

   The rebound effect generally refers to the additional energy consumption that may arise from
 the introduction of a more efficient, lower cost energy service which offsets, to some degree, the
 energy savings benefits of that efficiency improvement.6'7'8  In the context of light-duty vehicles
 (LDVs), rebound effects might occur when an increase in vehicle fuel efficiency encourages
 people to drive more as a result of the lower cost per mile of driving. Because this additional
 driving consumes fuel and generates emissions, the magnitude of the rebound effect is one
 determinant of the actual fuel savings and emission reductions that will result from adopting
 stricter fuel economy or GHG emissions standards.

   The rebound effect for personal vehicles can in theory be estimated directly from the change
 in vehicle use, in terms  of vehicle miles traveled (VMT), which results from a change in vehicle
 fuel efficiency.0  In practice, any attempt to quantify this "VMT rebound effect" (sometimes also
 labeled the "direct rebound effect," or "direct VMT rebound effect") is complicated by the
 difficulty in identifying an applicable data source from which the response to a significant
 improvement in fuel efficiency can be estimated. Analysts instead often estimate the VMT
 rebound indirectly as the change in vehicle use that results from a change in fuel cost per mile
 driven or a change in fuel price. When a fuel cost per mile approach is used, it does not
 distinguish the relative contributions of changes in fuel efficiency and changes in fuel price to
 the rebound effect, since both factors are determinants of fuel cost per mile.E

   When expressed as positive percentages, the elasticities of vehicle use with respect to fuel
 efficiency or per-mile fuel costs (or fuel prices) give the percentage increase in vehicle use that
 results from a doubling  of fuel efficiency (e.g., 100 percent increase), or a halving of fuel
 consumption or fuel price.  For example, a 10 percent rebound effect means that a 20 percent
 reduction in fuel consumption or fuel price (and the corresponding reduction in fuel cost per
 mile) is expected to result in a two percent increase in vehicle use.

   While we focus on the VMT rebound effect in our analysis of this program, there are at least
 two other types of rebound effects discussed in the transportation policy and economics
 literature. In addition to direct VMT rebound effect, there is the "indirect" rebound effect, which
 typically refers to the purchase  of other goods or services that consume energy with the costs
 savings from energy efficiency improvements. The last type of rebound effect is labeled the
 "economy-wide" rebound effect.  This  effect refers to the increased demand for energy
 throughout the whole economy in response to the reduced market price of energy that happens as
 a result of energy efficiency improvements.

   Research on indirect and economy-wide rebound effects is scant, and we have not identified
 any studies that attempt to quantify indirect or economy-wide rebound effects that result from
 improvements in the energy efficiency  of LDVs. In particular, the agencies are not aware of any
D Vehicle fuel efficiency is more often measured in terms of fuel consumption (gallons per mile) rather than fuel
  economy (miles per gallon) in rebound estimates.
E Fuel cost per mile is equal to the price of fuel in dollars per gallon divided by fuel economy in miles per gallon (or
  multiplied by fuel consumption in gallons per mile), so this figure declines when a vehicle's fuel efficiency
  increases.
                                               10-9

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

data to indicate that the magnitude of indirect or economy-wide rebound effects would be
significant for this National Program.  Therefore, the rebound effect discussed in this section
refers solely to the effect of increased fuel efficiency on vehicle use.  The terms, "VMT rebound
effect," "direct VMT rebound effect," and "rebound effect" can be used interchangeably,  and
they need to be distinguished from other rebound effects that could potentially impact the fuel
savings and emissions reductions from our standards such as the "indirect rebound effect."  To
restate, the rebound effect discussed in this section refers solely to the effect of increased fuel
efficiency on vehicle use.

   This section surveys previous studies on the LDV rebound effect,  summarizes recent work on
the rebound effect, and explains the basis for the 10 percent rebound  effect EPA and NHTSA are
using for the Draft TAR analyses.
10.4.2 Summary of Historical Literature on the LDV Rebound Effect

   It is important to note that a majority of the studies previously conducted on the rebound
effect rely on data from the 1950-1990s.  While these older studies provide valuable information
on the potential magnitude of the rebound effect, studies that include more recent information
(e.g., data within the last decade) may provide more reliable estimates of how the standards will
affect future driving behavior.  Recent studies on LDV rebound effects that have become
available since the 2012 final rule are summarized in Section 10.4.3 below.

   Estimates based on aggregate U.S. vehicle travel data published by the U.S.  Department of
Transportation, Federal Highway Administration, covering the period from  roughly 1950 to
1990, have found long-run rebound effects on the order of 10-30 percent. Some of these studies
are summarized in the following two Tables.
    Table 10.7  Estimates of the Rebound Effect Using U.S. Aggregate Time-Series Data on Vehicle Travel
Author
(year)
Mayo & Mathis (1988)
Gately (1992)
Greene (1992)
Jones (1992)
Schimek (1996)
Short-Run
22%
9%
Linear 5-19%
Log-linear 13%
13%
5-7%
Long-Run
26%
9%
Linear 5-19%
Log-linear 13%
30%
21-29%
Time Period
1958-84
1966-88
1957-89
1957-89
1950-94
 Source: Sorrell and Dimitropolous (2007) table 4.6.9
                                              10-10

-------
               	Economic and Other Key Inputs Used in the Agencies' Analyses

                Table 10.8 Estimates of the Rebound Effect Using U.S. State Level Data
Author
(year)
Haughton&Sarkar(1996)
Small and Van Dender
(2005 and 2007a)
Hymel, Small and Van
Dender (2010)
Short-Run
9-16%
4.5%
2.2%
4.7%
4.8%
Long-Run
22%
22.2%
10.7%
24.1%
15.9%
Time Period
1973-1992
1966-2001
1997-2001
1966-2004
1984-2004
Source: Sorrell and Dimitropolous (2007) table 4.7 and the agencies' addition of recent work by Small and Van Dender
(2007a) and Hymel, Small, and Van Dender (2010).


   While studies using national (Table 10.7) and state level (Table 10.8) data have found
relatively consistent long-run estimates of the rebound effect, household surveys display more
variability (Table 10.9).  One explanation is that these studies consistently find that the
magnitude of the rebound effect differs according to the number of vehicles a household owns,
and the average number of vehicles owned per household differs among the surveys used to
derive these estimates. Still another possibility is that it is difficult to distinguish the impact of
fuel cost per mile on vehicle use from that of other, unobserved factors. For example,
commuting distance might influence both the choice of the vehicle as well as VMT.  Residential
density may also influence both fuel cost per mile and VMT, since households in urban areas are
likely to simultaneously face both higher fuel prices and shorter travel distance.  Also, given that
household data tends to be collected on an annual basis, there may not be enough variability in
the fuel price data to estimate the magnitude of the rebound effect.10
                 Table 10.9 Estimates of the Rebound Effect Using U.S. Survey Data
Author
(year)
Goldberg (1996)
Greene, Kahn, and
Gibson (1999a)
Pickrell &Schimek
(1999)
Puller & Greening
(1999)
West (2004)
West and Pickrell
(2011)
Estimate of Rebound Effect
0%
23%
4-34%
49%
87%
9-34%
Time Period
CES 1984-90
EIA RTECS
1979-1994
NPTS 1995
Single year
CES 1980-90
Single year, cross-sectional
CES 1997
Single year
NHTS 2009
Single year
  Source: Sorrell and Dimitropolous (2007). The agencies added a more recent study by West and Pickrell (2011).

   It is important to note that some of these studies actually quantify the price elasticity of
gasoline demand (e.g., Puller & Greening (1999)11) or the elasticity of VMT with respect to the
price of gasoline (e.g., Pickrell & Schimek (1999)12), rather than the elasticity of VMT with
respect to fuel efficiency or the fuel cost per mile of driving.  These latter measures more closely
match the definition of the fuel economy rebound effect. In fact, most studies cited above do not
estimate the direct measure of the fuel economy rebound effect (i.e., the increase in VMT
attributable to an increase in fuel efficiency).
                                               10-11

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

   Another important distinction among studies of the rebound effect is whether they assume that
the effect is constant, or varies over time in response to the absolute levels of fuel costs, personal
income, or household vehicle ownership.  Most studies using aggregate annual data for the U.S.
assume a constant rebound effect, although some of these studies test whether the effect can vary
as changes in retail fuel prices or average fuel efficiency alter fuel cost per mile driven. Many
studies using household survey data estimate significantly different rebound effects for
households owning varying numbers of vehicles, with most finding that the rebound effect is
larger among households that own more vehicles.F

   In addition to the studies listed above, Bento et al.  (2009)13 combined demographic
characteristics of more than 20,000 U.S. households, the manufacturer and model of each vehicle
they owned,  and their annual usage of each vehicle from the 2001 National Household Travel
Survey with  detailed data on fuel economy and other attributes for each vehicle model obtained
from commercial publications. The authors aggregated vehicle models into 350 categories
representing combinations of manufacturer, vehicle type, and age, and use the resulting data to
estimate the  parameters of a complex model of households' joint choices of the number and
types of vehicles to own, and their annual use of each vehicle.

    Bento et  al. estimate the effect of vehicles' operating cost per mile, including fuel costs -
which depend in part on each vehicle's fuel economy - as well as maintenance and insurance
expenses, on households' annual use of each vehicle they own.  Combining the authors'
estimates of the elasticity of vehicle use with respect to per-mile operating costs with the
reported fraction of total operating costs accounted for by fuel (slightly less than one-half) yields
estimates of the rebound effect. The resulting values vary by household composition, vehicle
size and type, and vehicle age, ranging from 21 to 38 percent, with a composite estimate of 34
percent for all households, vehicle models, and ages.  The smallest values apply to new luxury
cars, while the largest estimates are for light trucks and households with children, but the implied
rebound effects  differ little by vehicle age.

   Wadud et al.  (2009)14 combine data on U.S. households' demographic characteristics and
expenditures on gasoline over the period 1984-2003 from the Consumer Expenditure Survey
with data on gasoline prices and an estimate of the average fuel economy of vehicles  owned by
individual households (constructed from a variety of sources).  They employ these data to
explore variation in the sensitivity of individual households' gasoline consumption to differences
in income, gasoline prices, the number of vehicles owned by each household, and their average
fuel economy. Using an estimation procedure intended to account for correlation among
unmeasured  characteristics of households and among estimation errors for successive years, the
F Six of the household survey studies evaluated in Table 10.6 found that the rebound effect varies in relation to the
  number of household vehicles. Of those six studies, four found that the rebound effect rises with higher vehicle
  ownership, and two found that it declines. The four studies with rebound estimates that increase with higher
  household vehicle ownership are: Greene, D., and Hu, P., "The Influence of the Price of Gasoline on Vehicle Use
  in Multi-vehicle Households," Transportation Research Record 988, pp. 19-24,; Hensher, D., Milthorpe, F. and
  Smith, N., "The Demand for Vehicle Use in the Urban Household Sector: Theory and Empirical Evidence,"
  Journal of Transport Economics and Policy, 24:2 (1990), pp. 119-137; Walls, M, Krupnick A., and Hood, H.,
  "Estimating the Demand for Vehicle-Miles Traveled Using Household Survey Data: Results from the 1990
  Nationwide Personal Transportation Survey," Discussion Paper ENR 93-25, Energy and Natural Resources
  Division, Resources for the Future, Washington, D.C., 1993; and West, R. and Pickrell, D., "Factors Affecting
  Vehicle Use in Multiple-Vehicle Households," 2009 National Household Travel Survey Workshop, June 2011.
                                               10-12

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

authors explore variation in the response of fuel consumption to fuel economy and other
variables among households in different income categories, and between those residing in urban
and rural areas.

   Dividing U. S. households into five equally-sized income categories, Wadud et al. estimate
rebound effects ranging from 1-25 percent, with the smallest estimates (8 percent and 1 percent)
for the two lowest income categories, and significantly larger estimates for the middle (18
percent) and two highest income groups (18 and 25 percent).  In a separate analysis, the authors
estimate rebound effects of seven percent for households of all income levels residing in U.S.
urban areas, and 21 percent for rural households.

   West and Pickrell (2011)15 analyzed data on more than 100,000 households and 300,000
vehicles from the 2009 Nationwide Household Transportation Survey to explore how households
owning multiple vehicles chose which of them to use and how much to drive each one on the day
the household was surveyed.  Their study focused on how the type and fuel economy of each
vehicle a household owned, as well as its demographic characteristics and location, influenced
household members' decisions about whether and how much to drive each vehicle. They also
investigated whether fuel economy and fuel prices exerted similar influences on vehicle use, and
whether households owning more than one vehicle tended to substitute use of one for another -
or vary their use of all of them similarly - in response to fluctuations in fuel prices and
differences in their vehicles' fuel economy.

   Their estimates of the fuel economy rebound effect ranged from as low as nine percent to as
high as 34 percent, with their lowest estimates typically applying to single-vehicle households
and their highest values to households owning three or more vehicles. They generally found that
differences in fuel prices faced by households who were surveyed on different dates or who lived
in different regions of the U.S. explained more of the observed variation in daily vehicle use than
did differences in vehicles' fuel economy. West and Pickrell also found that while the rebound
effect for households' use of passenger cars appeared to be quite large - ranging from 17 percent
to nearly twice that value - it was difficult to detect a consistent rebound effect for SUVs.

   In addition, some recent studies (Small and Van Dender (2007), Hymel, Small, and Van
Dender (2010), (2012)), using both state-level and national data, conclude that the rebound effect
varies directly in response to changes in personal  income, as well as fuel costs. These more
recent studies published between 2007 and 2012 indicate that the rebound effect has decreased
over time as incomes have risen and, until recently, fuel costs as a share of total monetary travel
costs have generally decreased.0 One theoretical argument for why the rebound effect should
vary over time is that the responsiveness to the fuel cost of driving will be larger when it is a
larger proportion of the total cost of driving. For example, as incomes rise, the responsiveness to
the fuel cost per mile of driving will decrease  if people view the time cost of driving - which is
likely to be related to their income levels - as  a larger component of the total cost.

   Small and Van Dender (2007)16 combined time series data for each of the 50 states and the
District of Columbia to estimate the rebound effect, allowing the magnitude of the rebound to
  While real gasoline prices have varied over time, fuel costs (which reflect both fuel prices and fuel efficiency) as a
  share of total vehicle operating costs declined substantially from the mid-1970s until the mid-2000s when the
  share increased modestly (see Greene (2012)). With the recent decline in world petroleum prices, total vehicle
  operating costs have declined recently as well.
                                              10-13

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

vary over time. For the time period from 1966-2001, their study found a long-run rebound effect
of 22.2 percent, which is consistent with previously published studies.  But for the five year
period (1997-2001) estimated in their study, the long-run rebound effect decreased to 10.7
percent.  Furthermore, when the authors updated their estimates with data through 2004, the
long-run rebound effect for the most recent five year period (2000-2004) dropped to six
percent.17

   Hymel, Small and Van Dender (2010)18 extended  the Small and Van Dender model by adding
congestion as an endogenous variable. Although controlling for congestion increased their
estimates of the rebound effect, Hymel, Small and Van Dender also found that the rebound effect
was declining over time.  For the time period from 1966-2004, they estimated a long-run
rebound effect of 24 percent, while for 2004 they estimated a long-run rebound effect of 13
percent.

   Research conducted by David Greene (2012)19 under contract with EPA further appears to
support the theory that the magnitude of the rebound  effect "is by now on the order of 10
percent."11  Like Small and Van Dender, Greene finds that the VMT rebound effect could decline
modestly over time as household income rises and travel costs increase.  Over the entire time
period analyzed (1966-2007), Greene found that fuel prices had a statistically significant impact
on VMT, while fuel efficiency did not, which is similar to Small and Van Dender's prior finding.
From this perspective, if the impact of fuel efficiency on VMT is not statistically significant, the
VMT rebound effect could be zero. When Small and Van Dender tested whether the elasticity of
vehicle travel with respect to the price of fuel was equal to the elasticity with respect to the rate
of fuel consumption (gallons per mile), they found that the data could not reject this hypothesis.
Therefore, Small and Van Dender estimated the rebound effect as the elasticity of travel with
respect to fuel cost per mile.

   In contrast, Greene's research rejected the hypothesis of equal elasticities for gasoline prices
and fuel efficiency.  In spite of this result, Greene also tested Small and Van Dender's
formulation which allows the elasticity of fuel cost per  mile to decrease with increasing per
capita income.  The results of estimation using national time series data confirmed the results
obtained by Small and Van Dender using a time series  of state level data. When using Greene's
preferred functional form, the projected rebound effect  is approximately 12 percent in 2008,  and
drops to 10 percent in 2020 and to nine percent in 2030.

   Since there has been little variation in fuel economy in the data over time,  isolating the impact
of fuel economy on VMT can be difficult using econometric analysis of historical data.
Therefore, studies that estimate the rebound effect using time-series data often examine the
impact of gasoline prices on VMT, or the combined impact of both gasoline prices and fuel
economy on VMT, as discussed above.  However, these studies may overstate the potential
impact of the rebound effect resulting from this rule,  if people are more responsive to changes in
fuel price than the variable directly of interest, fuel economy.

   There is some evidence in the literature that consumers are more responsive to an increase in
prices than to a decrease in prices. At the aggregate level, Dargay and Gately (1997) and
H p. 15, Greene, D., Rebound 2007: Analysis of U.S. light-duty vehicle travel statistics. Energy Policy (2010),
  doi:10.1016/j.enpol.2010.03.083.
                                              10-14

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

Sentenac-Chemin (2012)20 have provide some evidence that demand for transportation fuel is
asymmetric.  In other words, given the same size change in prices, the response to a decrease in
gasoline price is smaller than the response to an increase in gasoline price.  Gately (1993)21 has
shown that the response to an increase in oil prices can be on the order of five times larger than
the response to a price decrease.  Furthermore, Dargay and Gately and Sentenac-Chemin also
find evidence that consumers respond more to a large shock than a small, gradual change in fuel
prices.  Since these standards would decrease the cost of driving gradually over time, it is
possible that the rebound effect would be much smaller than some of the historical estimates
included in the literature. Greene also notes that the resultant data from such gradual changes
could make discernment of such an effect difficult.

10.4.3 Review of Recent Literature on LDV Rebound since the 2012 Final Rule

   A number of recent studies examining LDV rebound effects have been undertaken since
EPA/NHTSA's review of the LDV rebound literature for 2012 final rule. Only a limited amount
of work has been conducted to examine the rebound effect of electric vehicles so most of the
studies of light-duty vehicle rebound effects focus on a change in gasoline prices. Below is a
brief summary of the results of these recent studies.

   Using data on household characteristics and vehicle use from the 2009 Nationwide Household
Transportation Survey (NHTS), Su (2012)22 analyzes the effects of locational and demographic
factors on household vehicle use, and investigates how the magnitude of the rebound effect
varies with vehicles' annual use.  Using variation in the fuel economy and per-mile cost of and
detailed controls for the demographic,  economic, and locational characteristics of the households
that owned them (e.g., road and population density) and each vehicle's main driver (as identified
by survey respondents), the author employs specialized regression methods to capture the
variation in the rebound effect across ten different categories of vehicle use.

   Su estimated that the overall rebound effect for all vehicles in the sample averaged 13 percent,
and that its magnitude varied from 11-19 percent among the ten different categories of annual
vehicle use. The smallest rebound effects were estimated for vehicles at the two extremes  of the
distribution of annual use - those driven comparatively little, and those used most intensively —
while the largest estimated effects applied to vehicles that were driven slightly more than
average. Controlling for the possibility that high-mileage drivers respond to the increased
importance of fuel costs by choosing vehicles that offer higher fuel economy narrowed the range
of Su's estimated rebound effects slightly (to 11-17 percent), but did not alter the finding that
they are smallest for lightly- and heavily-driven vehicles and largest for those with slightly above
average use.

   Linn (2013)23 also uses the 2009 NHTS  to develop a linear regression approach to estimate
the relationship between the VMT of vehicles belonging to each household and a variety of
different factors: fuel costs, vehicle characteristics other than fuel economy (e.g., horsepower,
the overall "quality" of the vehicle),  and household characteristics (e.g., age, income). Linn
reports a fuel economy rebound effect with respect to VMT of between 20-40 percent.

   One interesting result of the study is that when the fuel efficiency of all vehicles increases,
which would be the long-run effect of rising fuel efficiency standards, two factors have opposing
effects on the VMT of a particular vehicle.  First, VMT increases when that vehicle's fuel
efficiency increases.  But the increase in the fuel efficiency of the household's other vehicles
                                              10-15

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

causes the vehicle's own VMT to decrease.  Since the effect of a vehicle's own fuel efficiency is
larger than the other vehicles' fuel efficiency, VMT increases if the fuel efficiency of all vehicles
increases proportionately.  Linn also finds that VMT responds much more strongly to vehicle
fuel economy than to gasoline prices, which is at variance with the Hymel et al. and Greene
results discussed above.

   Like Su and Linn, Liu et al. (2014)24 also employed the 2009 NHTS to develop an elaborate
model of an individual household's choices about how many vehicles to own, what types and
ages of vehicles to purchase, and how much combined driving to do using all of them. Their
analysis used a complex mathematical formulation and statistical methods to represent and
measure the interdependence among households' choices of the number, types, and ages of
vehicles to purchase, as well as how intensively to use them.

   Liu et al. employed their model to simulate variation in households' total vehicle use to
changes in their income levels, neighborhood characteristics, and the per-mile  fuel cost of
driving averaged over all vehicles each household owns. The complexity of the relationships
among the number of vehicles owned, their specific types and ages, fuel economy levels, and use
incorporated in their model required them to measure these effects by introducing variation in
income, neighborhood attributes, and fuel costs, and observing the response of households'
annual driving. Their results imply a rebound effect of approximately 40 percent in response to
significant (25-50 percent) variation in fuel costs, with almost exactly symmetrical responses to
increases and declines.

   Frondel and Vance (2013)25 use panel estimation methods and household diary travel data
collected in Germany between 1997 and 2009 to identify an estimate of a private transport
rebound value. The study focuses on single-car households that did not change their car
ownership over the maximum three years each household was surveyed.  Failing to reject the
null hypothesis of a symmetric price response, they find  a rebound effect for single-vehicle
households of 46-70 percent (though we discuss further  below the limitations in applying
findings of studies from other countries to U.S. rebound).

   Gillingham (2014)26 analyzed variation in the use of more than five million new vehicles
purchased in California during the years 2001-03 over the first several years of their lifetimes,
focusing particularly on the response of buyers' use of new vehicles to geographic and temporal
variation in fuel prices.  His  sample consists predominantly of personal vehicles (87 percent), but
also includes some purchased by businesses, rental car companies, and government. He
estimates the effect of differences in the average of monthly fuel prices on their monthly average
vehicle use over the time - at a county level, since being purchase - focusing his analysis on
vehicles that have been purchased new and have been in service for six to seven years. The
author also explores how the effect of fuel prices on vehicle use varies with vehicle use, buyer
type and household income.

   Gillingham relies exclusively on the effect of variation in fuel prices and does not involve
vehicles' fuel economy. He  reports an overall average effect of fuel prices on vehicle use that
corresponds to a rebound effect of 22 percent, rising to 23 percent when he controls for the
potential effect of gasoline demand on its retail price.  He finds little evidence  of variation  in the
rebound effect among buyer types.  Based on the nature of his data and estimation procedure, he
interprets his estimates as implying that vehicle use responds fully to changes in fuel prices after
approximately two years.
                                              10-16

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

   Gillingham's results suggest that the vehicle-level responsiveness to fuel price increases with
income.  Gillingham hypothesizes that the increase in the per-vehicle rebound effect with higher
incomes may relate to wealthier households having more discretionary driving or switching
between flying and driving.  Alternatively, wealthier households tend to own more vehicles and
it is possible that within-household switching of vehicles to other more efficient vehicles in the
household may account for the greater responsiveness at higher income levels.

   Hymel and Small (2015)27 revisit the simultaneous equations methodology of Small and Van
Dender (2007) and Hymel, Small and Van Dender (2010) to see whether their previous estimates
of the VMT rebound effect have changed by adding  in more recent data from the late 2000 time
period (e.g., 2005-2009).  Consistent with previous results, the VMT rebound effect declines
with increasing income and urbanization, and it increases with increasing fuel cost. By far the
most important of these sources of variation is income, whose effect is large enough to greatly
reduce the projected rebound effect for time periods  of interest to current policy decisions. The
best estimate of the long-run light-duty vehicle rebound effect over the years 2000-2009 is 17.8
percent, when evaluated at average values of income, fuel cost, and urbanization in the U.S.
during that time period.

   The recent study by Hymel and Small also finds a strengthening of the VMT rebound effect
for the years 2003-2009 compared to the results for time periods from their previous research,
suggesting that some additional unaccounted for factors have increased the rebound effect.
Three potential factors are hypothesized to have caused the upward shift in the VMT rebound
effect in the 2003-2009 time period: (1) media coverage, (2) price volatility, and (3) asymmetric
response to price changes.1 It should be noted that the while media coverage and volatility are
important to understand the rebound effect based upon fuel prices, they may not be as relevant to
the rebound effect due to fuel efficiency.  These results show strong evidence of asymmetry in
responsiveness to price increases and decreases. Results suggest that a rebound adjustment to
fuel price rises takes place quickly; the rebound response elasticity is large in the year of,  and the
first year following, a price rise, then diminishes to a smaller value.  The rebound response to
price decreases occurs more slowly.

   Hymel and Small find that there is an upward shift in  the rebound effect of roughly 2.5 to 2.8
percentage points starting in 2003. Results suggest that the media coverage and volatility
variables may explain about half of the upward shift in the  LDV rebound effect in the  2003-2009
time period. Nevertheless, these influences are small enough in magnitude that they do not fully
offset the downward trend in VMT response elasticities due to higher incomes and other factors.
Hence, even assuming that the variables retain their 2003-2009 values into the indefinite  future,
they would not prevent a further diminishing of the magnitude of the rebound effect if incomes
continue to grow at anything like historic rates.

   West et al.  (2015)28 attempt to estimate the VMT  rebound effect using household level data
from Texas using a discontinuity in the eligibility requirements for the 2009 U.S. "Cash for
1 The media coverage variable is measured by constructing measures of media coverage based upon gas-price related
  articles appearing in the New York Times newspaper. Using the ProQuest historical database, they tally the
  annual number of article titles containing the words gasoline (or gas) and price (or cost). They then form a
  variable equal to the annual fraction of all New York Times articles that are gas-price-related. This fraction
  ranged from roughly 1/4000 during the 1960s to a high of 1/500 in 1974.
                                              10-17

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

Clunkers" program, which incentivized eligible households to purchase more fuel-efficient
vehicles. Households that owned "clunkers" with a fuel economy of 18 miles per gallon (MPG)
or less were eligible for the subsidy, while households owning clunkers with an MPG of 19 or
more were ineligible. The empirical strategy of the paper is to compare the fuel economy of
vehicle purchases and subsequent vehicle miles traveled of "barely eligible" households to those
households who were "barely ineligible."

   The paper finds a meaningful discontinuity in the fuel economy of new vehicles purchased by
Cash for Clunker-eligible households relative to ineligible households. Those authors report that
the increases in fuel economy realized by households who scrapped low fuel economy vehicles
in response to the substantial financial incentives offered under the federal "Cash for Clunkers"
program were not accompanied by increased use of the higher-MPG replacement vehicles they
purchased because of the vehicle's other attributes. Households chose to buy cheaper, smaller
and lower-performing vehicles. As a result, they did not drive any additional miles after the
purchase of the fuel efficient vehicle.  They conclude there is no evidence of a rebound effect in
response to improved fuel economy from the Cash for Clunkers program.

   It may be difficult to generalize the VMT response from the Cash for Clunkers program to a
program for LDV GHG/fuel  economy standards. Throughout this and all previous analyses of
the likely effects of federal regulations to require increased fuel economy and reduce vehicles'
GHG emissions, EPA and NHTSA have stressed that manufacturers can achieve the required
improvements without compromising the performance, passenger-, cargo-carrying, and towing
capacity, safety, or other attributes affecting the utility buyers and owners derive from the
vehicles they choose to purchase. The Cash for Clunkers program was a one-time program for a
fixed fleet of existing vehicles with specific characteristics.  Their study may not provide useful
implications about the likely  response of vehicle use to required increases in fuel economy that
are achieved through temporary incentive programs offered during recessions.

   More recently, De Borger et al. (2016)29 analyze the response of vehicle use to changes in  fuel
economy among a sample of nearly 350,000 Danish households owning a single vehicle, of
which almost one-third replaced it with a different model  sometime during the period from 2001
to 2011.  By comparing the change in households' driving from the early years of this period to
its later years among those who replaced their vehicles during the intervening period to that
among households who kept  their original vehicles, the authors claim to isolate the effect of
changes in fuel economy on vehicle use from those of other factors. Their data allow them to
control for the effects of important household characteristics and vehicle features other than fuel
economy on vehicle use. They use complex statistical methods to account for the fact that some
households replacing their vehicles may have done so in anticipation of changes in their driving
demands (rather than the reverse), as well as for the possibility that some households who
replaced their cars may have  done so because their driving behavior was more sensitive to fuel
prices than other households.

   De Borger et al. measure the rebound effect from the change in households' vehicle use in
response to changes in fuel economy that are a consequence of their decisions to replace the
vehicles they owned previously. Thus they are able to directly estimate the fuel economy
rebound effect itself, in contrast to other research that relies on indirect measures.  Their
preferred estimates span a very narrow range - from 8-10 percent - and vary only minimally in
response to different statistical  estimation procedures. They also vary little depending on
                                             10-18

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

whether the data sample is restricted to households that replaced their vehicles, in which case the
rebound effect is identified exclusively by their responses to changes in fuel economy of varying
magnitudes, or also includes households that did not replace their vehicles, and is thus identified
partly by differences between their responses to varying fuel economy and changes in driving
among households with vehicles whose fuel economy remained unchanged. Finally, De Borger
et al. find no evidence that the rebound effect is smaller among lower-income households than
among their higher-income counterparts.

   Gillingham et al. (2016)30 undertake a summary and review of the general rebound literature
including, for example, rebound effects from LDVs as well as electricity used in stationary
applications. The literature suggests that differences in estimates of the rebound effect stem
from its varying definitions, as well as variation in the quality of data and empirical
methodologies used to estimate it. Gillingham et al. seek to clarify the definition of each of the
channels of the rebound effect and critically assess the state of the literature that estimates its
magnitude.

   Gillingham et al. provide a list of what they consider to be relevant rebound elasticities that
can provide guidance to policymakers, with a focus on studies of overall demand or household-
level demand.  According to the authors, the studies are selected both because they are more
recent and use rigorous empirical methods such as panel data methods, experimental designs,
and quasi-experimental approaches.

   Of the selected studies, four focus on VMT elasticities for light-duty vehicles in developed
countries. For the Frondel and Vance study (cited above), which reported a short-run elasticity of
VMT demand for Germany for the time period from 1997-2009, Gillingham et al. chose the 46
percent value/  Barla (2009)31 found  a short-run elasticity of VMT for Canada from 1990-2004
of eight percent. Gillingham (2014) (cited above) found a California medium-run new vehicle
elasticity of VMT demand for the time period 2001-2009 of 23 percent.  Small and Van Dender
(2007) (cited previously) found a U.S. short-run elasticity of VMT demand for the time period
from 1966-2001 of roughly five percent.

   It is not clear whether studies of LDV VMT rebound estimates  for countries different from the
U.S. would provide estimates that are appropriate to the U.S. context. For example, European
countries have higher fuel prices and  more transit options, both factors which would possibly
produce a VMT rebound effect that is higher than in the U.S. The agencies are planning to
undertake an updated literature review of recent studies on the rebound effect for LDVs.

10.4.4 Basis for Rebound Effect Used in the Draft  TAR

   As the preceding discussion indicates, there is a wide range of estimates for both the historical
magnitude  of the rebound effect and its projected future value, and there is some evidence that
the magnitude of the rebound effect appears to be declining over time for those studies that look
at VMT time trends. The recent literature is mixed, with some studies supporting relatively
modest direct VMT rebound estimates and other studies suggesting a higher rebound effect.
Some of these studies come to these varied conclusions despite using the same dataset. EPA and
NHTSA use a single point estimate for the direct VMT rebound effect as an input to the
1 Gillingham et al. believe that this value is derived by more successfully holding exogenous factors constant in the
  Frondel and Vance study.
                                              10-19

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

agencies' analyses, although a range of estimates can be used to test the sensitivity to uncertainty
about its exact magnitude. Based on a combination of historical estimates of the rebound effect
and more recent analyses, an estimate of 10 percent for the rebound effect is used for evaluating
the MY2022-2025 standards in this Draft TAR (i.e., we assume a 10 percent decrease in fuel
cost per mile from the standards would result in a 1 percent increase in VMT).

   As Table 10.7, Table 10.8, and Table 10.9 indicate, the 10 percent figure is on the low end of
the range reported in previous research.  Recent research by Small, Hymel and Van Dender, and
Greene reports evidence that the magnitude of the rebound effect is likely to be declining over
time as household incomes rise which would be consistent with Gillingham's (2014) results
showing that individual-vehicle rebound increases with household income. The values that are
more applicable to quantifying the impact of these standards are values based on overall
aggregate rebound effects. West and Pickrell, Su, Linn and Liu et al., each using NHTS 2009
data, find rebound effect estimates varying from 11 percent to 40 percent.

   Gillingham et al. (2016) cite four studies that focus on VMT elasticities for light-duty vehicles
in developed countries. Two of the four studies (for the U.S. and Canada) have VMT elasticity
values below the 10 percent figure. The study for California has per-vehicle rebound value of 23
percent, and does not reflect the reduced use of other vehicles in multi-vehicle household fleets.
A study for Germany has a considerably higher value, roughly 46 percent.  A recent study by De
Borger at al. found a rebound value in the range of 10 percent for Denmark. As noted
previously, it is not clear whether studies of VMT LDV rebound estimates for countries different
from the U.S. would provide estimates that are appropriate to the U.S. context.

   Most of the studies reviewed use changes in fuel prices or fuel cost/mile to derive estimates of
the VMT rebound effect instead of using the actual variable of interest, changes in fuel economy,
and its impact on VMT. It is not clear how reliable the use of changes in fuel prices/fuel costs
are in attempting to estimate the impacts of changes in fuel economy on VMT.

   As mentioned above, for the reasons described in Section 10.4.2, historical estimates of the
rebound effect may overstate the effect of a gradual decrease in the cost of driving due to the
standards.  As a consequence, a value on the low end of the historical estimates is likely to
provide a more reliable estimate of its magnitude during the period spanned by the analysis of
the impacts of the MYs 2022-2025 standards. Studies which produce an aggregate measure of
the rebound effect are most applicable to estimating the overall VMT effects of the LDV
standards.  The 10 percent estimate lies  at the bottom of the 10-30 percent range of estimates for
the historical, aggregate rebound effect in most research, and at the upper end of the 5-10
percent range of estimates for the future rebound effect reported in the relatively  recent studies
by Small, Hymel and Van Dender and Greene.  Both Greene and Small, Hymel and  Van Dender
find that the rebound effect decreases as household incomes rise. As incomes rise, the value of
time spent driving becomes a larger fraction of total travel costs so that vehicle use becomes less
responsive to variations in fuel costs.  Since the AEO 2015 projects that household incomes will
be rising throughout the analysis  period, the agencies believe that it is appropriate to factor in
studies that account for income on the rebound effect.  In summary, the 10 percent value was not
derived from a single point estimate from a particular study, but instead represents a reasonable
compromise between historical estimates of the rebound effect and forecasts of its projected
future value, based on an updated review of the literature on this topic.
                                              10-20

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

 10.5  Energy Security Impacts

   The National Program is designed to require improvements in the fuel economy of light-duty
 vehicles and, thereby, reduce fuel consumption and GHG emissions. In turn, the program helps
 to reduce U.S. petroleum imports.  A reduction of U.S. petroleum consumption and imports
 reduces both financial and strategic risks caused by potential sudden disruptions in global oil
 supply,  thus increasing U.S. energy security.  This section summarizes EPA's estimates of U.S.
 oil import reductions and energy security benefits of the GHG/fuel economy vehicle standards
 for model years 2022-2025.

 10.5.1 Implications of Reduced Petroleum Use on U.S. Imports

   U.S.  energy security is generally considered as the continued availability of energy sources at
 an acceptable, stable price. Most discussion of U.S. energy security revolves around the topic of
 the economic costs of U.S. dependence on oil imports. While the U.S. has reduced its
 consumption and increased its production of oil in recent years, it still relies on oil from
 potentially unstable sources outside of the U.S. and the U.S. oil price will remain tightly linked
 to the global oil market. In addition, oil exporters with a large share of global production have
 the ability to raise the price of oil by exerting the monopoly power associated with a cartel, the
 Organization of Petroleum Exporting Countries (OPEC), to restrict oil supply relative to demand.
 These factors contribute to the vulnerability of the U.S. economy to episodic oil shocks to either
 the global supply of oil or world oil price spikes.

   In 2014, U.S. expenditures for imports of crude oil and petroleum products, net of revenues
 for exports, were $178 billion and expenditures on both imported oil and domestic petroleum and
 refined products totaled $469 billion (2013$) (see Figure 10.2).32  Recently, as a result of strong
 growth in domestic  oil production mainly from tight shale formations, U.S. production of oil has
 increased while U.S. oil imports have decreased. For example, from 2012 to 2015, domestic oil
 production increased by 44 percent while oil imports decreased by 24 percent.33 While oil
 import costs have declined since 2011, total oil expenditures (domestic and imported) remained
 near historical highs through 2014. Post-2015 oil expenditures are projected (AEO 2015) to
 remain between double and triple the inflation-adjusted levels experienced by the U.S. from
 1986 to 2002.
                                             10-21

-------
                          Economic and Other Key Inputs Used in the Agencies' Analyses
                             U.S. Expenditures on Crude Oil
         soo
                                                                           2010    2015
                Figure 10.2 U.S. Expenditures on Crude Oil from 1970 through 201534

   Focusing on changes in oil import levels as a source of vulnerability has been standard
practice in assessing energy security in the past, but given current market trends both from
domestic and international levels, adding changes in consumption of petroleum to this
assessment may provide better information about U.S. energy security. The major mechanism
through which the economy sustains harm due to fluctuations in the (world) energy market is
through price, which itself is leveraged through both imports and consumption. However, the
United States, may be increasingly insulated from the physical effects of overseas oil disruptions,
though the price impacts of an oil disruption anywhere will continue to be transmitted to U.S.
markets. As of 2015, Canada accounted for 63 percent of U.S. net oil imports of crude oil and
petroleum products.35  The implications of the U.S. becoming a significant petroleum producer
have yet to  be discerned in the literature, but it can be anticipated that this will have some impact
on energy security.

   In 2010, just over 40 percent of world oil  supply came from OPEC nations. The AEO 201536
projects that this  share will stay high; dipping slightly from 37 percent by 2020 and then rising
gradually to over 40 percent by 2035 and thereafter.  Approximately 30 percent of global supply
is from Middle East and North African countries alone, a share that is also expected to grow  over
the long term. Measured in terms of the share of world oil resources or the share of global oil
export supply, rather than oil production, the concentration of global petroleum resources in
OPEC nations is  even larger. As another measure of concentration,  of the 137
countries/principalities that export either crude or refined products, the top 12 have recently
accounted for over 55 percent of exports.37 Eight of these countries are members of OPEC, and
                                             10-22

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

a ninth is Russia.K  In a market where even a 1-2 percent supply loss can raise prices noticeably,
and where a 10 percent supply loss could lead to an unprecedented price shock, this regional
concentration is of concern.L  Historically, the countries of the Middle East have been the source
of eight of the ten major world oil disruptions38, with the ninth originating in Venezuela, an
OPEC country, and the tenth being Hurricanes Katrina and Rita.

   EPA uses a processed combination of the MOVES and OMEGA models,  and DOT uses the
CAFE model, to estimate the reductions in U.S. fuel consumption due to the  LDV National
Program. Based on a detailed analysis of differences in U.S. fuel consumption, petroleum
imports, and imports of petroleum products, the agencies estimate that approximately 90 percent
of the reduction in fuel  consumption resulting from adopting improved GHG emission and fuel
economy standards is likely to be reflected in reduced U.S. imports of crude  oil and net imported
petroleum products.39  Thus, on balance, each gallon of fuel saved as a consequence of the LDV
GHG/fuel economy standards is anticipated to reduce total U.S.  imports of petroleum by 0.9
gallons.  Based upon the fuel savings estimated by the models and the 90 percent oil import
factor, the reduction in U.S. oil imports from the 2022-2025 LDV standards  are estimated for
selected years from 2022 to 2050 (in millions of barrels per day  (MMBD) in  Table 10.10 below.
For comparison purposes, Table 10.10 also shows U.S. oil exports/imports, U.S. net product
imports and U.S. net crude/product imports in selected years from 2022 to 2040, as projected by
DOE in the Annual Energy Outlook 2015 Reference Case. U.S. Gross Domestic Product (GDP)
is projected to grow by roughly 55 percent over the same time frame (e.g., from 2022 to 2040) in
the AEO 2015 projections.
K The other three are Norway, Canada, and the EU, an exporter of product.
L For example, the 2005 Hurricanes Katrina/Rita and the 2011 Libyan conflict both led to a 1.8 percent reduction in
  global crude supply. While the price impact of the latter is not easily distinguished given the rapidly rising post-
  recession prices, the former event was associated with a 10-15 percent world oil price increase. There are a range
  of smaller events with smaller but noticeable impacts. Somewhat larger events, such as the 2002/3 Venezuelan
  Strike and the War in Iraq, corresponded to about a 2.9 percent sustained loss of supply, and was associated with
  a 28 percent world oil price increase. Compiled from EIA oil price data, IEA2012 [IEA Response System for Oil
  Supply Emergencies
  (http://www.iea.org/publications/freepublications/publication/EPPD Brochure English 2012 02.pdf)  [EPA-
  HQ-OAR-2014-0827-0573] See table on P. 11.and Hamilton 2011 "Historical Oil Shocks,"
  (http://econweb.ucsd.edu/~ihamilto/oil history.pdfin rEPA-HQ-OAR-2014-0827-05981 Routledge Handbook of
  Major Events in Economic History*, pp. 239-265, edited by Randall E. Parker and Robert Whaples, New York:
  Routledge Taylor and Francis Group, 2013).
                                               10-23

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

 Table 10.10 Projected Trends in U.S. Oil Exports/Imports, and U.S. Oil Import Reductions Resulting from
         the Program in Selected Years from 2022 to 2050 (Millions of barrels per day (MMBD)
Year
2022
2023
2024
2025
2030
2035
2040
2050
U.S. Oil
Exports
0.63
0.63
0.63
0.63
0.63
0.63
0.63
**
U.S. Oil Imports
6.47
6.61
6.63
6.72
7.07
7.98
8.21
**
U.S. Net Product
Imports*
-3.08
-3.15
-3.20
-3.24
-3.56
-3.94
-4.26
**
U.S. Net Crude &
Product Imports
2.76
2.83
2.85
2.85
2.88
3.41
3.32
**
U.S. Reductions from
Oil Imports
0.019
0.055
0.106
0.169
0.420
0.685
0.880
1.119
 Notes:
 * Negative U.S. Net Product Imports imply positive exports.
 **The AEO 2015 only projects energy market and economic trends through 2040.
10.5.2 Energy Security Implications

   In order to understand the energy security implications of reducing U.S. oil imports, EPA has
worked with Oak Ridge National Laboratory (ORNL), which has developed approaches for
evaluating the social costs and energy security implications of oil use. The energy security
estimates provided below are based upon a methodology developed in a peer-reviewed study
entitled, "The Energy Security Benefits of Reduced Oil Use, 2006-2015", completed in March
2008. This ORNL study is an updated version of the approach used for estimating the energy
security benefits of U.S. oil import reductions developed in a 1997 ORNL Report.40  This
approach has been used to estimate energy security benefits for the LDV GHG/fuel economy
standards (2012-2016; 2017-2025) and the HDV GHG/fuel economy standards Phase I (2014-
2018)/Phase II proposal (2018 and later). For EPA and NHTSA rulemakings, the ORNL
methodology is updated periodically to account for forecasts of future energy market and
economic trends reported in the U.S. EIA's AEO.

   When conducting this analysis, ORNL considered the full cost of importing petroleum into
the U.S. The full economic cost is defined to include two components in addition to the
purchase price of petroleum itself.  These are: (1) the higher costs for oil imports resulting from
the effect of U.S. demand on the world oil price (i.e., the "demand" or "monopsony" costs); and
(2) the risk of reductions in U.S. economic output and disruption to the U.S. economy caused by
sudden disruptions in the supply of imported oil to the U.S. (i.e.,  macroeconomic
disruption/adjustment costs).

   For this Draft TAR, ORNL updated the energy security premiums by incorporating the most
recent oil price forecast and energy market trends, particularly regional oil supplies and
demands, from the AEO 2015 into its  model.41 Below are ORNL energy security premium
estimates for the selected years from 2022 to 2050,M as well as a breakdown of the components
M AEO 2015 forecasts energy market trends and values only to 2040. The post-2040 energy security premium
  values are assumed to be equal to the 2040 estimate.
                                             10-24

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

of the energy security premiums for each year. The components of the energy security premiums
and their values are discussed below.

       Table 10.11 Energy Security Premiums in Selected Years from 2022 to 2050, (2013$/Barrel)*
Year
(range)
2022
2023
2024
2025
2030
2035
2040
2050
Monopsony
(Range)
$2.31
($0.69 -$3.81)
$2.33
($0.71 -$3.92)
$2.40
($0.73 -$4.03)
$2.59
($0.76 -$4.14)
$2.83
($0.83 -$4.56)
$3.78
($1.10 -$6.17)
$4.09
($1.19 -$6.67)
$4.09
($1.19 -$6.67)
Avoided Macroeconomic
Disruption/ Adjustment Costs
(Range)
$5.69
($2.67 - $9.44)
$5.75
($2.75 -$9.70)
$5.89
($2.83 -$9.96)
$6.30
($2.92 -$10.22)
$7.26
($3.40 -$11.73)
$8.47
($3.99 -$13.58)
$9.61
($4.54 -$15.39)
$9.61
($4.54 -$15.39)
Total Mid-Point
(Range)
$7.99
($4.81 -$11.81)
$8.09
($4.94 -$12.15)
$8.29
($5.08-12.49)
$8.89
($5.22 -$12.83)
$10.09
($5.90 -$14.59)
$12.26
($7.28 -$17.59)
$13.69
($8. 12 -$19.64)
$13.69
($8. 12 -$19.64)
   Note:
   * The top values in each cell are the midpoints; the values in parentheses are the 90 percent confidence intervals.
10.5.2.1
Effect of Oil Use on the Long-Run Oil Price
   The first component of the full economic costs of importing petroleum into the U.S. follows
from the effect of U.S. import demand on the world oil price over the long-run.  Because the
U.S. is a sufficiently large purchaser of global oil supplies, its purchases can affect the world oil
price. This monopsony power means that increases in U.S. petroleum demand can cause the
world price of crude oil to rise,  and conversely, that reduced U.S. petroleum demand can reduce
the world price of crude oil.  Thus, one benefit of decreasing U.S. oil purchases due  to
improvements in the fuel  economy of light-duty vehicles is the potential decrease in the crude oil
price paid for all crude oil purchased.

   A variety of oil market and economic factors have  contributed to lowering the estimated
monopsony premium compared to monopsony premiums cited in the agencies' previous 2017-
2025 LDV GHG/fuel economy rulemakings. Three principal factors contribute to lowering the
monopsony premium: lower world oil prices, lower U.S. oil imports, and less responsiveness of
world oil prices to changes in U.S. oil demand. Below we consider differences in oil market
trends by comparing projections developed using the AEO 2012 (Early  Release) and the AEO
2015. The AEO 2012 (Early Release) was used for the 2012 final LDV rule and the AEO 2015
is being used for this Draft TAR assessment, so the comparison gives a  snapshot of how oil and
energy markets have changed since the 2012 final rule.

   The result of the comparison is that there has been  a general downward revision in world oil
price projections in the near term (e.g., a 35 percent reduction in 2020) and a sharp reduction in
projected U.S. oil imports in the near term due to increased U.S. supply (i.e., a 60 percent
                                             10-25

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

reduction in U.S. oil imports by 2020 and a 58 percent reduction in 2025).  Over the longer term,
based upon the AEO 2015 projections, oil's share of total U.S. imports is projected to gradually
increase after 2020 but still remain 50 percent below the AEO 2012 (Early Release) projected
level in 203 5.

   Currently some OPEC countries (e.g., Saudi Arabia) are increasing oil supply in an attempt to
price more expensive marginal suppliers, like the U.S., out of the market and regain market
share, exacerbating the worldwide oil supply glut which has resulted in lowering the world oil
price further. Lower world oil prices currently may reduce both production from existing
domestic oil resources and investment in new domestic oil sources increasing U.S. oil import
levels in the intermediate term.

   Another factor influencing  the monopsony premium is that U.S. demand on the global oil
market is projected to decline, suggesting diminished overall influence and some reduction in the
influence of U.S. oil demand on the world price of oil.  This is a result of the U.S. being a
smaller fraction of total world oil demand. Outside of the U.S., projected OPEC supply in the
AEO 2015 remains roughly steady as a share of world oil supply compared to the AEO 2012
(Early Release).  OPEC's share of world oil supply outside of the U.S. actually increases
slightly. Since OPEC  supply is estimated to be more price sensitive than non-OPEC supply, this
means that AEO 2015  projected world oil supply is slightly more responsive to changes in U.S.
oil demand.  Together, these factors suggest that changes in U.S. oil import reductions have a
somewhat smaller effect on the long-run world oil price than changes based on AEO 2012 (Early
Release) estimates.

   These changes in oil price and import levels lower the monopsony portion of energy security
premium since this portion of the security premium is related to the change in total U.S. oil
import costs that is achieved by a marginal reduction in U.S oil imports.  Since both the price and
the quantity of oil imports are lower, the monopsony premium component estimated in this
assessment is 60-75 percent lower over the years 2025-2040 than the estimates based upon the
AEO 2012 (Early Release) projections.

   The literature on the energy security for the last two decades has routinely combined the
monopsony  and the macroeconomic  disruption components when calculating the total value of
the energy security premium.  However, in the context of using a global value for the Social Cost
of Carbon (SCC) the question arises: how should the energy security premium be used when
some benefits from the rule, such as the benefits of reducing greenhouse gas emissions, are
calculated from a global perspective? Monopsony benefits represent avoided payments by U.S.
consumers to oil producers that result from a decrease in the world oil price as the U.S. decreases
its demand for oil. Although there is clearly an overall benefit to the U.S. when considered  from
a domestic perspective, the decrease  in price due to decreased demand in the U.S. also represents
a loss to oil producing countries, one of which is the U.S.

   Given the redistributive nature of this monopsony effect from a global perspective, it has been
excluded in the energy security benefits calculations in past rulemakings. In contrast, the other
portion of the energy security  premium, the avoided U.S. macroeconomic disruption and
adjustment cost that arises from reductions in U.S. petroleum imports, does not have offsetting
impacts outside of the U.S., and, thus, is included in the energy security benefits.  To summarize,
the agencies have included only the avoided macroeconomic disruption portion of the energy
security benefits to estimate the monetary value of the total energy security benefits.
                                             10-26

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

   There is disagreement in the literature about the magnitude of the monopsony component, and
its relevance for policy analysis. Brown and Huntington (2013)42, for example, argue that the
U.S.'s refusal to exercise its market power to reduce the world oil price does not represent a
proper externality, and that the monopsony component should not be considered in calculations
of the energy security externality.  However, they also note in their earlier discussion paper
(Brown and Huntington 2010)43 that this is a departure from the traditional energy security
literature, which includes sustained wealth transfers associated with stable but higher-price oil
markets.

   On the other hand, Greene (2010)44 and others in prior literature (e.g., Toman 1993)45 have
emphasized that the monopsony cost component is policy-relevant because the world oil market
is non-competitive and strongly  influenced by cartelized and government-controlled supply
decisions. Thus, while sometimes couched as an externality, Greene  notes that the monopsony
component is best viewed as stemming from a completely different market failure than an
externality (Ledyard 2008)46, yet still implying marginal social costs to importers.

   The Council on Foreign Relations47 (i.e., "the Council") (2015) recently released a discussion
paper that assesses NHTSA's analysis of the benefits and costs of CAFE in a lower-oil-price
world. In this paper, the Council notes that while NHTSA cites the monopsony effect of the
CAFE standards for 2017-2025, NHTSA does not include it when calculating the cost-benefit
calculation for the rule. The Council argues that the monopsony benefit should be included in
the CAFE cost-benefit analysis and that including the monopsony benefit is more consistent with
the legislators' intent in mandating CAFE standards in the first place.

   The recent National Academy of Science (NAS 2015) Report, "Cost, Effectiveness and the
Deployment of Fuel Economy Technologies for Light-Duty Vehicles,"48 suggests that the
agencies' logic about not accounting for monopsony benefits is inaccurate. According to the
NAS, the fallacy lies in treating  the two problems, oil dependence and climate change, similarly.
According to the NAS, "Like national defense, it [oil dependence] is inherently adversarial (i.e.,
oil consumers against producers using monopoly power to raise prices).  The problem of climate
change is inherently global and requires global action. If each nation considered only the benefits
to itself in determining what actions to take to mitigate climate change, an adequate solution
could not be achieved. Likewise, if the U.S.  considers the economic harm its reduced petroleum
use will do to monopolistic oil producers it will not adequately address its oil dependence
problem.  Thus, if the United States is to solve both of these problems it must take full account of
the costs and benefits of each, using the appropriate scope for each problem."  Based upon the
assessment of the monopsony premium in the Council of Foreign Relations  and NAS reports, we
are seeking public input on whether it is appropriate to consider monopsony in the societal
costs/benefits of the National Program.

   There is also a question about the ability of gradual, long-term reductions, such as those
resulting from the LDV GHG/fuel economy standards, to reduce the world oil price in the
presence of OPEC's monopoly power. OPEC is currently the world's marginal petroleum
supplier, and could conceivably  respond to gradual reductions in U.S. demand with gradual
reductions in supply over the course of several years as the fuel savings resulting  from this
Program grow.  However, if OPEC opts for a long-term strategy to preserve its market share,
rather than maintain a particular price level (as they have done recently in response to increasing
U.S. petroleum production) reduced demand would create downward pressure on the global
                                             10-27

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

price.  The Oak Ridge analysis assumes that OPEC does respond to demand reductions by
reducing its supply over the long run, but there is still a price effect in the model because the
supply reduction only partially offsets the demand  reduction, enough to maintain supply share.
Under the mid-case behavioral assumption used in the premium calculations,  OPEC responds by
gradually reducing supply to maintain market share (consistent with the long-term self-interested
strategy suggested by Gately (2004, 2007)).49

   It is important to note that the decrease in global petroleum prices resulting from this Program
could spur increased  consumption of petroleum in  other sectors and countries, leading to a
modest uptick in GHG emissions outside of the U.S.  This increase in global fuel consumption
could offset some portion of the GHG reduction benefits associated with these standards. The
agencies have not quantified this increase in global oil consumption  or GHG emissions outside
the U.S. due to world oil price changes resulting from the standards. Recent research has
quantified this type of effect in the context of biofuel policies (e.g., Drabik and de Goiter
(2011)50; Rajagopal,  Hochman and Zilberman (2011)51; Thompson, Whistance, and Meyer
(2011))52, pipeline construction (Erickson and Lazarus (2014))53, and fuel economy policies
(Karplusetal., (2015)54).

   Quantifying resulting GHG emissions may be challenging because other fuels, with varying
GHG intensities, could be displaced from the increasing use of oil worldwide, particularly
outside of the transportation sector. For example, if a decline in the world oil price causes an
increase in oil use in  China, India, or another country's industrial sector, this increase in oil
consumption may displace natural gas usage.  Alternatively, the increased oil use could result in
a decrease  in coal used to produce electricity.  We  seek comment on whether it is appropriate to
quantify changes in net global oil consumption and to consider the resulting GHG emissions in
the societal costs/benefits of the Program.  In particular, we are taking comments on any robust
methodologies that could be used to look at these impacts, a discussion on the strengths and
weaknesses of these methodologies, estimates of own and cross-price elasticities of demand for
fossil fuels and their  relative importance, and the appropriate level of regional and  sectoral
resolution for such an analysis.

10.5.2.2       Macroeconomic Disruption Adjustment Costs

   The second component of the oil import premium, "avoided macroeconomic
disruption/adjustment costs," arises from the effect of oil imports on the expected cost of supply
disruptions and accompanying price increases. A sudden increase in oil prices triggered by  a
disruption in world oil supplies has two main effects: (1) it increases the costs of oil imports in
the short-run and (2)  it can lead to macroeconomic contraction, dislocation and Gross Domestic
Product (GDP) losses. For example, ORNL estimates the combined value of these two factors to
be $6.30/barrel when U.S. oil imports are reduced in 2025, with a range from $2.92/barrel to
$10.22^arrel of imported oil reduced.

   Since future disruptions in foreign oil supplies are an uncertain prospect, each of the
disruption cost components must be weighted by the probability that the supply of petroleum to
the U.S. will actually be disrupted. Thus, the "expected value" of these costs - the product of the
probability that a supply disruption will occur and the sum of costs from reduced economic
output and the economy's abrupt adjustment to sharply higher petroleum prices - is the relevant
measure of their magnitude.  Further, when assessing the energy security value of a policy to
reduce oil use, it is only the change in the expected costs of disruption that results from the
                                              10-28

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

policy that is relevant. The expected costs of disruption may change from lowering the normal
(i.e., pre-disruption) level of domestic petroleum use and imports, from any induced alteration in
the likelihood or size of disruption, or from altering the short-run flexibility (e.g., elasticity) of
petroleum use.

   With updated oil market and economic factors, the avoided macroeconomic disruption
component of the energy security premiums is somewhat lower compared to the avoided
macroeconomic disruption premiums used in the 2017-2025 LDV GHG/fuel economy
rule.  Factors that contribute to moderately lowering the avoided macroeconomic disruption
component are lower U.S. imports (slightly reducing the U.S.' global reliance on unstable
supplies), lower real oil prices and slightly smaller price increases during prospective
shocks.  Oil price levels are 0-29 percent lower over the 2025-2040 period, and the likely
increase in oil prices in the event of an oil shock are somewhat smaller, reflecting small increases
in the responsiveness of global oil supply to changes in the world price of oil. However, over the
2025-2040 period AEO 2015 projected domestic oil demand, and real GDP levels, are little
changed from AEO 2012 (Early Release). So oil remains an important input to the U.S.
economy. Overall, the avoided macroeconomic disruption component estimates for the oil
security premiums  are 4-28 percent lower over the period  from 2025-2040 based upon different
projected oil market and economic trends in the AEO 2015 compared to the AEO 2012 (Early
Release).
   There are several reasons why the avoided macroeconomic disruption premiums change only
moderately. One reason is that the projected macroeconomic sensitivity to oil price shocks is
held unchanged from the historical average levels used in multiple prior estimates,  since
projected U.S. oil consumption levels and the expenditures on oil in the U.S. economy remain at
comparatively high levels under both AEO 2012 (Early Release) and AEO 2015. Figure 10.3
below shows that under AEO 2015,  projected U.S. real annual oil expenditures continue to rise
after 2015 to over $800 billion (2013$) by 2035. The value share of U.S. oil use, labeled in the
Figure below as U.S. oil expenditures as share of GDP, remains at three percent even as the
economy grows, lower than the AEO 2012 (Early Release) projection of 4.4 percent declining to
3.5 percent.  The value share of oil use in the AEO 2015 is still projected to be above the full
historical average (2.8 percent for 1970-2010), and well above the historical levels observed
from  1985 to 2005  (1.9 percent).  A second factor is that oil disruption risks are little changed.
The two factors influencing disruption risks are the probability of global supply interruptions and
the world oil supply share from OPEC. Both factors are not significantly different from previous
forecasts of oil market trends.
                                             10-29

-------
                          Economic and Other Key Inputs Used in the Agencies' Analyses
                       Projected and Historical  U.S. Expenditures,
                           and Expenditure Share, on Crude Oil
          900
     85
                             D Domestic
                             D Imported

                             BUS Oil Expenditures as Share of GDP
           0
0%
            1970  1975  1980 1985 1990 1995  2000  2005  2010  2015  2020  2025  2030  2035

                                             Year
     Figure 10.3 Projected and Historical U.S. Expenditures, and Expenditure Share, on Crude Oil55

       The energy security costs estimated here follow the oil security premium framework,
which is well established in the energy economics literature. The oil import premium gained
attention as a guiding concept for energy policy around the time of the second and third major
post-war oil shocks (Bohi and Montgomery (1982), EMF (1982)56, Plummer (1982))57 provided
valuable discussion of many of the key issues related to the oil import premium as well as the
analogous oil stockpiling premium. Bohi and Montgomery (1982)58 detailed the theoretical
foundations of the oil import premium and established many of the critical analytic relationships
through their thoughtful analysis.  Hogan (1981)59 and Broadman and Hogan (1986, 1988)60
revised and extended the established analytical framework to estimate optimal oil import
premium with a more detailed accounting of macroeconomic effects.

   Since the original work on energy security was undertaken in the 1980's, there have been
several reviews on this topic. For example,  Leiby, Jones, Curlee and Lee (1997)61 provided an
extended review of the literature and issues regarding the estimation of the premium. Parry and
Darmstadter (2004)62 also provided an overview of extant oil security premium estimates and
estimated of some premium  components.

   The recent economics literature on whether oil shocks are the threat to economic stability that
they once were is mixed.  Some of the current literature asserts that the macroeconomic
component of the energy security externality is small. For example, the National Research
Council (2009) argued that the non-environmental externalities associated with dependence on
foreign oil are small, and potentially trivial.63  Analyses by Nordhaus (2007) and Blanchard and
Gali (2010) question the impact of more recent oil price shocks on the economy.64  They were
motivated by attempts to explain why the economy  actually expanded immediately after the last
shocks, and why there was no evidence of higher energy prices being passed on through higher
                                             10-30

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

wage inflation. Using different methodologies, they conclude that the economy has largely
gotten over its concern with dramatic swings in oil prices.

   One reason, according to Nordhaus, is that monetary policy has become more accommodating
to the price impacts of oil shocks. Another is that consumers have simply decided that such
movements are temporary, and have noted that price impacts are not passed on as inflation in
other parts of the economy. He also notes that real changes to productivity due to oil price
increases are incredibly modest,N and that the general direction  of the economy matters a great
deal regarding how the economy responds to a shock. Estimates of the impact of a price shock
on aggregate demand are insignificantly different from zero.

   Blanchard and Gali (2010) contend that improvements in monetary policy (as noted above),
more flexible labor markets, and lessening of energy intensity in the economy, combined with an
absence of concurrent shocks, all contributed to lessen the impact of oil shocks after 1980. They
find "...  the effects of oil price shocks have changed over time, with steadily smaller effects on
prices and wages, as well as on output and employment."65 In a comment at the chapter's end,
this work is summarized as follows: "The message of this chapter is thus optimistic in that it
suggests a transformation in U.S. institutions has inoculated the economy against the responses
that we saw in the past."

   At the same time, the implications of the "Shale Oil Revolution" are now being felt in the
international  markets, with current prices at four year lows.  Analysts generally attribute this
result in  part to the significant increase in supply resulting from U.S. production, which has put
liquid petroleum production roughly on par with Saudi Arabia.  The price decline is also
attributed to the sustained reductions in U.S. consumption and global demand growth from fuel
efficiency policies and previously high oil prices. The resulting decrease in foreign imports,
down to  about one-third of domestic consumption (from 60 percent in 2005, for example66),
effectively permits U.S. supply to act as a buffer against artificial or other supply restrictions (the
latter due to conflict or a natural disaster, for example).

   However, other papers suggest that oil shocks, particularly sudden supply shocks, remain a
concern. Both Blanchard and Gali's and Nordhaus work were based on data and analysis
through 2006, ending with a period of strong global economic growth and growing global oil
demand. The Nordhaus work particularly stressed the effects of the price increase from 2002-
2006 that were comparatively gradual  (about half the growth rate of the 1973 event and one-third
that of the  1990 event). The Nordhaus study emphasizes the robustness of the U.S. economy
during a time period through 2006. This time period was just before rapid further increases in
the price of oil and other commodities with oil prices more-than-doubling to over $130/barrel by
mid-2008,  only to drop after the onset of the largest recession since the Great Depression.

   Hamilton  (2012)67 reviewed the empirical literature on oil shocks and suggested that the
results are mixed, noting that some work (e.g. Rasmussen and Roitman (2011) finds less
evidence for  economic effects of oil shocks, or declining effects of shocks  (Blanchard and Gali
N In fact, "... energy-price changes have no effect on multifactor productivity and very little effect on labor
  productivity." Page 19. He calculates the productivity effect of a doubling of oil prices as a decrease of 0.11
  percent for one year and 0.04 percent a year for ten years. Page 5.  (The doubling reflects the historical
  experience of the post-war shocks, as described in Table 7.1 in Blanchard and Gali, pp. 380) [EPA-HQ-OAR-
  2014-0827-0567].
                                              10-31

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

2010), while other work continues to find evidence regarding the economic importance of oil
shocks. For example, Baumeister and Peersman (2011) found that an oil price increase had a
decreasing effect over time. But they note that with a declining price-elasticity of demand that a
given physical oil disruption would have a bigger effect on price and a similar effect on output as
in the earlier data. Hamilton observes that "a negative effect of oil prices on real output has also
been reported for a number of other countries, particularly when nonlinear functional forms have
been employed." Alternatively, rather than a declining effect, Ramey and Vine (2010) found
"remarkable stability in the response of aggregate real variables to oil shocks once we account
for the extra costs imposed on the economy in the 1970s by price controls and a complex system
of entitlements that led to some rationing and shortages."68

   Some of the recent literature on oil price shocks has emphasized that economic impacts
depend on the nature of the oil shock, with differences between price increases caused by sudden
supply loss and those caused by rapidly growing demand.  Most recent analyses of oil price
shocks have confirmed that "demand-driven" oil price shocks have greater effects on oil prices
and tend to have positive effects on the economy while "supply-driven" oil shocks still have
negative economic impacts (Baumeister, Peersman and Van Robays (2010)).69 A recent paper
by Kilian and Vigfusson (2014)70, for example, assigned a more prominent role to the effects of
price increases that are unusual, in the sense of being beyond range of recent experience.  Kilian
and Vigfusson also conclude that the difference in response to oil  shocks may well stem from  the
different effects of demand- and supply-based price increases: "One explanation is that oil price
shocks are associated with a range of oil demand and oil supply shocks, some of which stimulate
the U.S. economy in the short run and some of which slow down U.S. growth (see Kilian
(2009)).  How recessionary the response to an oil price shock is thus depends on the average
composition of oil demand and oil supply shocks over the sample  period."

   The general conclusion that oil supply-driven shocks reduce economic output is also reached
in a recently published paper by Cashin et al. (2014)71 for 38 countries from 1979-2011. "The
results indicate that the economic consequences of a supply-driven oil-price shock are very
different from those of an oil-demand shock driven by global economic activity, and vary for oil-
importing countries compared to energy exporters", and "oil importers [including the U.S.]
typically face a long-lived fall in economic activity in response to a supply-driven surge in oil
prices" but almost all countries see an increase in real output for an oil-demand disturbance.
Note that the energy security premium calculation in this analysis  is based  on price shocks from
potential future supply events only.

   By early 2016, world oil prices were sharply lower than in 2014. Future prices remain
uncertain, but sustained markedly lower oil prices can have mixed implications for U.S. energy
security. Under lower prices U.S. expenditures on oil consumption are lower, and the
expenditures are a less prominent component of the U.S. economy. But sustained lower oil
prices encourage greater oil consumption, and reduce the competitiveness of new U.S. oil
supplies and alternative fuels.  The AEO 2015 low-oil price  outlook, for example, projects that
by 2030 total U.S. petroleum supply would be  10 percent lower and imports would be 78 percent
higher than the AEO 2015 Reference Case. Under the low-price case, 2030 prices are 35 percent
lower, so that U.S. import expenditures are 16 percent higher.
                                             10-32

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

   A second potential proposed energy security effect of lower oil prices is increased instability
of supply, due to greater global reliance on fewer suppling nations,0 and because lower prices
may increase economic and geopolitical instability in some supplier nations.72'73'74 The
International Monetary Fund reported that low oil prices are creating substantial economic
tension for Middle East oil producers on top of the economic costs of ongoing conflicts, and
noted the risk that Middle East countries including Saudi Arabia could run out of financial assets
without a substantial change in policy.75 The concern raised is that oil revenues are essential for
some exporting nations to fund domestic programs and avoid domestic unrest.

   Finally, despite continuing uncertainty about oil market behavior and outcomes and the
sensitivity of the U.S. economy to oil shocks, it is generally agreed that it is beneficial to reduce
petroleum fuel  consumption from an energy security standpoint.  It is not just imports alone, but
both imports and consumption of petroleum from all sources and their role in economic activity,
that may expose the U.S. to risk from price shocks in the world oil price. Reducing fuel
consumption reduces the amount of domestic economic activity associated with a commodity
whose price depends on volatile international markets.

   The relative significance of petroleum consumption  and import levels for the macroeconomic
disturbances that follow from oil price shocks is not fully understood. Recognizing that
changing petroleum consumption will change U.S. imports, this assessment of oil costs focuses
on those incremental social costs that follow from the resulting changes in imports, employing
the usual oil import premium measure. The agencies request comment on any published data or
literature that could help inform how the agencies might attempt to incorporate the impact of
changes in oil consumption, rather than imports exclusively, into our energy security analysis.
Most helpful would be the provision of specific methodologies that could be utilized to estimate
quantitatively how changes in  oil  consumption patterns influence energy security.

10.5.2.3      Cost of Existing U. S. Energy Security Policies

   The last often-identified component of the full economic costs of U.S. oil imports are the
costs to the U.S. taxpayers of existing U.S. energy  security policies.  The two primary examples
are maintaining the Strategic Petroleum Reserve (SPR) and maintaining a military presence to
help secure a stable oil supply  from potentially vulnerable regions of the world.  The SPR is the
largest  stockpile of government-owned emergency crude oil in the world. Established in the
aftermath of the 1973/1974 oil embargo, the SPR provides the U.S. with a response option
should  a disruption in commercial oil supplies threaten the U.S. economy. It also allows the U.S.
to meet part of its International Energy Agency obligation to maintain emergency oil stocks, and
it provides a national defense fuel reserve. While the costs for building and maintaining the SPR
are more clearly related to U.S. oil use and imports, historically these costs have not varied in
response to changes in U.S. oil import levels. Thus, while the effect of the SPR in moderating
price shocks is  factored into the ORNL analysis, the cost of maintaining the SPR is excluded.
0 Fatih Birol, Executive Director of the International Energy Agency warn that prolonged lower oil prices would
  trigger energy-security concerns by increasing reliance on a small number of low-cost producers "or risk a sharp
  rebound in price if investment falls short." "It would be a grave mistake to index our attention to energy security
  to changes in the oil price," Birol said. "Now is not the time to relax. Quite the opposite: a period of low oil prices
  is the moment to reinforce our capacity to deal with future energy security threats."
                                               10-33

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

10.5.2.4      Military Security Cost Components of Energy Security

   The agencies also attempted to assess the military security benefits components of energy
security in this Draft TAR. The recent literature on the military components of energy security
has included three broad categories of oil related military and national security costs all of which
are hard to quantify and provide estimates of their costs.  These include possible costs of U.S.
military programs to secure oil supplies from unstable regions of the world, the energy security
costs associated with the U.S.  military's reliance on petroleum to fuel its operations and possible
national security costs associated with expanded oil revenues to "rogue states."

   Of these categories listed above, the one that is most clearly connected to petroleum use and
is, in principle, quantifiable is the first, the cost of military programs to secure oil supplies and
stabilize oil supplying regions. There is a developing literature on the measurement of these
components of energy security but methodological and measurement challenges pose significant
challenges to providing a robust estimate of this component of energy security.

   Assessing the military component of the energy security cost has two major challenges:
attribution and incremental analysis. The  attribution  challenge is to determine which military
programs and expenditures can properly be attributed to oil supply protection, rather than  some
other objective.  The incremental  analysis challenge is to estimate how much the petroleum
supply protection costs might vary if U.S.  oil use were to be reduced or eliminated.

   Since "military forces are, to a great extent, multipurpose and fungible" across theaters and
missions (Crane et al. (2009))76, and because the military budget is presented along regional
accounts rather than by mission, the allocation to particular missions is not always clear.
Approaches taken usually either allocate "partial" military costs directly associated with
operations in a particular region, or allocate a share of total military costs (including some that
are indirect in the sense of supporting military activities overall) (Koplow and Martin (1998)).77

   The incremental analysis can estimate how military costs would vary if the oil security
mission is no longer needed, and many studies stop at this point.  It is substantially more difficult
to estimate how military costs would vary if U.S. oil  use or imports are partially reduced.  Partial
reduction of U.S. oil use diminishes the magnitude of the security problem,  but there is
uncertainty that  supply protection forces and their costs could be scaled down in proportion  (e.g.
Crane et al. (2009))78,  and there remains the associated goal of protecting supply and transit for
allies and important trade partners, and other importing countries, if they do not decrease their
petroleum use as well.

   The challenges of attribution and incremental analysis have led some to conclude that the
mission of oil supply protection cannot be clearly separated from others, and the military cost
component of oil security should be taken as near zero (Moore et al. (1997)).79  For example, the
Council on Foreign Relations takes the view that substantial foreign policy missions will remain
over the next 20 years, even without the oil security mission entirely. Stern, on the other hand,
argues that many of the other policy concerns in the Persian Gulf follow from oil, and the
reaction to U.S.  policies taken to protect oil.

   Most commonly, analysts estimate substantial military costs associated with the missions of
oil supply security and associated contingencies, but  avoid estimating specific cost reductions
from partial reductions in oil use. However, some relatively recent studies (Copulos (2003),
                                               10-34

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

Delucchi and Murphy (2008), Crane et al., Stern (2010))80 seek to update, and in some cases
significantly improve the rigor of analysis.

   Delucchi and Murphy sought to deduct from the cost of Persian Gulf military programs, the
costs associated with defending U.S. interests other than the objective of providing more stable
oil supply and price to the U.S. economy. Excluding an estimate of cost for missions unrelated
to oil, and for the protection of oil in the interest of other countries, Delucchi and Murphy
estimated military costs for all U.S. domestic oil interests of between $24 and $74 billion
annually.

   Crane et al. considered force reductions and cost savings that could be achieved if oil security
were no longer a consideration.  After reviewing documents supporting recent defense resource
allocations they concluded that the oil protection mission is prominent: "First, the United States
does include the security of oil supplies and global transit of oil as a prominent element in its
force planning." While they noted that the elimination of this mission of oil supply protection
might not lead to complete reduction of those costs, they concluded there is very likely to be
some cost reduction. Taking two approaches, and guided by post-Cold War force draw downs
and by a top-down look at the current U.S. allocation of defense resources, they concluded that
$75-$91 billion, or 12-15 percent of the current U.S. defense  budget, could be reduced if the oil
protection mission were completely eliminated.

   Stern presents an estimate of military cost for Persian Gulf force projection, addressing the
challenge of cost allocation with  an activity-based cost method. He used information on actual
naval force deployments rather than budgets, focusing on the costs of carrier deployment. As a
result of this different data set and these assumptions regarding allocation, the estimated costs are
much higher, roughly 4 to 10 times, than other recent estimates. For the 1976-2007 time frame,
Stern estimated an average military cost of $212 billion and for 2007, $500 billion.

   A study by the National Research Council (NRC) (2013)81  attempted to estimate the military
costs associated with U.S. imports and consumption of petroleum. The NRC  cites estimates of
the national defense costs of oil dependence from the literature that range from less than $5
billion to $50 billion per year or more. Assuming a range of approximate range of $10 billion to
$50  billion per year, the NRC divided national defense costs by a projected U.S.  consumption
rate  of approximately 6.4 billion barrels per year (EIA, 2012).  This procedure yielded a range of
average national defense cost of $1.50-$8.00 per barrel (rounded to the nearest $0.50), with a
mid-point of $5/barrel (in 2009).  However, as discussed above, it is unclear that incremental
reductions in either U.S. imports, or consumption of domestic petroleum, would  produce
incremental changes to the military expenditures related to the oil protection mission (Crane, et
al.).  The agencies continue to review newer studies and literature to better estimate the military
components of the energy security benefits associated with this Draft TAR, but as of this date,
have not been able to identify a robust methodology that can be used to quantify  the military cost
component of energy security.

10.6  Non-GHG Health and Environmental Impacts

   This section discusses the economic benefits from reductions in health and environmental
impacts resulting from  non-GHG emission reductions (such as criteria and toxic  air pollutants)
that  can be expected to occur as a result of the  light-duty 2022-2025 GHG standards. CCh
emissions are predominantly the byproduct of fossil fuel combustion processes that also produce
                                              10-35

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

criteria and hazardous air pollutant emissions.  The vehicles that are subject to this program are
also significant sources of mobile source air pollution such as direct PM, NOx, VOCs and air
toxics, which are regulated by separate emissions standards programs.  The program will affect
exhaust emissions of these pollutants from vehicles and will also affect emissions from upstream
sources that occur during the refining and distribution of fuel. Changes in ambient
concentrations of ozone,  PM2.5, and air toxics that will result from the program are expected to
affect human health by reducing premature deaths and other serious human health effects, as
well as other important improvements in public health and welfare. Children especially benefit
from reduced exposures to criteria and toxic pollutants, because they tend to be more sensitive to
the effects of these respiratory pollutants. Ozone and particulate matter have been associated
with increased incidence of asthma and other respiratory effects in children, and parti culate
matter has been associated with a decrease in lung maturation.

   It is important to quantify the co-pollutant-related health and environmental impacts
associated with the GHG standards because a failure to adequately consider these ancillary
impacts could lead to an  incorrect assessment of the  standards' costs and benefits. Moreover, the
health and other impacts  of exposure to criteria air pollutants and airborne toxics tend to occur in
the near term, while most effects from reduced climate change are likely to occur only over a
time frame of several decades  or longer.

   For purposes of this Draft TAR, EPA has applied PM-related benefits per-ton values to its
estimated emission reductions as an interim approach to estimating only the PM-related benefits
of the program.82'1" However, there are several health benefit categories that EPA was unable to
quantify due to limitations  associated with using benefits-per-ton estimates, several of which
could be substantial. For example, we have not quantified a number of known or suspected
health benefits linked to reductions in ozone and other criteria pollutants, as well as health
benefits linked to reductions in air toxics. Additionally, we are unable to quantify a number of
known welfare effects, including reduced acid and paniculate deposition damage to cultural
monuments and other materials, and environmental benefits due to reductions  of impacts of
eutrophication in coastal  areas. As a result, the health benefits quantified in this section are
likely underestimates of total benefits.  If necessary, EPA will quantify and monetize the health
and environmental impacts related to both PM and ozone later in the midterm evaluation process,
which would entail photochemical air quality modeling.

10.6.1 Economic Value of Reductions in  Particulate Matter

   As presented in Chapter 12, the standards would reduce emissions of several criteria and toxic
pollutants and their precursors. In this analysis, however, EPA only estimates the economic
value of the human health benefits associated with the resulting reductions in PM2.5 exposure
(related to both directly emitted PM2.5 and secondarily-formed PM2.s).  Due to analytical
limitations with the benefit per-ton method, this analysis does not estimate benefits resulting
p See also: hftii^ywww^^                           The current values available on the webpage have
  been updated since the publication of the Farm et al., 2012 paper. For more information regarding the updated
  values, see: btti)^/www.ei^^
  (accessed June 9, 2016).
                                              10-36

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

from reductions in population exposure to other criteria pollutants such as ozone.Q  Furthermore,
the benefits per-ton method, like all air quality impact analyses, does not monetize all of the
potential health and welfare effects associated with reduced concentrations of PM2.5.

   This analysis uses estimates of the benefits from reducing the incidence of the specific PM2.5-
related health impacts described below. These estimates, which are expressed per ton of PM2.5-
related emissions eliminated by the standards, represent the total monetized value of human
health benefits (including reduction in both premature mortality and premature morbidity) from
reducing each ton of directly emitted PM2.5, or its precursors (SCh and NOx), from a specified
source. Ideally, the human health benefits would be estimated based on changes in ambient
PM2.5 as determined by full-scale air quality modeling.  However, the length of time needed to
prepare the necessary emissions inventories, in addition to the processing time associated with
the modeling itself, has precluded us from performing air quality modeling for the Draft
TAR. If necessary, EPA will conduct this modeling later in the midterm evaluation process.

   The PM-related dollar-per-ton benefit estimates used in this analysis are provided in Table
10.12. As the table indicates, these values differ among directly emitted PM and PM precursors
(SO2 and NOx), and also depend on their original source, because emissions from different
sources can result in different degrees of population  exposure and resulting health impacts. In
the summary of costs and benefits, Chapter 12, EPA presents the monetized value of total PM-
related improvements associated with the standards summed across sources (on-road and
upstream)  sources and across PM-related pollutants (direct PM2.5 and PM precursors SCh and
NOx).
                  Table 10.12 PM-Related Benefits-per-ton Values (thousands, 2012$)a
Year0
On-road Mobile Sources
Direct PM2.5
S02
NOX
Upstream Sourcesd
Direct PM2.5
S02
NOX
Estimated Using a 3 Percent Discount Rateb
2016
2020
2025
2030
$380-$850
$400-$910
$440-$ 1,000
$480-$ 1,100
$20-$45
$22-$49
$24-$55
$27-$61
$7.7-$18
$8.1-$18
$8.8-$20
$9.6-$22
$330-$750
$350-$790
$390-$870
$420-$950
$69-$160
$75-$170
$83-$190
$91-$200
$6.8-$16
$7.4-$17
$8.1-$18
$8.7-$20
Estimated Using a 7 Percent Discount Rateb
2016
2020
2025
2030
$340-$770
$370-$820
$400-$910
$430-$980
$18-$41
$20-$44
$22-$49
$24-$55
$6.9-$16
$7.4-$17
$8.0-$18
$8.6-$20
$290-$670
$320-$720
$350-$790
$380-$850
$63-$140
$67-$150
$75-$170
$81-$180
$6.2-$14
$6.6-$15
$7.3-$17
$7.9-$18
Notes:
a The benefit-per-ton estimates presented in this table are based on a range of premature mortality estimates derived
from the ACS study (Krewski et al., 2009) and the Six-Cities study (Lepeule et al., 2012).
b The benefit-per-ton estimates presented in this table assume either a 3 percent or 7 percent discount rate in the
valuation of premature mortality to account for a twenty-year segmented cessation lag.
Q The air quality modeling that underlies the PM-related benefit per ton values also produced estimates of ozone
   levels attributable to each sector. However, the complex non-linear chemistry governing ozone formation
   prevented EPA from developing a complementary array of ozone benefit per ton values. This limitation
   notwithstanding, we anticipate that the ozone-related benefits associated with reducing emissions of NOx and
   VOC could be substantial.
                                                10-37

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

0 Benefit-per-ton values were estimated for the years 2016, 2020, 2025 and 2030.  We hold values constant for
intervening years (e.g., the 2016 values are assumed to apply to years 2017-2019; 2020 values for years 2021-2024;
2025 values for years 2026-2029; and 2030 values for years 2031 and beyond).
d We assume for the purpose of this analysis that "upstream emissions" are most closely associated with refinery sector
benefit per-ton values. The majority of upstream emission reductions associated with the standards are related to
domestic onsite refinery emissions and domestic crude production. While upstream emissions also include storage
and transport sources, as well as upstream refinery sources, we have chosen to simply apply the refinery values.
   The benefit per-ton technique has been used in previous analyses, including EPA's 2017-2025
Light-Duty Vehicle Greenhouse Gas Rule,83 the Reciprocating Internal Combustion Engine
rules,84'85 and the Residential Wood Heaters NSPS.86  Table 10.13 shows the quantified PM2.5-
related co-benefits captured in those benefit per-ton estimates, as well  as unquantified effects the
benefits per-ton estimates are unable to capture.
                       Table 10.13  Human Health and Welfare Effects of PM2.s
  Pollutant
Quantified and Monetized
in Primary Estimates
Unquantified Effects
Changes in:
  PM2
Adult premature mortality
Acute bronchitis
Hospital admissions: respiratory and
cardiovascular
Emergency room visits for asthma
Nonfatal heart attacks (myocardial infarction)
Lower and upper respiratory illness
Minor restricted-activity days
Work loss days
Asthma exacerbations (asthmatic population)
Infant mortality
Chronic and subchronic bronchitis cases
Strokes and cerebrovascular disease
Low birth weight
Pulmonary function
Chronic respiratory diseases other than chronic
bronchitis
Non-asthma respiratory emergency room visits
Visibility
Household soiling
   Readers interested in reviewing the complete methodology for creating the benefit-per-ton
estimates used in this analysis can consult EPA's "Technical Support Document: Estimating the
Benefit per Ton of Reducing PM2.5 Precursors from 17 Sectors."R  Readers can also refer to Fann
et al. (2012) for a detailed description of the benefit-per-ton methodology. As described in the
documentation, EPA uses a method that is consistent with the cost-benefit analysis that
accompanied the 2012 PM NAAQS revision.  The benefit-per-ton estimates utilize the
concentration-response functions as reported in the epidemiology literature.s>87  To calculate the
total monetized impacts associated with quantified health impacts, EPA applies values derived
from a number of sources. For premature mortality, EPA applies a value of a statistical life
(VSL) derived from the mortality valuation literature. For certain health impacts, such as
R For more information regarding the updated values, see:
   http://www.epa.gov/airquality ^enmap/models/Source_Apportionment_BPT_TSD_l_3 l_13.pdf (accessed
   September 9, 2014).
s Although we summarize the main issues in this chapter, we encourage interested readers to see the benefits chapter
   of the RIA that accompanied the PM NAAQS for a more detailed description of recent changes to the
   quantification and monetization of PM benefits. Note that the cost-benefit analysis was prepared solely for
   purposes of fulfilling analysis requirements under Executive Order 12866 and was not considered, or otherwise
   played any part, in the decision to revise the PM NAAQS.
                                                  10-38

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

respiratory-related ailments, EPA applies willingness-to-pay estimates derived from the
valuation literature. For the remaining health impacts, EPA applies values derived from current
cost-of-illness and/or wage estimates.

   The documentation cited above also describes that national per-ton estimates were developed
for selected PM-related pollutant/source category combinations.  The per-ton values calculated
therefore apply only to tons reduced from those specific PM-related pollutant/source
combinations (e.g., NCh emitted from on-road mobile sources; direct PM emitted from electricity
generating units). EPA's estimate of PM2.5 benefits is therefore based on the total direct PIVb.s
and PM-related precursor emissions controlled by sector and multiplied by each per-ton value.

   As Table 10.12 indicates, EPA projects that the per-ton values for reducing emissions of non-
GHG pollutants from both vehicle use and upstream sources such as fuel refineries will increase
over time.1 These projected increases reflect rising income levels, which increase affected
individuals' willingness to pay for reduced exposure to health threats from air pollution.11  They
also reflect future population growth and increased life expectancy, which expands the size of
the population exposed to air pollution in both urban and rural areas, especially among older age
groups with the highest mortality risk.v
   The benefit-per-ton estimates are subject to a number of assumptions and uncertainties:

        •   The benefit-per-ton estimates used in this analysis reflect specific geographic patterns
           of emissions reductions and specific air quality and benefits modeling assumptions
           associated with the derivation of those estimates (see the TSD describing the
           calculation of the national benefit-per-ton  estimates). 88'w Consequently, these
           estimates may not reflect local variability in population density, meteorology,
           exposure, baseline health incidence  rates, or other local factors associated with the
           current analysis. Therefore, use of these benefit-per-ton values to estimate non-GHG
           benefits may lead to higher or lower benefit estimates than if these benefits were
           calculated based on direct air quality modeling. EPA plans to conduct full-scale air
           quality modeling later in the midterm evaluation process in an effort to capture this
           variability.
        •   This analysis assumes that all fine particles, regardless of their chemical composition,
           are equally potent in causing premature mortality. This is an important  assumption,
           because PM2.5 produced via transported precursors emitted from stationary sources
           may differ significantly from direct  PM2.5 released from diesel engines and other
T As we present in Chapter 12, the standards would yield emission reductions from upstream refining and fuel
   distribution due to decreased petroleum consumption.
u The issue is discussed in more detail in the 2012 PM NAAQS RIA, Section 5.6.8.  See U.S. Environmental
   Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the National Ambient Air
   Quality Standards for Particulate Matter, Health and Environmental Impacts Division, Office of Air Quality
   Planning and Standards, EPA-452-R-12-005, December 2012. Available on the internet:
   http://www.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf.
v For more information about EPA's population projections, please refer to the following:
   http://www.epa.gov/air/benmap/models/BenMAPManualAppendicesAugust2010.pdf (See Appendix K)
w See also: httBI//wsQ«g3a^^                          The current values available on the webpage have
   been updated since the publication of the Farm et al., 2012 paper. For more information regarding the updated
   values, see: hfiBI/Aw^y^
   (accessed September 9, 2014).
                                                10-39

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

 industrial sources.  The PM ISA, which was twice reviewed by SAB-CASAC,
 concluded that "many constituents of PM2.5 can be linked with multiple health effects,
 and the evidence is not yet sufficient to allow differentiation of those constituents or
 sources that are more closely related to specific outcomes."89 PM composition and
 the size distribution of those particles vary within and between areas due to source
 characteristics.  Any specific location could have higher or lower contributions of
 certain PM species and  other pollutants than the national average, meaning potential
 regional differences in health impact of given control strategies.  Depending on the
 toxicity of each PM species reduced by the proposed standards, assuming equal
 toxicity could over or underestimate benefits.
 When estimating the benefit-per-ton values, EPA assumes that the underlying health
 impact functions for fine particles are linear within the range of ambient
 concentrations under consideration. Thus, the estimates include health benefits from
 reducing fine particles in areas with varied concentrations of PM2.5, including regions
 that are in attainment with the fine particle standard.  The direction of bias that
 assuming a linear-no threshold model (or an alternative model) introduces depends
 upon the "true" functional from  of the relationship and the specific assumptions and
 data in a particular analysis.  For example, if the true function identifies a threshold
 below which health effects do not occur, benefits may be overestimated if a
 substantial portion of those benefits were estimated to occur below that threshold.
 Alternately, if a substantial portion of the benefits occurred above that threshold, the
 benefits may be underestimated  because an assumed linear no-threshold function may
 not reflect the steeper slope above that threshold to account for all health effects
 occurring above that threshold.
 There are several health benefit  categories that EPA was unable to quantify due to
 limitations associated with using benefits-per-ton estimates, several of which could be
 substantial. Because the NOx and VOC emission reductions associated with the
 standards are also precursors to  ozone, reductions in NOx and VOC would also
 reduce ozone formation and the  health  effects associated with ozone exposure.
 Unfortunately, ozone-related benefits-per-ton estimates do not exist due to issues
 associated with the complexity of the atmospheric air chemistry and nonlinearities
 associated with ozone formation. The PM-related benefits-per-ton estimates also do
 not include any human welfare or ecological benefits.
 There are many uncertainties associated with the health impact functions that underlie
 the benefits-per-ton estimates. These include: within-study variability (the precision
 with which a given study  estimates the relationship between air quality changes and
 health effects); across-study variation (different published studies of the same
 pollutant/health effect relationship typically do not report identical findings and in
 some instances the differences are substantial); the application of concentration-
 response functions nationwide (does not account for any relationship between region
 and health effect, to the extent that such a relationship exists); extrapolation of impact
 functions across population (we  assumed that certain health impact functions applied
 to age ranges broader than that considered in the original epidemiological study); and
 various uncertainties in the concentration-response function, including causality and
 thresholds.  These uncertainties  may under- or over-estimate benefits.
                                     10-40

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

       •  EPA has investigated methods to characterize uncertainty in the relationship between
          PM2.5 exposure and premature mortality. EPA's final PIVh.sNAAQS analysis
          provides a more complete picture about the overall uncertainty in PM2.5 benefits
          estimates. For more information, please consult the PIVh.sNAAQS RIA.90
       •  The benefit-per-ton unit values used in this analysis incorporate projections of key
          variables, including atmospheric conditions, source level emissions, population,
          health baselines, incomes, and technology. These projections introduce some
          uncertainties to the benefit per ton estimates.


   As mentioned above, emissions changes and benefits-per-ton estimates alone are not a good
indication of local or regional air quality and health impacts, as there may  be localized impacts
associated with the standards. Additionally, the atmospheric chemistry related to ambient
concentrations of PIVfo.s, ozone and air toxics is very complex. Full-scale photochemical
modeling is therefore necessary to provide the needed spatial and temporal detail to more
completely and accurately estimate the changes in ambient levels of these  pollutants and  their
associated health and welfare impacts.  As discussed above, timing constraints precluded EPA
from conducting a full-scale photochemical air quality modeling analysis in time for the Draft
TAR. Later in the midterm evaluation process, EPA plans to quantify and monetize the health
and environmental impacts related to both PM and ozone, which entails photochemical air
quality modeling.

10.7  Greenhouse Gas Emission Impacts

   We estimate the global social benefits of CCh emission reductions expected from the 2022-
2025 final standards using the SC-CCh estimates presented in the Technical Support Document:
Technical Update of the Social Cost of Carbon for Regulatory Impact Analysis Under Executive
Order 12866 (May 2013, Revised July 2015) ("current TSD").91  We refer to these estimates,
which were developed by the U.S. government, as "SC-CCh estimates." The SC-CCh is a metric
that estimates the monetary value of impacts associated with marginal changes in CCh  emissions
in a given year. It includes a wide range of anticipated climate impacts, such as net changes in
agricultural productivity and human health, property damage from increased flood risk, and
changes in energy system costs, such as reduced costs for heating and increased costs for air
conditioning. It is typically used to assess the avoided damages as a result of regulatory actions
(i.e., benefits of rulemakings that lead to an incremental reduction in cumulative global CCh
emissions).

   The SC-CCh estimates used in the final 2017-2025 RIA and in this analysis were developed
over many years, using the best science available,  and with input from the public.  Specifically,
an interagency working group (IWG) that included the EPA and other executive branch agencies
and offices used three integrated assessment models (lAMs) to develop the SC-CCh estimates
and recommended four global values for use in regulatory analyses. The SC-CCh estimates were
first released in February 2010 and were used to estimate the value of CCh benefits in the final
2017-2025 rulemaking.

   These SC-CCh estimates were developed using an ensemble of the three most widely cited
integrated assessment models in the economics literature with the ability to estimate the SC-CCh.
A key objective of the IWG was to draw from the insights of the three models while respecting
                                             10-41

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

the different approaches to linking GHG emissions and monetized damages taken by modelers in
the published literature.  After conducting an extensive literature review, the interagency group
selected three sets of input parameters (climate sensitivity, socioeconomic and emissions
trajectories, and discount rates) to use consistently in each model. All other model features were
left unchanged, relying on the model developers' best estimates and judgments, as informed by
the literature. Specifically, a common probability distribution for the equilibrium climate
sensitivity parameter, which informs the strength of climate's response to atmospheric GHG
concentrations, was used across all three models. In addition,  a common range of scenarios for
the socioeconomic parameters and emissions forecasts were used in all three models.  Finally,
the marginal damage estimates from the three models were estimated using a consistent range of
discount rates, 2.5, 3.0, and 5.0 percent. See Technical Support Document: Technical Update of
the Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866
(February 2010) ("2010 TSD") for a complete discussion of the methods used to develop the
estimates and the key uncertainties, and the current TSD for the latest estimates.92

   In 2013, and after the final LD 2017-2025 rulemaking, the IWG updated the SC-CCh
estimates using new versions of each LAM.  The 2013 update did not revisit the 2010 modeling
decisions with regards to the discount rate, reference case socioeconomic and emission scenarios,
and equilibrium climate sensitivity distribution. Rather, improvements in the way damages are
modeled are confined to those that have been incorporated into the latest versions of the models
by the developers themselves and published in the peer-reviewed literature. The model updates
that are relevant to the SCC estimates include: an explicit representation of sea level rise
damages in the Dynamic Integrated Climate and Economy (DICE) and Policy Analysis of the
Greenhouse Effect (PAGE) models; updated adaptation assumptions, revisions to ensure
damages are constrained by GDP, updated regional scaling of damages, and a revised treatment
of potentially abrupt shifts in climate damages in the PAGE model; an updated carbon cycle in
the DICE model; and updated damage functions for sea level rise impacts, the agricultural sector,
and reduced space heating requirements, as well as changes to the transient response of
temperature to the buildup of GHG concentrations and the inclusion of indirect effects of
methane emissions in the Climate Framework for Uncertainty, Negotiation, and Distribution
(FUND) model. The current TSD presents and discusses the 2013 update (including recent
minor technical corrections to the estimates).x

   The updated estimates continue to represent global measures because of the distinctive nature
of the climate change, which is highly unusual in at least three respects. First, emissions of most
GHGs contribute to damages around the world independent of the country in which they are
emitted. The SC-CCh must therefore incorporate the full (global) damages caused by GHG
emissions to address the global nature of the problem.  Second, the U.S. operates in a global and
highly interconnected economy, such that impacts on the other side of the world can affect our
economy.  This means that the true costs of climate change to the U.S. are larger than the direct
impacts that simply occur within the U.S.  Third, climate change represents a classic public
goods problem because each country's reductions benefit everyone else and no country can be
excluded from enjoying the benefits  of other countries' reductions, even if it provides no
reductions itself. In this situation, the only way to achieve an economically efficient level of
x Both the 2010 TSD and the current TSD are available at: https://www.whitehouse.gov/omb/oira/social-cost-of-
  carbon.
                                              10-42

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

emissions reductions is for countries to cooperate in providing mutually beneficial reductions
beyond the level that would be justified only by their own domestic benefits.  In reference to the
public good nature of mitigation and its role in foreign relations, thirteen prominent academics
noted that these "are compelling reasons to focus on a global SCC" in a recent article on the SCC
(Pizer et al., 2014).  In addition, as noted in OMB's Response to Comments on the SC-CCh, a
document discussed further below, there is no bright line between domestic and global damages.
Adverse impacts on other countries can have  spillover effects on the United States, particularly
in the areas of national  security, international trade, public health and humanitarian concerns.93

   The 2010 TSD noted a number of limitations to the SC-CCh analysis, including the
incomplete way in which the integrated assessment models capture catastrophic and non-
catastrophic impacts, their incomplete treatment of adaptation and technological change,
uncertainty in the extrapolation of damages to high temperatures, and assumptions regarding risk
aversion. Currently integrated assessment models do not assign value to all of the important
physical, ecological, and economic impacts of climate change recognized in the climate change
literature due to a lack of precise information on the nature of damages and because the science
incorporated into these models understandably lags behind the most recent research/ The
limited amount of research linking climate impacts to economic damages makes the modeling
exercise even more difficult. These individual limitations do not all work in the same direction
in terms of their influence on the SC-CCh estimates, though taken together they suggest that the
SC-CCh estimates are likely conservative. In particular, the IPCC Fourth Assessment Report
(2007), which was the most current IPCC assessment available at the time of the IWG's 2009-
2010 review, concluded that "It is very likely that [SC-CCh estimates] underestimate the damage
costs because they cannot include many non-quantifiable impacts." Since then, the peer-
reviewed literature has  continued to support this conclusion.  For example, the IPCC Fifth
Assessment report observed that SC-CCh estimates continue to omit various impacts that would
likely increase damages.

   The EPA and other agencies have continued to consider feedback on the SC-CCh estimates
from stakeholders through a range of channels, most recently including public comments on the
Clean Power Plan rulemaking94 and others that use the SC-CCh in supporting analyses and
through regular interactions with stakeholders and research analysts implementing the SC-CCh
methodology used by the interagency working group.  Commenters have provided constructive
recommendations for potential opportunities to improve the SC-CCh estimates in future updates.
In addition, OMB sought public comment on the approach used to develop the SC-CCh estimates
through a separate comment period and published a response to those comments in 2015.z

   After careful evaluation of the full range of comments submitted to OMB, the IWG continues
to recommend the use of the SC-CO2 estimates in regulatory impact analysis. With the release of
the response to comments, the IWG announced plans in July 2015 to obtain expert independent
Y Climate change impacts and SCC modeling is an area of active research. For example, see: (1) Howard, Peter,
  "Omitted Damages: What's Missing from the Social Cost of Carbon." March 13, 2014,
  http://costofcarbon.org/files/Omitted Damages Whats Missing From the Social  Cost of Carbon.pdf: and (2)
  Electric Power Research Institute, "Understanding the Social Cost of carbon: A Technical Assessment," October
  2014, www.epri.com.
z See https://www.whitehouse.gov/sites/default/files/omb/inforeg/scc-response-to-comments-final-iuly-2015.pdf.
                                              10-43

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

advice from the National Academies of Sciences, Engineering and Medicine to ensure that the
SC-CCh estimates continue to reflect the best available scientific and economic information on
climate change.AA The Academies then convened a committee, "Assessing Approaches to
Updating the Social Cost of Carbon," (Committee) that is reviewing the state of the science on
estimating the SC-CCh, and will provide expert, independent advice on the merits of different
technical approaches for modeling and highlight research priorities going forward. EPA will
evaluate its approach based upon any feedback received from the Academies' panel.

   To date, the Committee has released an interim report, which recommended against doing a
near term update of the SC-CCh estimates.  For future revisions, the Committee recommended
the IWG move efforts  towards a broader update of the climate system module consistent with the
most recent, best available science, and also offered recommendations for how to enhance the
discussion and presentation of uncertainty in the SC-CCh estimates.  Specifically, the Committee
recommended that "the IWG provide guidance in their technical support documents  about how
[SC-CCh] uncertainty should be represented and discussed in individual regulatory impact
analyses that use the [SC-CCh]" and that the technical support document for each update of the
estimates present a section discussing the uncertainty in the overall approach, in the models used,
and uncertainty that may not be included in the estimates.BB At the time of this writing, the IWG
is reviewing the interim report and considering the recommendations. EPA looks forward to
working with the IWG to respond to the recommendations and will continue to follow IWG
guidance on SC-CCh.
   The current SC-CCh estimates are as follows: $14,  $47, $70, and $140 per ton of CCh
emissions in the year 2022 (2013$).cc The first three values are based on the average  SC-CCh
from the three lAMs, at discount rates of 5, 3, and 2.5 percent, respectively. SC-CCh estimates
for several discount rates are included because the literature shows that the SC-CCh is  quite
sensitive to assumptions about the discount rate, and because no consensus exists on the
appropriate rate to use in an intergenerational context (where costs and benefits are incurred by
different generations).  The fourth value is the 95th percentile of the SC-CCh from all three
models at a 3 percent discount rate.  It is included to represent lower probability but higher -
impact outcomes from climate change, which are captured further out in the tail of the SC- CCh
distribution, and while less likely than those reflected by the average SC- CCh estimates, would
be much more harmful to society and therefore, are relevant to policy makers.
AA The Academies' review will be informed by public comments and focus on the technical merits and challenges of
  potential approaches to improving the SC-CCh estimates in future updates. See
  h!ttiLL//wjft5;Lld^
BB National Academies of Sciences, Engineering, and Medicine. (2016). Assessment of Approaches to Updating the
  Social Cost of Carbon: Phase 1 Report on a Near-Term Update. Committee on Assessing Approaches to
  Updating the Social Cost of Carbon, Board on Environmental Change and Society. Washington, DC: The
  National Academies Press, doi: 10.17226/21898. See Executive Summary, page 1, for quoted text.
cc The current version of the TSD is available at: https://www.whitehouse.gov/sites/default/files/omb/inforeg/scc-
  tsd-fmal-july-2015.pdf.  The 2010 and 2013 TSDs present SC-CO2 in 2007$ per metric ton. The unrounded
  estimates from the current TSD were adjusted to 2013$ using GDP Implicit Price Deflator (1.097),
  http://www.bea.gov/iTable/index_nipa. The estimates presented in this document were rounded to two significant
  digits.
                                               10-44

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

   The current estimates are higher than those used to analyze the CCh impacts in the final LD
2017-2025 rulemaking, which preceded the 2013 SC-CCh update and were published in the 2010
SC-CCh TSD. By way of comparison, the four SC-CCh estimates used to analyze the CCh
impacts for the final LD 2017-2015 rulemaking were $8.1, $30, $48, and $93 per metric ton in
2022 (2013$).DD  As previously noted, the IWG updated these estimates in 2013 using new
versions of each integrated assessment model but did not Table 10.14 presents the current global
SC-CCh estimates for select years between 2022 and 2050. In order to calculate the dollar value
for emission reductions, the  SC-CCh estimate for each emissions year would be applied to
changes in CCh emissions for that year, and then discounted back to the analysis year using the
same discount rate used to estimate the SC-CCh.  The SC-CCh increases over time because future
emissions are expected to produce larger incremental damages as physical and economic systems
become more stressed in response to greater climate change.  Note that the interagency group
estimated the growth rate of the SC-CCh directly using the three integrated assessment models
rather than assuming a constant annual growth rate.  This helps to ensure that the estimates are
internally consistent with other modeling assumptions. Chapter 12 reports the updated GHG
benefits in  select model years and calendar years.
                 Table 10.14 Social Cost of CCh, 2015-2050 (in 2013$ per metric ton)*
           Discount Rate and Statistic
 Year      5% Average         3% Average         2.5% Average          3% (95th percentile)
2022
2023
2024
2025
2030
2040
$14
$14
$14
$15
$18
$23
$47
$48
$49
$50
$55
$66
$70
$71
$72
$75
$80
$92
$140
$140
$150
$150
$170
$200
 2050      $29	$76	$100	$230	
Note:
* These SC-CCh values are stated in $/metric ton and rounded to two significant figures. The estimates vary
depending on the year of COa emissions and are defined in real terms, i.e., adjusted for inflation using the GDP
implicit price deflator.


   One limitation of the primary benefits analysis in the 2017-2025 final rulemaking is that it did
not include the valuation of non-CCh GHG impacts (CH4, N2O, HFC-134a).  Specifically, the
IWG did not estimate the social costs of non-CCh GHG emissions using an approach analogous
to the one used to estimate the SC-CCh.  While there were other estimates of the social cost of
non- CCh GHGs in the peer review literature, the methodologies underlying those estimates were
inconsistent with the methodology the IWG used to estimate the SC-CCh.  As discussed in the
DD The 2010 and 2013 TSDs present SC-CO2in $2007; see fatlesi/Awwj^
   carbon for both TSDs. The estimates used in the final 2017-2025 rulemaking were adjusted to $2010 using GDP
   Implicit Price Deflator. The estimates have been adjusted to 2013$ here for consistency with the Draft TAR. See
   National Income and Product Accounts Tables, Table 1.1.9 at Ililil^MvB^^                       for
   GDP Implicit Price Deflators.
                                              10-45

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

2017-2025 final rulemaking, there is considerable variation among these published estimates in
the models and input assumptions they employ.EE These studies differ in the emission
perturbation year, employ a wide range of constant and variable discount rate specifications, and
consider a range of baseline socioeconomic and emissions scenarios that have been developed
over the last 20 years. EPA also determined that the estimates in the literature were most likely
underestimates due to changes in the underlying science since their publication.FF

   However, EPA recognized that non-CCh GHG impacts associated with these standards (e.g.,
net reductions in CH4, N2O, and HFC-134a) would provide benefits to society.  To understand
the potential implication of omitting these benefits, EPA conducted sensitivity analysis using an
approximation approach based on global warming potential (GWP) gas comparison metrics that
has been used in previous rulemakings. The EPA also sought public comments on the valuation
of non-CCh GHG impacts in the proposed LD 2017-2025 rulemaking and other previous
rulemakings (e.g., U.S.  EPA 2012).95 In general, the commenters strongly  encouraged the EPA
to incorporate the monetized value of non-CCh GHG impacts into the benefit cost analysis,
however they noted the challenges associated with the GWP-approach,  as discussed further
below, and encouraged the use of directly-modeled estimates of the SC-CH4to overcome those
challenges.

   Subsequent to the 2017-2025 final rule, a paper by Marten et al. (2014) provided the first set
of published SC-CH4 and SC-N2O estimates that are consistent with the modeling assumptions
underlying the SC-CCh.96  Specifically, the estimation approach of Marten  et al. used the same
set of three lAMs, five socioeconomic and emissions scenarios, equilibrium climate sensitivity
distribution, three constant discount rates, and aggregation approach used by the IWG to develop
the SC-CCh estimates. The aggregation method involved distilling the 45 distributions of the
SC-CH4 and of the SC-N2O produced for each emissions year into four estimates: the mean
across all models  and scenarios using a 2.5 percent, 3 percent, and 5 percent discount rate, and
the 95th percentile of the pooled estimates from all models and scenarios using a 3 percent
discount rate.  Marten et al. also used the same rationale as the IWG to develop global estimates
of the SC-CH4 and SC-N2O, given that methane and N2O are global pollutants.

   The atmospheric lifetime and radiative efficacy of methane and N2O used by Marten et al. is
based on the estimates reported by the IPCC in their Fourth Assessment Report (AR4, 2007),
including an adjustment in the radiative efficacy of methane to account for its role as a precursor
for tropospheric ozone and stratospheric water.  These values represent the same ones used by
the IPCC in AR4 for calculating GWPs. At the time Marten et al. developed their estimates of
the SC-CH4, AR4 was the latest assessment report by the IPCC. The IPCC updates  GWP
estimates with each new assessment, and in the most recent assessment, AR5, the latest estimate
of the methane GWP ranged from 28-36, compared to a GWP of 25 in AR4. The updated values
reflect a number of changes: changes in the lifetime and radiative  efficiency estimates for CCh,
changes in the lifetime estimate for methane, and changes in the correction factor applied to
EE The researchers cited in the 2017-2015 RIA include: Fankhauser (1994); Kandlikar (1995); Hammitt et al. (1996);
  Tol et al. (2003); Tol (2004); and Hope and Newbeny (2006).
FF See the 2017-2025 RIA, page 7-7, for complete discussion. Literature included studies primarily from the mid-
  1990s through early 2000s. !lUlK//!LWs3j3^
                                             10-46

-------
 	Economic and Other Key Inputs Used in the Agencies' Analyses

 methane's GWP to reflect the effect of methane emissions on other climatically important
 substances such as tropospheric ozone and stratospheric water vapor. In addition, the range
 presented in the latest IPCC report reflects different choices regarding whether to account for
 how biogenic and fossil methane have different carbon cycle effects, and for whether to account
 for climate feedbacks on the carbon cycle for both methane and CCh (rather than just for CCh as
 was done in AR4).97'GG

    Marten et al. (2014) discuss these estimates, (SC-CH4 and SC-N2O estimates presented below
 in Table  10.15), and compare them with other recent estimates  in the literature. The authors
 noted that a  direct comparison of their estimates with all of the  other published estimates is
 difficult, given the differences in the models and socioeconomic and emissions scenarios, but
 results from three relatively recent studies offer a better basis for comparison (see Hope (2006),
 Marten and Newbold (2012), Waldhoff et al. (2014)). Marten et al. found that, in general, the
 SC-CH4 estimates from their 2014 paper are higher than previous estimates and the SC-N2O
 estimates from their 2014 paper fall within the range from Waldhoff et al.  The higher SC-CH4
 estimates are partially driven by the higher effective radiative forcing due to the inclusion of
 indirect effects from methane emissions in their modeling. Marten et al., similar to other recent
 studies, also find that their directly modeled SC-CH4 and SC-N2O estimates are higher than the
 GWP-weighted estimates. More detailed discussion of the SC-CH4 and SC-N2O estimation
 methodology,  results and a comparison to other published estimates can be found in Marten et al.

    The resulting SC-CH4 and SC-N2O estimates are presented in Table 10.15. The tables do not
 include HFC-134a because EPA is unaware of analogous estimates.
        Table 10.15 Social Cost of CH4 and Social Cost of N20,2015-2050 (in 2013$ per metric ton)

Year
2022
2023
2024
2025
2030
2040
2050
Social Cost of CH4
5% (Avg)
$640
$660
$690
$710
$830
$1,100
$1,400
3% (Avg)
$1,400
$1,400
$1,500
$1,500
$1,800
$2,200
$2,700
2.5% (Avg)
$1,800
$1,900
$1,900
$2,000
$2,200
$2,900
$3,400
3% (95th
percentile)
$3,700
$3,800
$3,900
$4,100
$4,600
$6,000
$7,300
Social Cost of N2O
5%
(Avg)
$5,500
$5,700
$5,900
$6,000
$6,900
$9,200
$12,000
3% (Avg)
$17,000
$18,000
$18,000
$19,000
$21,000
$25,000
$30,000
2.5% (Avg)
$25,000
$25,000
$26,000
$26,000
$30,000
$35,000
$41,000
3% (95th
percentile)
$45,000
$46,000
$47,000
$48,000
$54,000
$66,000
$79,000
Note:
* These SC-CH4 and SC-N2O values are stated in $/metric ton and rounded to two significant figures.  The estimates
 vary depending on the year of emissions and are defined in real terms, i.e., adjusted for inflation using the GDP
 implicit price deflator. In addition, the estimates in this table have been adjusted to reflect the minor technical
 corrections to the SC-CO2 estimates described above. See Corrigendum to Marten et al. (2014) for more details
  } Note that the Draft TAR uses 100-year GWP values for CC>2 equivalency calculations that are consistent with the
   GHG emissions inventories and the IPCC Fourth Assessment Report (AR4), i.e., 25 for methane. The IPCC
   reported the same 100-year GWP for N2O (298) in AR4 and AR5.
                                               10-47

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

   Today's publication updates the analysis of non-CCh GHG benefits presented in the 2017-
2025 final rule by using Marten et al. (2014) estimates of SC-CH4 sand SC-N2O.  In particular,
the application of directly modeled estimates from Marten et al. (2014) to benefit-cost analysis of
a regulatory action is analogous to the use of the SC-CCh estimates. Specifically, the SC-CH4
and SC-N2O estimates in Table 10.15 are used to monetize the benefits of reductions in methane
and N2O emissions, respectively, expected as a result of the 2022-2025 standards.  Forecast
changes in methane (or N2O) emissions in a given year, expected as a result of the standards, are
multiplied by the SC-CFLt (or SC-ISbO) estimate for that year.  To obtain a present value
estimate, the monetized stream of future non-CCh GHG benefits are discounted back to the
analysis year using the same discount rate used to estimate the social cost of the non-CCh GHG
emission changes. In addition, the limitations for the SC-CCh estimates discussed above likewise
apply to the SC-CH4 and SC-ISbO estimates, given the consistency in the methodology.

   The EPA recently conducted a peer review of the application of the Marten et al. (2014) non-
CO2 social cost estimates in regulatory analysis and received responses that supported this
application.  Three reviewers considered seven charge questions that covered issues such as the
EPA's interpretation of the Marten et al. estimates, the consistency of the estimates with the SC-
CO2 estimates, the EPA's characterization of the limits of the GWP-approach to value non-CCh
GHG impacts, and the appropriateness of using the Marten et al. estimates in regulatory impact
analyses.  The reviewers agreed with the EPA's interpretation of Marten et al.'s estimates;
generally found the estimates to be consistent with the SC-CCh estimates; and concurred with the
limitations of the GWP approach, finding directly modeled estimates to be more appropriate.
While outside of the scope of the review, the reviewers briefly considered the limitations in the
SC-CCh methodology (e.g., those discussed earlier in this section) and noted that because the
SC-CCh and SC-CH4 (SC-N2O) methodologies are similar, the limitations also apply to the
resulting SC-CH4 (SC-ISbO) estimates.  Two of the reviewers concluded that use in RIAs of the
SC-CH4 (SC-N2O) estimates developed by Marten et al. and published in the peer-reviewed
literature is appropriate, provided that the agency discusses the limitations, similar to the
discussion provided for SC-CCh and other economic analyses. All three reviewers encouraged
continued improvements in the SC-CCh estimates and suggested that as those improvements are
realized they should also be reflected in the SC-CH4 (SC-N2O) estimates, with one reviewer
suggesting the SC-CH4 (SC-N2O) estimates should lag this process. The EPA supports
continued improvement in the SC-CCh estimates developed by the U.S. government and agrees
that improvements in the SC-CCh estimates should also be reflected in the SC-CH4 (SC-N2O)
estimates. The fact that the reviewers agree that the SC-CH4 (SC-N2O) estimates are generally
consistent with the SC-CCh estimates that are recommended by OMB's guidance on valuing CCh
emissions reductions, leads the EPA to conclude that use of the SC-CH4 (SC-N2O) estimates is
an analytical improvement over excluding methane emissions from the monetized portion of the
benefit cost analysis.

   In light of the favorable peer review and past comments urging the EPA to value non-CCh
GHG impacts in its rulemakings, the agency has used the Marten et al. (2014)  SC-CH4 and SC-
N2O estimates to value methane and N2O impacts, respectively, expected from the 2022-2025
standards.

   The summary of GHG (CCh, methane, N2O) benefits are presented for select model years and
calendar years is in Chapter 12.
                                             10-48

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

   EPA is unaware of estimates of the social cost of HFC-134a that are analogous to the SC-
CCh, SC-CH4, and SC-N2O estimates discussed above.  In the 2017-2025 final rulemaking, EPA
used the GWP for HFC-134a to convert the emissions of this gas to CCh equivalents, which were
then valued using the SC-CCh estimates.  These estimates were presented in a sensitivity analysis
due to the limitations associated with using the GWP  approach to value changes in non-CCh
GHG emissions.

   The GWP measures the cumulative radiative forcing from a perturbation of a non-CCh GHG
relative to a perturbation of CCh over a fixed time horizon, often 100 years.  The GWP mainly
reflects differences in the radiative efficiency of gases and differences in their atmospheric
lifetimes.  While the GWP is a simple, transparent, and well-established metric for assessing the
relative impacts of non-CCh emissions compared to CCh on a purely physical basis, there are
several well-documented limitations in using it to value non-CCh GHG benefits, as discussed in
the 2010 SC-CCh TSD and previous rulemakings.98 In particular, several recent studies found
that  GWP-weighted benefit estimates for methane are likely to be lower than the estimates
derived using directly modeled social cost estimates for these gases. Gas comparison metrics,
such as the GWP, are designed to measure the impact of non-CCh GHG emissions relative to
CCh at a specific point along the pathway from emissions to monetized  damages (depicted in
Figure 10.4), and this point may differ across measures.
Emissions
h.
w
Atmospheric
Concentration
^
w
Radiative
Forcing
^
w
Climate
Impacts
h.
f
Environmental
and Socio-
Economic
Impacts
h.
r
Monetized
Damages
       Figure 10.4 Path from GHG Emissions to Monetized Damages (Source: Marten et al., 2014)
   The GWP is not ideally suited for use in benefit-cost analyses to approximate the social cost
of non-CCh GHGs because it ignores important nonlinear relationships beyond radiative forcing
in the chain between emissions and damages. These can become relevant because gases have
different lifetimes and the SC-CCh takes into account the fact that marginal damages from an
increase in temperature are a function of existing temperature levels.  Another limitation of gas
comparison metrics for this purpose is that some environmental and socioeconomic impacts are
not linked to all of the gases under consideration, or radiative forcing for that matter, and will
therefore be incorrectly allocated.  For example, the economic impacts associated with increased
agricultural productivity due to higher atmospheric CCh concentrations included in the SC-CCh
would be incorrectly allocated to methane emissions with the GWP-based valuation approach.

   Also of concern is the fact that the assumptions made in estimating the GWP are not
consistent with the assumptions underlying SC-CCh estimates in general, and the SC-CCh
estimates developed by the IWG more specifically. For example, the 100-year time horizon
usually used in estimating the GWP is less than the approximately 300-year horizon the IWG
used in developing the SC-CCh estimates.  The GWP approach also treats all impacts within the
time horizon equally, independent of the time at which they occur. This is inconsistent with the
role of discounting in economic analysis, which accounts for a basic preference for earlier over
later gains in utility and expectations regarding future levels of economic growth.
                                             10-49

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

   The changes in HFC-134a emissions occur through model year 2021, at which point use of
HFC-134a in new vehicles is prohibited under the Significant New Alternatives Policy (SNAP).
As discussed in Chapter 5.2.9.2, EPA expects that HFC-134a will be entirely replaced by
refrigerants with lower GWPs by model year 2021. In other words, there will be no further
reductions in HFC-134a emissions after model year 2021. Given that this midterm review
considers years after 2021, there are no changes in impacts to report for HFC-134a.  See Chapter
5.2.9.2 for complete discussion, including EPA's assessment about the transition to use of low-
GWP alternative refrigerants.

10.8  Benefits from Reduced Refueling Time

   The total time spent pumping and paying for fuel, and driving to and from fueling stations,
represents an economic cost to drivers and other vehicle occupants.  Increased driving range
provides a benefit to individuals arising from the value of the time saved when refueling events
are eliminated.  As described in this section, the EPA calculates this benefit by applying DOT-
recommended values of travel time savings to estimates of how much time is saved.

   The increases in fuel economy resulting from the standards are expected to lead to some
increase in vehicle driving range. The extent of this increase depends on manufacturers'
decisions to apply reduced fuel consumption requirements towards increasing range, rather than
reducing tank size while maintaining range.  For the 2012 FRM, EPA conducted a regression
analysis to identify the relationship between fuel economy and fuel tank size for different vehicle
classes based on historical data. Trends in fuel tank size for a number of redesigned vehicles
were also investigated. Based on these analyses, fuel economy improvements were assumed to
be entirely realized as improvements in driving range, due to insufficient evidence to indicate
that fuel tank size is reduced as vehicle fuel economy is improved. For this Draft TAR analysis,
EPA is again using the FRM assumption that fuel tank sizes remain constant; however, we will
continue to monitor trends in fuel tank designs and vehicle range.

   No direct estimates of the value of extended vehicle range or reduced fuel tank size are
readily available.  Instead, this analysis calculates the reduction in the annual amount of time a
driver would spend filling its fuel tank; this reduced time could result either from fewer refueling
events, if new fuel tanks stay the same size, or  from less time spent filling the tank during each
refueling  stop, if new fuel tanks are made proportionately smaller.  As discussed in Section 10.4,
the average number of miles each type of vehicle is driven annually would likely increase under
the regulation, as drivers respond to lower fuel expenditures (the "rebound effect").  The
estimates of refueling time in effect allow for this increase in vehicle use. However, the estimate
of the rebound effect does not account for any reduction in net operating costs from lower
refueling time.  Because the rebound effect should measure the change in VMT with respect to
the net change in overall operating costs, refueling time costs would ideally factor into this
calculation.  The effect of this omission is expected to be minor because refueling time savings
are generally small relative to the value of reduced fuel expenditures.

   The savings in refueling time are calculated as the total amount of time the driver of a typical
vehicle would save each year as a consequence of pumping less fuel into the vehicle's tank.  The
calculation also includes a fixed time per refill  event of 3.5 minutes which would not occur as
frequently due to the fewer number of refills.
                                             10-50

-------
                          Economic and Other Key Inputs Used in the Agencies' Analyses
   The calculation uses the reduced number of gallons consumed by truck type and divides that
value by the tank volume and refill amount to get the number of refills, then multiplies that by
the time per refill to determine the number of hours saved in a given year.  The calculation then
applies DOT-recommended values of travel time savings to convert the resulting time savings to
their economic value.  The input metrics used in the EPA analysis are included in Table 10.16.
                Table 10.16 Metrics Used in Calculating the Value of Refueling Time
Metric
Average tank refill percentage
Average tank volume
Fuel dispense rate
Fixed time per refill
Wage rate for the value of refill time
Number of people in vehicle
Wage growth rate, 2014 baseyear
Value
65%
15 gallons
lOgal/min
3.5 minutes
$25.00
1.2
1.1%
   The equation used by EPA to calculate refueling benefits is shown below.
                                              Gal per refill
  Refueling Benefit =
                         Gal per refill
x
  \Fuel dispense rate
              \  / *
+ time per refill x  —
              /  \hr
                                                                                  labor
            Table 10.17 Metrics Used in Calculating the Value of Refueling Time by NHTSA
Metric
Average tank refill percentage
Average tank volume
Fuel dispense rate
Fixed time per refill
Wage rate for the value of refill time
Number of people in vehicle
Wage growth rate, 2014 base year
Value
65%
15 gallons
lOgal/min
3.5 minutes
$18.07/$18.37
1.2
1.1%
   The economic value of refueling time savings was calculated by applying DOT-recommended
valuations for travel time savings to estimates of how much time is saved.HH The value of travel
time depends on average hourly valuations of personal and business time, which are functions of
annual household income and total hourly compensation costs to employers. The nationwide
median annual household income, $51,939 in 2013, is divided by 2,080 hours to yield an income
of $25.00 per hour.  The total hourly compensation cost to employers, inclusive of benefits, in
2013$ is  $24.40.n Table 10.18 demonstrates the agency's approach to estimating the value of
travel time ($/hour) for both urban and rural (intercity) driving.  This approach relies on the use
of DOT-recommended weights that assign a lesser valuation to  personal travel time than to
HHhttps://www.transportation.gov/administrations/office-policy/2015-value-travel-time-guidance.
11 Ibid.
                                             10-51

-------
                           Economic and Other Key Inputs Used in the Agencies' Analyses
business travel time, as well as weights that adjust for the distribution between personal and
business travel.

     Table 10.18 Estimating the Value of Travel Time for Urban and Rural (Intercity) Travel ($/hour)
Urban Travel

Wage Rate ($/hour)
DOT- Recommended Value of Travel Time Savings, as% of Wage Rate
Hourly Valuation (=Wage Rate * DOT-Recommended Value)
% of Total Urban Travel
Hourly Valuation (Adjusted for % of Total Urban Travel)
Rural (Intercity) Travel

Wage Rate ($/hour)
DOT- Recommended Value of Travel Time Savings, as% of Wage Rate
Hourly Valuation (=Wage Rate * DOT-Recommended Value)
% of Total Rural Travel
Hourly Valuation (Adjusted for % of Total Rural Travel)

Personal Travel
$25.00
50%
$12.50
95.4%
$11.93

Personal Travel
$25.00
70%
$17.50
78.6%
$13.76

Business Travel
$24.40
100%
$24.40
4.6%
$1.12

Business Travel
$24.40
100%
$24.40
21.4%
$5.22

Total
-
-
-
100%
$13.05

Total



100%
$18.98
   The estimates of the hourly value of urban and rural travel time ($13.05 and $18.98,
respectively) shown in Table 10.18 must be adjusted to account for the nationwide ratio of urban
to rural driving. By applying this adjustment (as shown in Table 10.19), an overall estimate of
the hourly value of travel time - independent of urban or rural status - may be produced. Note
that the calculations above assume only one adult occupant per vehicle. To fully estimate  the
average value of vehicle travel time, the agency must account for the presence of additional adult
passengers during refueling trips. NHTSA applies such an adjustment as shown in Table 10.19;
this adjustment is performed separately for passenger cars and for light trucks, yielding
occupancy-adjusted valuations of vehicle travel time during refueling trips for each fleet. Note
that children (persons under age 16) are excluded from average vehicle occupancy counts, as it is
assumed that the opportunity cost of children's time is zero.
           Table 10.19 Estimating the Value of Travel Time for Light-Duty Vehicles ($/hour)

Urban Travel
Rural Travel
Total


Average Vehicle Occupancy During
Refueling Trips (persons)
Weighted Value of Travel Time
($/hour)
Occupancy-Adjusted Value of Vehicle
Travel Time During Refueling Trips
($/hour)
Unweighted Value of
Travel Time ($/hour)
$13.05
$18.98
-

Passenger Cars
1.21
$14.93
$18.07
Weight (% of Total
Miles Driven)
68.2%
31.8%
100.0%

2b3 Light Trucks
1.23
$14.93
$18.37
Weighted Value of
Travel Time ($/hour)
$8.90
$6.03
$14.93


                                              10-52

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

 10.9  Benefits and Costs from Additional Driving

 10.9.1 Travel Benefit

   The increase in travel associated with the rebound effect produces additional benefits to
 vehicle drivers, which reflect the value of the added (or more desirable) social and economic
 opportunities that become accessible with additional travel. The analysis estimates the economic
 benefits from increased rebound-effect driving as the sum of fuel expenditures incurred plus the
 vehicle owner/operator surplus from the  additional accessibility it provides. As evidenced by the
 fact that vehicles make more frequent or longer trips when the cost of driving declines, the
 benefits from this added travel exceed added expenditures for the fuel consumed. Note that the
 amount by which the benefits from this increased driving exceed its increased fuel costs
 measures the net benefits from the additional travel, usually referred to as increased consumer
 surplus or, in this case, increased driver surplus. The equation for the calculation of the total
 travel benefit is shown below:
                                / $ \       /i\            [/ $
     Travel Benefit = (VMTrehound) I — 1    + ^-J (VMTrehound) I —
                                \   'policy                 IA
                                                              -
                                                         i .mileI          \mile   ,.
                                    policy                  \_^   ' reference   \    ' policy


   The agencies' analysis estimates the economic value of the increased owner/operator surplus
provided by added driving using the conventional approximation, which is one half of the
product of the decline in vehicle operating costs per vehicle-mile and the resulting increase in the
annual number of miles driven. Because it depends on the extent of improvement in fuel
economy, the value of benefits from increased vehicle use changes by model year and varies
among alternative standards. Under even those alternatives that would impose the highest
standards, however, the magnitude of the surplus from additional vehicle use represents a small
fraction of this benefit.
10.9.2 Costs Associated with Crashes, Congestion and Noise

   In contrast to the benefits of additional driving are the costs associated with that driving. If net
operating costs of the vehicle decline, then we expect a positive rebound  effect. Increased
vehicle use associated with a positive rebound effect also  contributes to increased traffic
congestion, motor vehicle crashes, and highway noise. Depending on how the additional travel
is distributed throughout the day and on where it takes place, additional vehicle use can
contribute to traffic congestion and delays by increasing traffic volumes on facilities that are
already heavily traveled during peak periods. These added delays impose higher costs on drivers
and other vehicle occupants in the form of increased travel time and operating expenses.
Because drivers do not take these added costs into account in deciding when and where to travel,
they must be accounted for separately as a cost of the added driving associated with the rebound
effect.

   EPA and NHTSA rely on estimates of congestion, crash, and noise costs caused light-duty
vehicles developed by the Federal Highway Administration to estimate the increased external
costs caused by added driving due to the rebound effect. The FHWA estimates are intended to
measure the increases in costs from added congestion, property damages  and injuries in traffic
crashes, and noise levels caused by various classes  of vehicles that are borne by persons other
than their drivers (or "marginal" external costs).  EPA and NHTSA employed estimates from this
                                              10-53

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

source previously in the analysis accompanying the light-duty 2012-2016 vehicle rulemaking.
The agencies continue to find them appropriate for this analysis after reviewing the procedures
used by FHWA to develop them and considering other available estimates of these values.

   FHWA's congestion cost estimates focus on freeways because non-freeway effects are less
serious due to lower traffic volumes and opportunities to re-route around the congestion.  The
agencies, however, applied the congestion cost to the overall VMT increase, though the fraction
of VMT on each road type used in MOVES range from X to Y  percent of the vehicle miles on
freeways for light-duty vehicles. The results of this analysis potentially overestimate the
congestions costs associated with increased vehicle use, and thus lead to a conservative estimate
of net benefits.

   The agencies are using FHWA's "Middle"  estimates for marginal congestion, crash, and noise
costs caused by increased travel from vehicles. This approach is consistent with the
methodology used in both LD and FID GHG rules.  These costs are multiplied by the annual
increases in vehicle miles travelled from the rebound effect to yield the estimated increases in
congestion, crash, and noise externality costs during each future year. The values used are shown
in Table 10.20.
  Table 10.20 Metrics Used to Calculate the Costs Associated with Congestion, Crashes and Noise Linked to
                                   Rebound Miles Traveled
Metric
Congestion
Crashes
Noise
Value
$0.0583 per mile
$0.0252 per mile
$0.0008 per mile
 10.10 Discounting Future Benefits and Costs

   The benefits and costs are analyzed using 3 percent and 7 percent discount rates, consistent
 with current OMB guidance.99'" These rates are intended to represent consumers' preference for
 current over future consumption (3 percent), and the real rate of return on private investment (7
 percent) which indicates the opportunity cost of capital. However, neither of these rates
 necessarily represents the discount rate that individual decision-makers use, nor do they reflect
 the rates in OMB Circular A-94 Appendix C, which are revised annually.100  The 2015 Appendix
 lists real (i.e., inflation-adjusted) discount rates between 0.3 percent (for a 3-year period) and  1.5
 percent (for a 30-year time horizon). All costs and benefits are discounted to 2015 except for
 those considered in payback analyses where costs and benefits are discounted to the first year of
 a vehicle's life.
11 Discounting involving the Social Cost of Carbon (SC-CCh) values uses several discount rates because the
  literature shows that the SC-CCh is quite sensitive to assumptions about the discount rate, and because no
  consensus exists on the appropriate rate to use in an intergenerational context (where costs and benefits are
  incurred by different generations). Refer to Section 10.7 for more information.
                                               10-54

-------
                          Economic and Other Key Inputs Used in the Agencies' Analyses
10.11 Additional Costs of Vehicle Ownership
10.11.1
Maintenance & Repair Costs
   We continue to believe that the maintenance estimates used in the FRM are still reasonable
and have therefore used them again in this analysis. We distinguish maintenance from repair
costs as follows: maintenance costs are those costs that are required to keep a vehicle properly
maintained and, as such, are usually recommended by auto makers to be conducted on a regular,
periodic schedule. Examples of maintenance costs are oil and air filter changes, tire
replacements, etc. Repair costs are those costs that are unexpected and, as such, occur randomly
and uniquely for every driver, if at all. Examples of repair costs would be parts replacement
following a crash or a mechanical failure, etc.

   In Chapter 3.6 of the final joint TSD supporting the 2012 FRM, the agencies presented a
lengthy  discussion of maintenance costs and the impacts projected as part of that rule.101 Table
10.21 shows the results of that analysis, the maintenance impacts used in the 2012 FRM and
again in this analysis, although the costs here have been updated to 2013$. Note that the
technologies shown in Table 10.21 are those for which we believe that maintenance costs would
change; it is clearly not a complete list of technologies expected to meet the MY2025 standards.
                     Table 10.21  Maintenance Event Costs & Intervals (2013$)
New Technology
Low rolling resistance tires level 1
Low rolling resistance tires level 2
Diesel fuel filter replacement
EVoil change
EV air filter replacement
EV engine coolant replacement
EV spark plug replacement
EV/PHEV battery coolant replacement
EV/PHEV battery health check
Reference
Technology
Standard tires
Standard tires
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Gasoline vehicle
Cost per Maintenance Event
$6.71
$51.55
$51.93
-$40.78
-$30.16
-$62.21
-$87.52
$123.37
$40.78
Maintenance Interval
(miles)
40,000
40,000
20,000
7,500
30,000
100,000
105,000
150,000
15,000
   Note that many of the maintenance event costs for EVs are negative. The negative values
represent savings since EVs do not incur these costs while their gasoline counterparts do. Note
also that the 2010 FRM is expected to result in widespread use of low rolling resistance tires
level 1 (LRRT1) on the order of 85 percent penetration. Therefore, as 2012 FRM results in
increasing use of low rolling resistance tire level 2 (LRRT2), there is a corresponding decrease in
the use of LRRT1.  As such, as LRRT2 maintenance costs increase with increasing market
penetration, LRRT1 maintenance costs decrease. The technology penetrations of these
technologies are those  shown in Section 12.2. The resultant maintenance costs are as shown in
Section 12.4.
10.11.2
Sales Taxes
   When consumers consider their total cost of ownership of a vehicle, or its potential payback,
they may consider the sales taxes they have to pay at the time of purchasing the vehicle. As
these costs are transfer payments, they are not included in the societal costs of the program, but
                                             10-55

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

they are included as one of the increased costs to the consumer for these standards when we
calculate costs that the consumer pays out for vehicle ownership as part of our payback analysis.
In the 2012 FRM, the agencies took the most recent auto sales taxes by state and weighted them
by population by state to determine a national weighted-average sales tax of 5.46 percent.KK  The
agencies sought to weight sales taxes by new vehicle sales by state; however, such data were,
and continue to be, unavailable.  It is recognized that for this purpose, new vehicle sales by state
is a superior weighting mechanism to Census population; in an effort to approximate new vehicle
sales by state, during the 2012 FRM, a study of the change in new vehicle registrations (using
R.L. Polk data) by state across recent years was conducted, resulting in a corresponding set of
weights. Use of the weights derived from the study of vehicle registration data resulted in a
national weighted-average sales tax rate almost identical to that resulting from the use of Census
population estimates as weights, just slightly above 5.5 percent. The agencies opted to utilize
Census population rather than the registration-based proxy of new vehicle sales as the basis for
computing this weighted average, as the end results were negligibly different and the analytical
approach involving new vehicle registrations had not been as thoroughly reviewed. We have
used the same value in this Draft TAR as was used in the 2012 FRM.

10.11.3       Insurance Costs

   The agencies considered the standards' impact to consumers' auto insurance expenses over
vehicle lifetimes.  More expensive vehicles will require more expensive collision and
comprehensive (e.g., theft) car insurance.  The scope of this analysis is to estimate the increased
cost to the consumer for these standards, not the increase in societal costs due to collision and
property damage.  The increase in insurance costs was estimated from the average value of
collision plus comprehensive insurance as a proportion of average new vehicle price.  Collision
plus comprehensive insurance represent the portion of insurance costs that depend on vehicle
value. In the 2012 FRM, a study by Quality Planning provided the average value of collision
plus comprehensive insurance for new vehicles, in 2010$, as $521 ($396 of which was collision
and $125 of which was comprehensive).102 The average consumer expenditure for a new
passenger car in 2011, according to the Bureau of Economic Analysis was $24,572 and the
average price of a new light truck was $31,721 in $2010.103 Using sales volumes from the
Bureau, we determined an average passenger car and an average light truck price was  $27,953 in
$2010 dollars.104

   Dividing the  cost to insure a new vehicle by the average price of a new vehicle gives the
proportion of comprehensive plus collision insurance as 1.86 percent of the price of a vehicle.
As vehicles' values decline with vehicle age, comprehensive and collision insurance premiums
likewise decline. Data on the change in insurance  premiums as a function of vehicle age are
scarce; however, the agencies utilized data from the aforementioned Quality Planning study that
KK See htlfi^/BQBviactOTwa^                           (last accessed April 5, 2012). Note that county,
  city, and other municipality-specific taxes were excluded from the weighted averages, as the variation in locality
  taxes within states, lack of accessible documentation of locality rates, and lack of availability of weights to apply
  to locality taxes complicate the ability to reliably analyze the subject at this level of detail. Localities with
  relatively high automobile sales taxes may have relatively fewer auto dealerships, as consumers would endeavor
  to purchase vehicles in areas with lower locality taxes, therefore reducing the impact of the exclusion of
  municipality-specific taxes from this analysis.
                                               10-56

-------
	Economic and Other Key Inputs Used in the Agencies' Analyses

cite the cost to insure the average vehicle on the road today (average age 10.8 years) to enable a
linear interpolation of the change in insurance premiums during the first 11 years of a typical
vehicle's life.LL To illustrate, as a percentage of the base vehicle price of $27,953, the cost of
collision and comprehensive insurance in each of the first five years of a vehicle's life is 1.86
percent, 1.82 percent, 1.75 percent,  1.64 percent, and 1.50 percent, respectively, or 8.57 percent
in aggregate.  The agencies additionally utilized data from the same Quality Planning study that
cite average insurance costs for vehicles greater than 10 years of age (for which the agencies
estimated age to be 18, as this is the age at which half of vehicles in service at age 10 remain in
service) to extrapolate insurance costs to age 18. Discounting is applied to future insurance
payments in the model's calculations, and all calculations are adjusted by projected vehicle
survival rates.

   The agencies considered whether to estimate incremental comprehensive and collision
insurance premiums only to year 18. As vehicles age, it becomes increasingly impractical to
purchase these forms of insurance, and the Quality Planning study indicates that many owners
drop these forms of insurance much earlier - in some cases upon repayment of the initial auto
loan.  The agencies nevertheless use the 30-year lifetime of the vehicle because  we use survival-
weighted values, which take into account the probability that some vehicles are  no longer
incurring costs because they no longer exist. This approach may tend to overstate insurance
costs, because many owners are not paying insurance collision/comprehensive premiums even on
vehicles that continue to  exist.  Therefore, the insurance premiums were age-adjusted to year 30
using the assumption that by end-of-life, no vehicle would remain on comprehensive or collision
insurance. This approach provides the agencies with our estimates of the impact of insurance
costs on vehicle owners based on the expected increase in MSRP resulting from the standards.

   As discussed earlier, the scope of this analysis is to estimate the increased cost to the
consumer in the context of our payback analysis, not the increase in societal costs or benefits.
LL Insurance data did not differentiate between passenger cars and light trucks. Therefore, a 30-year lifetime was
  assumed in this analysis. Due to several factors, among them discounting, decreased vehicle value with age, and
  limited vehicle survival in later years of vehicles' lifetimes, this assumption is of minimal impact on the results.
                                               10-57

-------
  	Economic and Other Key Inputs Used in the Agencies' Analyses

   References
1 Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975 Through 2015,"
Table 10.1, page 128, epa.gov/otaq/fetrends.htm, December 2015.
2 Ibid.
3 How Do Motorists' Own Fuel Economy Estimates Compare with Official Government Ratings? A Statistical
Analysis," David Greene et al, University of Tennessee, October 1, 2015.
4 US EPA. MOVES Motor Vehicle Emissions Simulator. hUjK^yiQvwljgg^^
Version 2014a. November, 2015.
5 US EPA. Population and Activity of On-road Vehicles in MOVES2014. EPA-420-R-16-003a. March 2016.
6 Winebrake, J.J., Green, E.H., Comer, B., Corbett, J.J., Froman, S., 2012. "Estimating the direct rebound effect for
on-road freight transportation," Energy Policy 48, 252-259.
7 Greene, D.L., Kahn, J.R., Gibson, R.C., 1999, "Fuel economy rebound effect for U.S. household vehicles," The
Energy Journal, 20.
8 Greening, L.A., Greene, D.L., Difiglio, C., 2000, "Energy efficiency and consumption — the rebound effect — a
survey," Energy Policy, 28, 389-401.
9 Sorrell,  S. and Dimitropoulos, J., "UKERC Review of Evidence of the Rebound Effect, Technical Report 2:
Econometric Studies," Sussex Energy Group, Working Paper, 2007.
10 Pickrell, D. and Schimek, P., 1999. "Growth in Motor Vehicle Ownership and Use: Evidence from the
Nationwide Personal Transportation Survey," Journal of Transportation and Statistics, vol. 2, no. 1, pp. 1-17. [EPA-
HQ-OAR-2010-0799-00027].
11 Puller,  S. and Greening, L., 1999. "Household Adjustment to Gasoline Price Change: An Analysis Using Nine
Years of U.S. Survey Data," Energy Economics 21(l):37-52. [EPA-HQ-OAR-2010-0799-0754].
12 Pickrell, D. and Schimek, P., 1999. "Growth in Motor Vehicle Ownership and Use: Evidence from the
Nationwide Personal Transportation Survey," Journal of Transportation and Statistics, vol. 2, no. 1, pp. 1-17.  [EPA-
HQ-OAR-2010-0799-0027] .
13 Bento,  A.,  Goulder, L., Jacobsen, M. and Haefen, R., 2009, "Distributional and Efficiency Impacts of Increased
U.S. Gasoline Taxes, American Economic Review, 99:3 667-699.
14 Wadud, Z., Graham, D.J., Noland, R.B., 2009. Modelling fuel demand for different socio-economic groups.
Applied Energy 86, 2740-2749.
15 West, R. and Pickrell, D., "Factors  Affecting Vehicle Use in Multiple-Vehicle Households," 2009, NHTS
Workshop, June, 2011.
16 Small, K. and Van Dender, K., 2007a. "Fuel Efficiency and Motor Vehicle Travel: The Declining Rebound
Effect." The Energy Journal, vol. 28,  no. 1, pp. 25-51.  [EPA-HQ-OAR-2010-0799-0755].
17 Small, K. and Van Dender, K., 2007b. "Long Run Trends in Transport Demand, Fuel Price Elasticities and
Implications  of the Oil Outlook for Transport Policy,"  OECD/ITF Joint Transport Research Centre Discussion
Papers 2007/16, OECD, International Transport Forum. [EPA-HQ-OAR-2010-0799-0756].
18 Hymel, K.  M., Small, K. A., and Van Dender, K., "Induced demand and rebound effects in road transport,"
Transportation Research Part B: Methodological, Volume 44, Issue 10, December 2010, Pages 1220-1241, ISSN
0191-2615, DOI: 10.1016/j.trb.2010.02.007. [EPA-HQ-OAR-2010-2010-0799-0758].
19 Greene, David, 2012. "Rebound 2007: Analysis of U.S. light-duty vehicle travel statistics," Energy Policy, vol.
41, pp. 14-28. [EPA-HQ-OAR-2010-0799-0759].
20Dargay, J.M.,Gately,  D., 1997. "The demand for transportation fuels: imperfect price-reversibility?"
Transportation Research Part B 31(1), [EPA-HQ-OAR-2010-0799-076], Sentenac-Chemin, E., 2012, "Is the price
effect on fuel consumption symmetric: Some evidence from an empirical study," Energy Policy, Volume 41, pp. 59-
65 [EPA-HQ-OAR-2010-0799-0762].
21 Gately, D., 1993. "The Imperfect Price-Reversibility of World Oil Demand," The Energy Journal, International
Association for Energy Economics, vol.  14(4), pp. 163-182. [EPA-HQ-OAR-2010-0799-0763].
22 Su, Q., 2012, A quantile regression analysis of the rebound effect: Evidence from the 2009 National Household
Transportation Survey in the United States, Energy Policy 45, pp. 368-377.
23 Linn, J., 2013, "The Rebound Effect of Passenger Vehicles," RFF Discussion Paper, No. 13-19 [EPA-HQ-OAR-
2010-0799-0761].
24 Lui, Y.,Tremblay, J. and Cirillo, C., 2014. "An integrated model for discrete and continuous decisions with
application to vehicle ownership, type and usage," Transportation Research Part A, pp. 315-328.
                                                    10-58

-------
                              Economic and Other Key Inputs Used in the Agencies' Analyses
25 Frondel, M., and Vance, C., 2013. Re-Identifying the Rebound: What about Asymmetry? Energy Journal 34
(4):43-54.
26 Gillingham, K., 2014, Identifying the Elasticity of Driving: Evidence from a Gasoline Price Shock. Regional
Science & Urban Economics 47 (4): 13-24.
27 Hymel, K. M. and Small, K. A., 2015, "The rebound effect for automobile travel: Asymmetric response to price
changes and novel features of the 2000s," Energy Economics, 49 (2015) 93-103 [EPA-HQ-OAR-2015-0827-0034].
28 West, J., Hoekstra, M., Meer, J., Puller, S., 2015, "Vehicle Miles (Not) Traveled: Why Fuel Economy
Requirements Don't Increase Household Driving" National Bureau of Economic Research (NBER), NBER Working
Paper Series, Working Paper 21194, http://www.nber.org/papers/w21194.
29 De Borger, B. Mulalic, I., and Rouwendal, J.,2016, Measuring the rebound effect with micro data: A first
difference approach, Journal of Environmental Economics and Management, 79, 1-17.
30 Gillingham, K., Rapson, D., Wagner, G., 2016, "The Rebound Effect and Energy Efficiency Policy," Review of
Environmental Economics and Policy, 10 (1), pp. 68 - 88.
31 Barla, P., Lamonde, B., Miranda-Moreno, L. and Boucher, N., 2009. Traveled Distance, Stock and Fuel
Efficiency of Private Vehicles in Canada: Price Elasticities and Rebound Effect. Transportation 36 (4):389-402.
32 EIA Annual Energy Review, various editions. For data 2012-2014, and projected data: EIA Annual Energy
Outlook (AEO) 2015 (Reference Case). See Table 11, file "aeotab_ll.xls"  [EPA-HQ-OAR-2014-0827-0619].
33 EIA Annual Energy  Outlook 2015, Table 11, aeotab_ll.xlsx [EPA-HQ-OAR-2014-0827-0619].
34 See EIA Annual Energy Review, various editions. For data 2012-2014, and projected data: EIA Annual Energy
Outlook (AEO) 2015 (Reference Case). See Table 11, file "aeotab_ll.xls"  [EPA-HQ-OAR-2014-0827-0619].
35 EIA, Annual Energy Outlook 2015, International petroleum and other liquids supply, disposition, and prices.
Table 21, yearbyyear.xlsx.
36 EIA Annual Energy Outlook 2015, Table 21, yearbyyear.xlsx.
37 Based on data from the CIA, combining various recent years, Mjj2S^MTja¥.ciajjOT^

38IEA 2011 'TEA Response System for Oil Supply Emergencies." [EPA-HQ-OAR-2014-0827-0573].
39 We looked at changes in U.S. crude oil imports and net petroleum products in the AEO 2015 Reference Case in
comparison the Low (i.e., Economic Growth) Demand Case to undertake this analysis. See the spreadsheet "Impact
of Fuel Demand on Imports AEO2015.xlsx." We also considered a paper entitled "Effect of a U.S. Demand
Reduction on Imports and Domestic Supply Levels" by Paul Leiby, 4/16/2013. This paper suggests that "Given a
particular reduction in oil demand stemming from a policy or significant technology change, the fraction of oil use
savings that shows up as reduced U.S. imports, rather than reduced U.S. supply, is actually quite close to 90 percent,
and probably close to 95 percent".  [EPA-HQ-OAR-2014-0827-0572]  [EPA-HQ-OAR-2014-0827-0574].
40 Leiby, Paul N., Donald W. Jones, T. Randall Curlee, and Russell Lee, 1997, Oil Imports: An Assessment of
Benefits and Costs, ORNL-6851, Oak Ridge National Laboratory, [EPA-HQ-OAR-2014-0827-0594].
41 Leiby, P., Factors Influencing Estimate of Energy Security Premium for the GHG Program, Oak Ridge National
Laboratory. [EPA-HQ-OAR-2014-0827-0595].
42 Brown, S.and Huntington., H., 2013, Assessing the U.S. Oil Security Premium, Energy Economics, vol. 38, pp.
118 -  127. [EPA-HQ-OAR-2014-0827-0571].
43 Brown, S. and Huntington, H., 2010, Reassessing the Oil Security Premium. RFF Discussion Paper Series, (RFF
DP 10-05). doi: RFF DP 10-05  [EPA-HQ-OAR-2014-0827-0602].
44 Greene, D., 2010, Measuring energy security: Can the United States achieve oil independence? Energy Policy,
38(4), 1614-1621. EPA-HQ-OAR-2014-0827-0576].
45 Toman, M., 1993, The economics of energy security: theory, evidence and policy, Chapter 25, Handbook of
Natural Resource and Energy Economics, Volume 3, pp. 1167-1218.
46 Ledyard, J. O. "Market Failure." The New Palgrave Dictionary of Economics. Second Edition. Eds. Steven N.
Durlauf and Lawrence E. Blume. Palgrave Macmillan, 2008. [EPA-HQ-OAR-2014-0827-0596].
47 Sivaram, Varan; Levi, Michael A., 2015, "Automobile Fuel Standards in Lower-Oil-Price World", Council on
Foreign Relations.
48 National Academy of Sciences,  2015, "Cost, Effectiveness and Deployment of Fuel Economy Technologies for
Light-Duty Vehicles," Committee on the Assessment of Technologies for Improving Fuel, Economy of Light-Duty
Vehicles, Phase 2; Board on Energy and Environmental Systems; Division on Engineering and Physical Sciences;
National Research Council.
49 Gately, D., 2004, "OPEC's Incentives for Faster Output Growth," The Energy Journal, 25  (2):75-96; Gately, D.,
2007. "What Oil Export Levels Should We Expect From OPEC?" The Energy Journal, 28(2):151-173.  [EPA-HQ-
OAR-2014-0827-0599].

                                                    10-59

-------
                              Economic and Other Key Inputs Used in the Agencies' Analyses
50 Drabik, D. and de Gorter, H., 2011, "Biofuel Policies and Carbon Leakage", AgBioForum, 14(3): 104-110
51 Rajagopal, D., Hochman, G. and Zilberman, D., 2011, "Indirect fuel use policies (IFUC) and the lifecycle
environmental impact of biofuel policies", Energy Policy, 39: 228-233.
52 Thompson, W., Whistance, J. and Meyer, S., 2011, "Effects of US biofuel policies on US and world petroleum
product markets and consequences for greenhouse gas emissions", Energy Policy, 39: 5509-5518.
53 Erickson, P. and Lazarus, M, 2014, "Impact of Keystone Pipeline on global oil markets and greenhouse gas
emissions", Stockholm Environmental Institute.
54 Karplus, V., Kishimoto, P. and Paltsev, S., 2015, "The Global Energy, CO2 Emissions and Economic Impacts of
Vehicle Fuel Economy Standards", Journal of Transport Economics and Policy, 49(4): 517-538.
55 Historical data are from EIA Annual Energy Review, various editions. For data since 2012 and projected data:
source is EIA Annual Energy Outlook (AEO) 2015 (Reference Case). See Table 11, file "aeotab_ll.xlsx" and Table
20 (Macroeconomic Indicators," (file "aeotab_20.xlsx"). [EPA-HQ-OAR-2014-0827-0619].
56 Bohi, D. R. and Montgomery, D., 1982. Social Cost of Imported and U.S. Import Policy, Annual Review of
Energy, 7:37-60. Energy Modeling Forum, 1981. World Oil, EMF Report 6 (Stanford University Press: Stanford 39
CA. https//emf.stanford.edu/publications/emf-6-world-oil.
57 Plummer, J. (Ed.), 1982. Energy Vulnerability, "Basic Concepts, Assumptions and Numerical Results," pp. 13 -
36, (Cambridge MA: Ballinger Publishing Co.).  [EPA-HQ-OAR-2014-0827-0600].
58 Bohi, D., and Montgomery, D., 1982, Social Cost of Imported and U.S. Import Policy, Annual Review of Energy,
7:37-60.
59 Hogan, W., 1981, "Import Management and Oil Emergencies," Chapter 9 in Deese, 5 David and Joseph Nye, eds.
Energy and Security. Cambridge, MA: Ballinger Publishing Co. [EPA-HQ-OAR-2014-0827-0578].
60 Broadman, H. G. 1986, "The Social Cost of Imported Oil," Energy Policy 14(3):242-252. BroadmanH. and W.
Hogan, 1988. "Is an Oil Import Tariff Justified? An American Debate: The Numbers Say 'Yes'." The Energy
Journal 9: 7-29.  [EPA-HQ-OAR-2014-0827-0569]  [EPA-HQ-OAR-2014-0827-0570].
61 Leiby, P., Jones, D., Curlee, R. and Lee, R., Oil Imports: An Assessment of Benefits and Costs, ORNL-6851, Oak
Ridge National Laboratory, 1997.  [EPA-HQ-OAR-2014-0827-0594].
62 Parry, I. and Darmstadter J., 2004, "The Costs of U.S. Oil Dependency," Resources for the Future, November 17,
2004 (also published as NCEP Technical Appendix Chapter 1: Enhancing Oil Security, the National Commission on
Energy Policy 2004 Ending the Energy Stalemate - A Bipartisan Strategy to Meet America's Energy Challenges).
[EPA-HQ-OAR-2014-0827-0603].
63 National Research Council, 2009, Hidden Costs of Energy: Unpriced Consequences of Energy Production and
Use. National Academy of Science, Washington, DC.  [EPA-HQ-OAR-2014-0827-0597].
64 See, William Nordhaus, "Who's Afraid of a Big Bad Oil Shock?"and Blanchard, O. and Gali, J., "The
macroeconomic Effects of Oil price Shocks: Why are the 2000s so different from the 1970s?," pp.  373-421,  in The
International Dimensions of Monetary Policy. Gali, J., and Gertler,  M., editors, University of Chicago Press,
February 2010, available at http://www.nber.org/chapters/c0517.pdf [EPA-HQ-OAR-2014-0827-0567].
65 Blanchard, O.  and Gali, J., PP. 414. [EPA-HQ-OAR-2014-0827-0568].
66 See, Oil Price Drops on Oversupply, http://www.oil-price.net/en/articles/oil-price-drops-on-oversupplv.php.
10/6/2014.  [EPA-HQ-OAR-2014-0827-0566].
67 Hamilton, J. D., 2012, Oil Prices, Exhaustible Resources, and Economic Growth. In Handbook of Energy and
Climate Change. Retrieved from http://econweb.ucsd.edu/~ihamilto/handbook climate.pdf  [EPA-HQ-OAR-2014-
0827-0577].
68 Ramey, V.and Vine, D., 2010, "Oil, Automobiles, and the U.S. Economy: How Much have Things Really
Changed?" National Bureau of Economic Research Working Papers, WP 16067. Retrieved from
http://www.nber.org/papers/wl6067.pdf [EPA-HQ-OAR-2014-0827-0601].
69 Baumeister, C.; Peersman, G.; VanRobays, I., 2010, "The Economic Consequences of Oil Shocks: Differences
across Countries and Time", Workshop and Conference on Inflation Challenges in the Era of Relative Price Shocks.
70Kilian, L.; Vigfusson, R., 2014, "The Role of Oil Price Shocks in Causing U.S. Recessions", Board of Governors
of the Federal Reserve System.  International Finance Discussion Papers.
71 Cashin, P., Mohaddes, K., Raissi, Maziar, and Raissi, M.,2014, "The differential effects of oil demand and supply
shocks on the global economy". Energy Economics.
72 Batovic, A., 2015, Low oil prices fuel political and economic instability. Global Risk Insights, 18-19. Retrieved
from http://globalriskinsights.com/2015/09/low-oil-prices-fuel-political-and-economic-instability/.

73 Monaldi, F., 2015, The Impact of the Decline in Oil Prices on the Economics, Politics and Oil Industry of
Venezuela. Columbia Center on Global Energy Policy Discussion Papers, (September). Retrieved from

                                                    10-60

-------
                               Economic and Other Key Inputs Used in the Agencies' Analyses
http://energypolicy.columbia.edu/sites/default/files/energy/Impact of the Decline in Oil Prices on Venezuela,
September2015.pdf.
74 Even, S., & Guzansky, Y., 2015, Falling oil prices and Saudi stability - Opinion. Jerusalem Post, (September 30).
Retrieved from http://www.jpost.com/Opinion/Falling-oil-prices-and-Saudi-stability-419534.
75 International Monetary Fund (IMF), 2015, IMFRegional Economic Outlook- Middle East and Central Asia.
Regional Economic Outlook (Vol. 33). Tomkiw, L., 2015, Oil Rich Saudi Arabia Running Out Of Assets? IMF
Report Says It's Possible In Next 5 Years. International Business Times, October 21, 19-22. Retrieved from
http://www.ibtimes.com/oil-rich-saudi-arabia-running-out-assets-imf-report-says-its-possible-next-5-years-215017.
76 Crane, K., Goldthau, A., Toman, M., Light, T., Johnson, S., Nader, A., Rabasa, A. and Dogo, H., Imported oil and
U.S. national security. RAND Corporation, 2009, http://www.stormingmedia.us/62/6279/A627994.pdf.
77 Koplow, D., Martin A., 1998, Fueling Global Warming: Federal Subsidies to Oil in the United States.
Greenpeace, Washington, DC.
78 Crane et al.,  2009, "Imported Oil and U.S. National Security", RAND Corporation.
79 Moore, J., Behrens, C., and Blodgett, I, "Oil Imports: An Overview and Update of Economic and Security
Effects." CRS Environment and Natural Resources Policy Division report 98, no. 1 (1997): 1-14.
80 Copulos, M R. "America's Achilles Heel: The Hidden Costs of Imported Oil." Alexandria VA: The National
Defense Council Foundation, September (2003): 1-153. Copulos, M R. "The Hidden Cost of Imported Oil~An
Update." The National Defense Council Foundation (2007).  Delucchi, Mark A. and James J. Murphy. "US military
expenditures to protect the use of Persian Gulf oil for motor  vehicles." Energy Policy 36, no. 6, 2008: 2253-2264,
Crane et al. /RAND, as above, Stern, Roger J. "United States cost of military force projection in the Persian Gulf,
1976-2007." Energy Policy 38, no. 6  (June 2010): 2816-2825.
http ://linkinghub. elsevier. com/retrieve/pii/S03 01421510000194.
81 "Transitions to Alternative Vehicles and Fuels," Committee on Transitions to Alternative Vehicles and Fuels,
National Research Council, 2013.
82 Farm, N., Baker, K.R., and Fulcher, C.M. (2012). Characterizing the PM2.5-related health benefits of emission
reductions for 17 area and mobile emission sectors across the  U.S., Environment International, 49, 241-151,
published online September 28, 2012.
83 U.S. Environmental Protection Agency (U.S. EPA).  (2012/ Regulatory Impact Analysis: Final Rulemaking for
2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy
Standards, Assessment and Standards Division, Office of Transportation and Air Quality, EPA-420-R-12-016,
August 2012. Available on the Internet at: http://www.epa.gov/otaq/climate/documents/420rl2016.pdf.
84 U.S. Environmental Protection Agency (U.S. EPA).  (2013). Regulatory Impact Analysis for the Reconsideration
of the Existing Stationary Compression Ignition (CI) Engines NESHAP, Office of Air Quality Planning and
Standards, Research Triangle Park, NC. January. EPA-452/R-13-001. Available at
http://www.epa.gov/ttnecasl/regdata/RIAs/RICE NESHAPreconsideration Compression Ignition Engines RIAfi
na!2013 EPA.pdf.
85 U.S. Environmental Protection Agency (U.S. EPA).  (2013). Regulatory Impact Analysis for Reconsideration of
Existing Stationary Spark Ignition (SI) RICE NESHAP, Office of Air Quality Planning and Standards, Research
Triangle Park, NC.  January. EPA-452/R-13-002. Available  at
http://www.epa.gov/ttnecasl/regdata/RIAs/NESHAP RICE Spark IgnitionRIA finarreconsideration2013 EPA.p
df.
86 U.S. Environmental Protection Agency (U.S. EPA).  (2015). Regulatory Impact Analysis for Residential Wood
Heaters NSPS Revision.  Office of Air Quality Planning and Standards, Research Triangle Park, NC.  February.
EPA-452/R-15-001.  Available at http://www2.epa.gov/sites/production/files/2015-02/documents/20150204-
residential-wood-heaters-ria.pdf.
87 U.S. Environmental Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter, Health and Environmental Impacts Division, Office
of Air Quality Planning and Standards, EPA-452-R-12-005,  December 2012. Available on the internet:
http://www.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf.
88 Farm, N., Baker, K.R., and Fulcher, C.M. (2012). Characterizing the PM2.5-related health benefits of emission
reductions for 17 industrial, area and mobile emission sectors across the U.S., Environment International, 49, 241-
151, published online September 28, 2012.
89 U.S. Environmental Protection Agency (U.S. EPA).  (2009). Integrated Science Assessment for Particulate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment—RTF Division. December.
Available at http://cfpub.epa.gov/ncea/cfm/recordisplav.cfm?deid=216546.

                                                     10-61

-------
                               Economic and Other Key Inputs Used in the Agencies' Analyses
90 U.S. Environmental Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter, Health and Environmental Impacts Division, Office
of Air Quality Planning and Standards, EPA-452-R-12-005, December 2012.  Available on the internet:
http://www.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf.
91 Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact Analysis
Under Executive Order 12866, Interagency Working Group on Social Cost of Carbon, with participation by Council
of Economic Advisers, Council on Environmental Quality, Department of Agriculture, Department of Commerce,
Department of Energy, Department of Transportation, Environmental Protection Agency, National Economic
Council, Office of Energy and Climate Change, Office of Management and Budget, Office of Science and
Technology Policy, and Department of Treasury (May 2013, Revised July 2015). Available at:
 Accessed 7/11/2015.
92 See https://www.whitehouse.gov/omb/oira/social-cost-of-carbon for both TSDs.
93 See July 2015 Response to Comments document on the SCC,
https://www.whitehouse.gov/sites/default/files/omb/inforeg/scc-response-to-comments-final-july-2015.pdf. See also
(1) Endangerment and Cause or Contribute Findings for Greenhouse Gases under Section 202(a) of the Clean Air
Act, 74 Fed. Reg. 66,496, 66,535 (Dec. 15, 2009) and (2) National Research Council. 2013. Climate and Social
Stress: Implications for Security Analysis. National Academies Press. Washington, DC.
94 Clean Power Plan final rule, see 80 FR 64661; 10/23/15. See also Clean Power Plan Response to Comments,
Section 8.7.2, Document ID EPA-HQ-OAR-2013-0602-37106..
95 U.S. Environmental Protection Agency (U.S. EPA). 2012.  Regulatory Impact Analysis Final New Source
Performance Standards and Amendments to the National Emissions Standards for Hazardous Air Pollutants for the
Oil and Natural Gas Industry. Office of Air Quality Planning and Standards, Health and Environmental Impacts
Division. April. < http://www.epa.gov/ttn/ecas/regdata/RIAs/oil_natural_gas_final_neshap_nsps_ria.pdf >.
96 Marten, A. L., E.  A. Kopits, C. W. Griffiths, S. C. Newbold & A. Wolverton (2014, online publication; 2015,
print publication). Incremental CH4 and N2O mitigation benefits consistent with the U.S. Government's SC-CO2
estimates, Climate Policy, DOI: 10.1080/14693062.2014.912981.
97 Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change [Stacker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K.
Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge,
United Kingdom and New York,  NY, USA.
98 For example, see  (1) U.S. Environmental Protection Agency (U.S. EPA). (2012). Regulatory Impact Analysis:
Final Rulemaking for 2017-2025  Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average
Fuel Economy Standards, Assessment and Standards Division, Office of Transportation and Air Quality, EPA-420-
R-12-016, August 2012.  Available on the Internet at: http://www.epa.gov/otaq/climate/documents/420rl2016.pdf.
and (2) U.S. Environmental Protection Agency (U.S. EPA). 2012. Regulatory Impact Analysis Final New Source
Performance Standards and Amendments to the National Emissions Standards for Hazardous Air Pollutants for the
Oil and Natural Gas Industry. Office of Air Quality Planning and Standards, Health and Environmental Impacts
Division. April. < http://www.epa.gov/ttn/ecas/regdata/RIAs/oil_natural_gas_final_neshap_nsps_ria.pdf>.
99 Office of Management and Budget (2003).  "Circular A-4." https://www.whitehouse.gov/omb/circulars_a004_a-
4/.
100 office of Management and Budget (2015). "Circular A-94 Appendix C, Revised November 2015."
https://www.whitehouse.gov/omb/circulars_a094/a94_appx-c.
101 "Joint Technical Support Document: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards," EPA-420-R-12-901, August 2012.
102 "During Recession, American Drivers Assumed More Risk to  Reduce Auto Insurance Costs," Quality Planning,
March 2011. See https://www.qualitvplanning.com/media/4312/110329%20tough%20times f2.pdf (last accessed
April 4,  2012).
103 U.S. Department of Commerce, Bureau of Economic Analysis, Table 7.2.5S. Auto and Truck Unit Sales,
Production, Inventories, Expenditures, and Price, Available at (www.bea.gov. last accessed May 4,
2012)http://www.bea.gov/national/nipaweb/nipa  underlying/Table7.2.5shttp://www.bea.gov/national/nipaweb/nipa
 underlying/Table7-025S.txt (last accessed May 4 July 11, 20162), see docket item "BEA Data 7_9_12.xlsx."
104 http://www.bls.gov/cpi/cpidllav.pdf. Table 1A. Consumer Price Index for All Urban Consumers (CPI-U): U.S.
city average, by  expenditure category and commodity and service group, for new vehicles.
                                                     10-62

-------
                                                       Credits, Incentives and Flexibilities
Table of Contents

Chapter 11: Credits, Incentives and Flexibilities	11-2
   11.1   Overview	11-2
   11.2   Averaging, Banking, and Trading Provisions	11-3
   11.3   Air Conditioning System Credits	11-4
   11.4   Off-cycle Technology Credits	11-5
   11.5   Incentives for Advanced Technology Vehicles	11-6
   11.6   Advanced Technology Incentives for Large Pickups	11-8
   11.7   Harmonized CAFE Incentives and Flexibilities	11-9

Table of Tables
Table 11.1 Incentive Multipliers for EV, FCV, PHEVs, and CNG Vehicles	11-6
Table 11.2 Penetration Rate Requirements by Model Year for Full-size Pickup Credits (% of Production)	11-9

-------
                                                      Credits, Incentives and Flexibilities
Chapter llrCredits, Incentives and Flexibilities

11.1   Overview

   The National Program was designed with a wide range of optional flexibilities to allow
manufacturers to maintain consumer choice, spur technology development, and minimize
compliance costs, while achieving significant GHG and oil reductions.  The National Program
also includes several EPA temporary incentives that encourage the use of advanced technologies
such as electric, hybrid, and fuel cell vehicles and these vehicles are also included in the
performance calculations for CAFE. This section provides an overview of all of these
compliance flexibilities.

   Averaging, banking, and trading (ABT) provisions, including credit carry-forward  and carry-
back provisions, define how credits may be used and are integral to the program. ABT
provisions are described in Chapter 11.2.  Credits for improvements to air conditioning systems
that increase efficiency and reduce refrigerant leakage, and credits for using technologies that
reduce emissions and improve fuel consumption that aren't captured on EPA tests ("off-cycle"
technologies) are discussed in Chapter 11.3 and 11.4, respectively. These credit opportunities
currently do not sunset, remaining a part of the program through MY2025 and beyond unless the
program is changed as part of a future regulatory action.

   As noted above, the GHG program includes temporary incentives for advanced technology
vehicles including incentives for large pickups using advanced technologies.  The CAFE
program also includes credits for large pickups using advanced technologies.  These provisions
are described below in Chapter 11.4 and 11.5. In the final rule, the agencies recognized that
temporary regulatory incentives will reduce the  short-term benefits of the program,  but believed
that it is worthwhile to have a limited short-term loss of benefits to increase the potential for far-
greater game changing benefits in the longer run. The agencies also believed  that the  temporary
regulatory incentives may help bring some technologies to market more quickly than in the
absence of incentives.1

   The use of the optional credit and incentive provisions varies from manufacturer to
manufacturer (some manufacturers have not availed themselves of the extra credit options, while
others have used some combination of, or all, options available under the regulations).2
Although a manufacturer's use of the credit and incentive provisions is optional, EPA projected
that the standards would be met on a fleet-wide  basis by using a combination of reductions in
tailpipe CCh and some use of the additional optional credit and incentive provisions in the
regulations.3  NHTSA is limited by its statutory authority to not include credits flexibilities in the
setting of CAFE standards.

   The discussion in this chapter is focused on compliance flexibilities which are integral to the
National Program. There are numerous other programs at the national, state, and local level
which provide incentives to consumers and manufacturers to develop, produce, and buy vehicles
with advanced technologies for reducing emissions and oil use. For example, tax incentives and
                                              11-2

-------
                                                        Credits, Incentives and Flexibilities
HOV lane access to incentive the purchase of electrified vehicles, and loan programs to
encourage investment in the development and manufacturing of advanced technologies^

11.2  Averaging, Banking, and Trading Provisions

   Both the CAFE and GHG programs include provisions for how credits may be used within the
programs. These averaging, banking, and trading (ABT) provisions include credit carry-forward,
credit carry-back (also called deficit carry-forward), credit transfers  (within a manufacturer), and
credit trading (across manufacturers). Credit carry-forward refers to banking (saving) credits for
future use, after satisfying any needs to offset pre-existing debits within a vehicle category (car
fleet or truck fleet).  Credit carry-back refers to using credits to offset any deficit in meeting the
fleet average standards that had accrued in a prior model year.  A manufacturer may have a
deficit at the end of a model year (after averaging across its fleet using credit transfers between
cars and trucks)—that is, a manufacturer's fleet average level may fail to meet the required fleet
average standard.  The EPCA/EISA statutory framework for the CAFE program limits credit
carry-forward to 5 years and credit carry-back to 3 years. Although  the Clean Air Act does not
include such limitations on the duration of credit provisions, in the MYs 2012-2016 and 2017-
2025 programs, EPA chose to adopt 5-year credit carry-forward (generally, with  an exception
noted below) and 3-year credit carry-back provisions as a reasonable approach that maintained
consistency between the agencies' provisions.

   Although the credit carry-forward and carry-back provisions generally remain in place for
MY2017  and later, EPA finalized provisions allowing all unused (banked) credits generated  in
MY2010-2016 (but not MY2009 early credits) to be carried forward through MY2021.  See §
86.1865-12(k)(6)(ii) and 77 FR 62788. This amounts to the normal 5 year carry-forward for
MY2016  and later credits, but provides additional carry-forward years for credits generated in
MYs 2010-2015.  Extending the life for MY2010-2015 credits provides greater flexibility for
manufacturers in using the credits they have generated. This provision helps facilitate the
transition to increasingly more stringent standards through MY2021  by helping manufacturers
resolve lead-time issues they might face in the early model years of the program. The one-time
extension of credit carry-forward also provides additional incentive for manufacturers to
generate credits earlier, for example in MYs 2014 and 2015, thereby encouraging the earlier  use
of additional CCh reducing technologies.  It does not change the overall CCh benefits of the
National Program, as EPA would not expect that any of the credits at issue would otherwise have
A The Advanced Technology Vehicles Manufacturing (ATVM) Loan Program provides long-term, low-interest rate
  loans to support the domestic manufacturing of advanced technology vehicles and automotive components. The
  ATVM Loan Program is administered by the U.S. Department of Energy's (DOE) Loan Programs Office (LPO).
  It was authorized concurrently with the first Congressionally-mandated increase in CAFE standards in thirty years
  and was designed to ensure that rising fuel economy standards did not disadvantage domestic manufacturing.
ATVM can finance a wide range of project costs, including the construction of new manufacturing facilities;
  retooling, reequipping, modernizing, or expanding an existing facility in the U.S; and the engineering integration
  costs necessary to manufacture eligible vehicles and components.
With more than $16 billion in remaining loan authority, the ATVM program can provide the financing needed to
  support the manufacturing of fuel-efficient technologies and components. By comparison, commercial lenders
  may be unwilling to lend at rates that allow automakers and suppliers to fully build out manufacturing capacity or
  ensure that new facilities are located in the U.S.
                                                11-3

-------
                                                     Credits, Incentives and Flexibilities
been allowed to expire. Rather, the credits would be used or traded for use by other
manufacturers.

   Transferring credits in the EPA program refers to exchanging credits between the two
averaging sets, passenger cars and light trucks, within a manufacturer.  For CAFE, credit
transfers can occur between compliance fleets (i.e., domestic and import passenger cars and light
trucks). For example, credits accrued by over-compliance with a manufacturer's car fleet
average standard could be used to offset debits accrued due to that manufacturer not meeting the
truck fleet average standard in a given year. (Put another way, a manufacturer's car and truck
fleets are, in essence, a single averaging set in the EPA program).  For NHTSA, transferring
credits between compliance fleets is possible but must to done using an adjustment which
ensures "total oil savings" are preserved because of differences in CAFE performance and
standards for compliance fleets and the amount of credits which can be transferred are capped by
statutory requirements.

   Finally, accumulated credits may be traded to another manufacturer.  Credit trading is now
occurring on a regular basis for the first time in an EPA vehicle program and has existed for
NHTSA since 2011. As of the end of MY2014, four manufacturers have sold credits and three
manufacturers  have purchased credits under the EPA program.4  For NHTSA, since 2011, six
manufacturers  have traded 151 million (unadjusted) CAFE credits. Manufacturers are acquiring
credits to offset immediate credit shortfalls and to bank for future compliance use.  As standards
become more stringent and total credit shortfalls increase, NHTSA projects an increase in credit
trades and carry-forwards and a reduction in civil penalty payments as a result of these changes
in flexibility usage.

   The EPA ABT provisions are generally consistent with those included in the CAFE program,
with a few notable exceptions. As with EPA's approach (except for the provision just discussed
above for a one-time extended carry-forward of MY2010-2016 credits), under EISA, credits
generated in the CAFE program can be carried forward for 5 model years or back for 3 years,
and can also be transferred between a manufacturer's fleets or traded to another manufacturer.
Transfers of credits across a manufacturer's car and truck compliance fleets are  also allowed
under CAFE, but with  limits established by EISA on the use of transferred credits.  The amount
of transferred credits that can be used in a year is limited under CAFE,  and transferred credits
may not be used to meet the CAFE minimum domestic passenger car standard, also per statute.
CAFE allows credit trading, but again, traded credits cannot be used to meet the minimum
domestic passenger car standard.5

   The ABT provisions are an integral part of the GHG and CAFE programs and the agencies
expect that manufacturers will continue to fully utilize these provisions into the  future. EPA's
annual GHG Manufacturers Performance Report provides details on the use of these provisions
in the GHG program thus far.6 Details on final compliance for model year 2014 for the NHTSA
and EPA programs are also summarized in Chapter 3.

11.3  Air Conditioning System Credits

   There are two mechanisms by which air conditioning (A/C) systems contribute to the
emissions of greenhouse gases: through leakage of hydrofluorocarbon refrigerants into the
atmosphere (sometimes called "direct emissions") and through the consumption of fuel to
provide mechanical power to the A/C system (sometimes called "indirect emissions").7  The high
                                             11-4

-------
                                                     Credits, Incentives and Flexibilities
global warming potential of the current automotive refrigerant, HFC-134a, means that leakage of
a small amount of refrigerant will have a far greater impact on global warming than emissions of
a similar amount of CCh.  The impacts of refrigerant leakage can be reduced significantly by
systems that incorporate leak-tight components, or, ultimately, by using a refrigerant with a
lower global warming potential.  The A/C system also contributes to increased tailpipe CCh
emissions through the additional work required to operate the compressor, fans, and blowers.
This additional power demand is ultimately met by using additional fuel, which is converted into
CCh by the engine during combustion and exhausted through the tailpipe.  These emissions can
be reduced by increasing the overall efficiency of an A/C system, thus reducing the additional
load on the engine from A/C operation, which in turn means a reduction in fuel consumption and
a commensurate reduction in GHG emissions.

   Manufacturers may generate credits for improved A/C systems in complying with the CCh
fleet average standards in the MY2012 and later model years. Manufacturers may generate fuel
consumption improvement credits for A/C efficiency improvement under the CAFE program
equivalent to the CCh credits beginning in MY2017. EPA expected manufacturers to generate
A/C credits and accounted for those credits in developing the final CCh standards by adjusting
the standards to make them more stringent. EPA's A/C credits program is also related to EPA
action under the Significant New Alternatives Policy (SNAP) program which on July, 20, 2015
changed the listing status of HFC-134a to unacceptable for newly manufactured light-duty
vehicles beginning with MY2021 due to the refrigerant's high global warming potential (GWP).8
This action effectively requires auto manufacturer's  to choose an alternative refrigerant with a
lower GWP beginning with MY2021. Prior to MY2021, the use of low GWP refrigerants in
light-duty vehicles is encouraged by EPA's credit program. A detailed discussion of A/C credits
and technologies is provided in Chapter 5.2.

11.4 Off-cycle Technology Credits

   "Off-cycle" emission reductions can be achieved by employing technologies that result in
real-world benefits, but where that benefit is not adequately captured on the test procedures used
by manufacturers to demonstrate compliance with emission standards.  EPA's light-duty vehicle
greenhouse gas program acknowledges these benefits by giving automobile manufacturers
several options for generating "off-cycle" technology CCh  credits.  Starting in MY2017,
manufacturers may also generate equivalent fuel consumption improvement credits in the CAFE
program.

   There are three pathways by which a manufacturer may accrue off-cycle technology  credits.
The first is a predetermined list or "menu" of credit  values for specific off-cycle technologies
that may be used beginning for model year 2014.9 This pathway allows manufacturers to use
conservative credit values established by EPA for a  wide range of off-cycle technologies, with
minimal data submittal or testing requirements. In cases where additional laboratory testing can
demonstrate emission benefits, a second pathway allows manufacturers to use a broader array of
emission tests (known as "5-cycle" testing because the methodology uses five different  testing
procedures) to demonstrate and justify off-cycle CCh credits.10  The additional emission tests
allow emission benefits to be demonstrated over some elements of real-world driving not
captured by the GHG compliance tests, including high speeds, rapid accelerations, and cold
temperatures. Credits determined according to this methodology do not undergo additional
public review.  The third and last pathway allows manufacturers to seek EPA approval to use an
                                             11-5

-------
                                                      Credits, Incentives and Flexibilities
alternative methodology for determining the off-cycle technology CCh credits.11 This option is
only available if the benefit of the technology cannot be adequately demonstrated using the 5-
cycle methodology. Manufacturers may also use this option for model years prior to 2014 to
demonstrate off-cycle CCh reductions for off-cycle technologies that are on the menu, or to
demonstrate reductions that exceed those available via use of the menu.  As with other emissions
controls, off-cycle technologies are subject to full useful life requirements.

   Chapter 5.2 provides a detailed description of the off-cycle technology program including
what off-cycle technologies manufacturers have used to date to generate credits and the
magnitude of those credits. Chapter 5.2 also discusses how the agencies have considered off-
cycle credits in the Draft TAR analysis.

11.5  Incentives for Advanced Technology Vehicles

   EPA included incentives for advanced technologies to promote the commercialization of
technologies that have the potential to transform the light-duty vehicle sector by achieving zero
or near-zero GHG emissions and oil consumption in the longer term, but which face major near-
term market barriers. Providing temporary regulatory incentives for certain advanced
technologies will decrease the overall GHG emissions reductions  associated with the program in
the near term.  However, in setting the 2017-2025 standards, EPA believed it is worthwhile to
forego modest additional emissions reductions in the near term in order to lay the foundation for
the potential for much larger "game-changing" GHG emissions and oil reductions in the longer
term. EPA also believed that temporary regulatory incentives may help bring some technologies
to market more quickly than in the absence of incentives.  See  77  FR 62811 et seq. EPA
accounts for the higher real world GHG emissions and lower GHG emissions reductions
associated with these temporary regulatory incentives in all of our regulatory analyses, as well as
in this Draft TAR.

   A multiplier incentive is available for MY2017-2021 electric vehicles (EVs), plug-in hybrid
electric vehicles (PHEVs), fuel cell vehicles (FCVs) and compressed natural gas (CNG)
vehicles.12  The multiplier allows a vehicle to "count" as more than one vehicle in  the
manufacturer's compliance calculation. Table 11.1 provides the multipliers for the various
vehicle technologies included in the 2012 final rule for MY2017-2021 vehicles.13  Since the
GHG performance for these vehicle types is  significantly better than that of conventional
vehicles, the multiplier provides a significant benefit to the manufacturer. The specific
multiplier levels were picked to be large enough to provide a meaningful incentive, but not be so
large as to promote vehicles being produced only to take advantage of the incentive. The
multipliers for EVs and FCVs are larger because they face greater market barriers.
              Table 11.1 Incentive Multipliers for EV, FCV, PHEVs, and CNG Vehicles
Model Years
2017-2019
2020
2021
EVs and FCVs
2.0
1.75
1.5
PHEVs and CNG
1.6
1.45
1.3
   Although EPA does not view CNG as a game changing technology from a GHG tailpipe
emissions perspective, EPA included a multiplier incentive for dedicated and dual-fueled CNG
vehicles because EPA considered investments in CNG technology and refueling infrastructure to
                                              11-6

-------
                                                     Credits, Incentives and Flexibilities
be a valuable, indirect step towards hydrogen FCVs, which can be a game-changer in terms of
GHG emissions.14 In this way, EPA believed that CNG could be a critical facilitator of a next-
generation technology.

   EPA included a second incentive for EVs, PHEVs, and FCVs by allowing temporary and
limited 0 g/mile treatment of the electric operation of those vehicles.15  The tailpipe GHG
emissions from EVs, from PFIEVs operated on grid electricity, and from hydrogen-fueled FCVs
are zero,  and traditionally the emissions of the vehicle itself are all that EPA takes into account
for purposes of compliance with standards set under Clean Air Act section 202(a). Focusing on
vehicle tailpipe emissions has not raised any issues for criteria pollutants, as upstream criteria
emissions associated with production and distribution of the fuel are addressed by
comprehensive regulatory programs focused on the upstream sources of those emissions. At the
time of the final rule,  however, there was no such comprehensive program addressing upstream
emissions of GHGs,16 and the upstream GHG emissions associated with production and
distribution of electricity are higher, on a national average basis, than the corresponding
upstream GHG emissions of gasoline or other petroleum based fuels.

   Therefore, EPA placed limits on the use of 0 g/mile for MY2022-2025 vehicles and the use of
0 g/mile is currently not allowed after MY2025.  EPA included per-company vehicle production
caps for use of 0 g/mile in MYs 2022-2025, and 0 g/mile cannot be used for production that
exceeds these caps. The cumulative per-company caps for MYs 2022-2025 are 600,000
EV/PHEV/FCVs for those manufacturers that produce a total of 300,000 or more
EV/PHEV/FCVs in MYs 2019-2021, and 200,000 EV/ PHEV/FCVs for all other manufacturers.
The structure of these per-company caps was based on a balancing of promoting game-changing
technologies, while minimizing the short-term loss in overall GHG savings.  Once the production
cap is met, the manufacturer must include net upstream emissions associated with electricity
generation on a g/mile basis in their compliance calculations.  Currently, U.S. annual sales of
advanced technology  vehicles are well below the per manufacturer thresholds.  Tesla's 2015
annual sales are estimated to be just under 26,000 vehicles and GM, Ford, and Nissan 2015 sales
were in the 17,000-19,000 vehicles per year range.17

   The final rule provides a methodology for determining the net upstream GHG emissions value
to be assigned to a vehicle for purposes of vehicle certification and compliance calculations.18
EPA concluded in the MY2017-2025 final rule that the "compliance treatment finalized for
EV/PHEV/FCVs strikes a reasonable balance between promoting the commercialization of
EV/PHEV/FCVs, which have the potential to achieve game-changing GHG emissions reductions
in the future, and accounting for upstream emissions once  such vehicles reach a reasonable
threshold in the market." 19

   EPA recognized that the mid-term evaluation would provide an opportunity to review the
status of advanced vehicle technology commercialization,  the status of upstream GHG emissions
control programs, and other relevant factors.20 At the time of the MY2017-2025 final rule, part
of the rationale for including upstream emissions associated with electricity production, for
production volumes in excess of the per-company production volume caps, was because these
upstream GHG emissions values are generally higher than the upstream GHG emissions values
associated with gasoline vehicles, and because there was then no federal program in place to
reduce GHG emissions from electric power  plants.  EPA also stated that in the future, if there
were a program to comprehensively address upstream GHG emissions, then the zero tailpipe
                                             11-7

-------
                                                      Credits, Incentives and Flexibilities
levels from these vehicles have the potential to contribute to very large GHG reductions, and to
transform the transportation sector's contribution to nationwide GHG emissions (as well as oil
consumption).

   Since the MY2017-2025 final rule, EPA has adopted GHG controls for electricity generation.
On August 3, 2015, EPA issued final GHG emissions regulations addressing both existing
(referred to as the Clean Power Plan21) and new electricity generating units. These rules are
expected to markedly decrease GHG emissions associated with future electricity generation. In
the MY2017-2025 final rule, EPA used the Office of Atmospheric Programs' Integrated Planning
Model, along with assumptions for the 2030 timeframe about total light-duty vehicle demand for
electricity, geographical distribution of EVs and PHEVs, and on-peak versus off-peak charging,
to project that the average power plant electricity GHG emissions factor in 2030 for vehicle
electricity use would be 0.445 grams/watt-hour.22  The overall vehicle electricity GHG emissions
factor was projected to be 0.534 grams/watt-hour when using a multiplicative value of 1.20 to
account for feedstock-related GHG emissions upstream of the power plant. EPA is currently
exploring whether there are appropriate updates to these projected emissions factors for the
incremental electricity that would be necessary for electric vehicle operation in the 2030
timeframe, which we plan to assess in more detail further during the midterm evaluation process.
EPA also plans to develop a similar methodology for net upstream GHG emissions associated
with hydrogen fuel production and distribution.

11.6  Advanced Technology Incentives for Large Pickups

   The agencies recognized that the MY2017-2025 standards will be challenging for large
vehicles, including full-size pickup trucks that are often used for commercial purposes.  In the
MY2017-2025 final rule, EPA and NHTSA included a per-vehicle credit provision for
manufacturers that hybridize a significant number of their full-size pickup trucks, or use other
technologies that comparably reduce CO2 emissions and fuel consumption. The agencies' goal
was to incentivize the penetration into the marketplace of "game changing" technologies for
these pickups. The incentives provide an opportunity in the program's early years to begin
penetration of advanced technologies into this category of vehicles, and in turn creates more
opportunities for achieving the more stringent later year standards. Full-size pickup trucks using
mild hybrid technology will be eligible for a per-truck 10 g/mi CO2 credit  (equivalent to 0.0011
gal/mi for a gasoline-fueled truck) during MYs 2017-2021.  Full-size pickup trucks using strong
hybrid technology will be eligible for a per-truck 20 g/mi CO2 credit (0.0023 gal/mi) during MYs
2017-2025.23 Eligibility for both the mild and strong hybrid credit is dependent on the
manufacturer reaching the minimum technology penetration thresholds discussed below. The
agencies established definitions for full-size pickup and mild and strong hybrid for the
program.24

   Alternatively, manufacturers may generate performance-based credits for full-size pickups.
This performance-based credit is 10 g/mi CO2 (equivalent to 0.0011 gal/mi for the CAFE
program) or 20 g/mi CO2 (0.0023 gal/mi) for full-size pickups achieving 15 percent or 20
percent, respectively, better CO2 than their footprint-based targets in a given model year.25'26
This second option incentivizes other, non-hybrid, advanced technologies that can reduce pickup
truck GHG emissions and fuel consumption at rates comparable to strong and mild hybrid
technology.  These performance-based credits have no specific technology or design
requirements; automakers can use any technology or set of technologies as long as the vehicle's
                                              11-8

-------
                                                      Credits, Incentives and Flexibilities
CCh performance is at least 15 or 20 percent below the vehicle's footprint-based target.
However, a vehicle cannot receive both hybrid and performance-based credits, since that would
be double-counting.

   The 10 g/mi performance-based credit is available for MYs 2017 to 2021. In recognition of
the nature of automotive redesign cycles, a vehicle model meeting the requirements in a model
year will receive the credit in subsequent model years through 2021 unless its CCh level
increases or its production level drops below the penetration threshold described below, even if
the year-by-year reduction in standards levels causes the vehicle to fall below the 15 percent
over-compliance threshold. Not doing so would reduce substantially the incentive to introduce
advanced technology in earlier model years if the incentive wasn't available for the design cycle
period.  The 10 g/mi credit is not available after MY2021 because the stringency of the post-
MY2021 standards quickly overtake designs that were originally 15 percent over-compliant,
making the awarding of credits to them inappropriate.  See also 80 FR at 40253 (advanced
technology credits from phase 1  heavy duty GHG  rules inappropriate for phase 2, since these
technologies are now part of the compliance basis  for the proposed phase 2 standards).  The 20
g/mi CCh performance-based credit will  be available for a maximum of 5 consecutive model
years (the typical redesign cycle period)  within the 2017 to 2025 model year period, provided the
vehicle model's CCh level does not increase from the level determined in its first qualifying
model year, and subject to the technology penetration requirement described below. A
qualifying vehicle model that subsequently undergoes a major redesign can requalify for the
credit for an additional period starting in the redesign model year, not to exceed 5 model years
and not to extend beyond MY2025.27

   Access to any of these large pickup credits requires that the technology be used on a
minimum percentage of a manufacturer's full-size pickups.28 These minimum percentages,
established in the 2012 final rule, are set to encourage significant penetration of these
technologies, leading to long-term market acceptance.  Meeting the penetration threshold in one
model year does not ensure credits in subsequent years; if the production level in a model year
drops below the required threshold, the credit is not earned for that model year. The required
penetration levels are shown in the table below.29
  Table 11.2 Penetration Rate Requirements  by Model Year for Full-size Pickup Credits (% of Production)

Strong hybrid
Mild Hybrid
20% better performance
15% better performance
2017
10
20
10
15
2018
10
30
10
20
2019
10
55
10
28
2020
10
70
10
35
2021
10
80
10
40
2022
10
N/A
10
N/A
2023
10
N/A
10
N/A
2024
10
N/A
10
N/A
2025
10
N/A
10
N/A
11.7  Harmonized CAFE Incentives and Flexibilities

   Since issuing standards in the October 2012 final rule (see 77 FR 62624) for model year 2017
to 2025 light duty vehicles, the Alliance of Automobile Manufacturers and some individual
automobile manufacturers have reached out to NHTSA to discuss several programmatic
differences between NHTSA's CAFE and EPA's GHG programs.  Many of the incentives and
flexibilities available under the EPA program are not statutorily available to the CAFE program
because of prescribed limitations establish by Congress in EISA and EPCA. The issues
                                             11-9

-------
                                                      Credits, Incentives and Flexibilities
identified by the Alliance are contained in a presentation shared with NHTSA, available in
NHTSA's docket.
                                             11-10

-------
                                                             Credits, Incentives and Flexibilities
References
1 77 FR 62812, October 15, 2012.
2  Greenhouse Gas Emission Standards for Light-duty Vehicles, Manufacturer Performance Report for the 2014
Model Year, EPA-420-R-15-026, December 2015.
3 See Tables III-2 and III-3, 77 FR 62772, October 15, 2012.
4 Greenhouse Gas Emission Standards for Light-duty Vehicles, Manufacturer Performance Report for the 2014
Model Year, EPA-420-R-15-026, December 2015.
5 See generally 49 U.S.C. § 32903.
6 Greenhouse Gas Emission Standards for Light-duty Vehicles, Manufacturer Performance Report for the 2014
Model Year, EPA-420-R-15-026, December 2015.
7 40 CFR 1867-12 and 40 CFR 86.1868-12.
8 80 FR 42870, July 20, 2015.
9See40CFR86.1869-12(b).
10See40CFR86.1869-12(c).
11 See40CFR86.1869-12(d).
12 77 FR 62810, October 15, 2012.
13 77 FR 62813-62816, October 15, 2012.
14 77 FR 62815-62816, October 15, 2012.
15 77 FR 62816, October 15, 2012.
16 EPA had proposed but had not yet adopted a New Source Performance Standard for greenhouse gas emissions
from new electricity generating units, see 77 FR 22392.
17 Monthly Plug-In Sales Scorecard, Insideevs.com, June 8, 2016.
18 77 FR 62820-62822, October 15, 2012.
19 77 FR 62820, October 15, 2012.
20 77 FR 62820, October 15, 2012.
1 80 FR 64661; October 23, 2015.
2 77 FR 62821, October 15, 2012.
3 77 FR 62825, October 15, 2012.
4 77 FR 62825, October 15, 2012.
5 77 FR 62826, October 15, 2012.
6 For additional discussion of the performance thresholds, see Section 5.3.4 of the "Joint Technical Support
Document: Final Rulemaking for 2017-2025 Light-duty Vehicle Greenhouse Gas Emission Standards and Corporate
Average Fuel Economy Standards" for the Final Rule," EPA-420-R-12-901, August 2012.
27 77 FR 62826, October 15, 2012.
28 77 FR 62826, October 15, 2012.
29 40 CFR 86.1870-12.
                                                   11-11

-------
                                    EPA's Analysis of the MY2022-2025 GHG Standards
Table of Contents

Chapter 12: EPA's Analysis of the MY2022-2025 GHG Standards	12-1
  12.1   EPA's Estimates of Costs per Vehicle & Technology Penetrations Based on OMEGA
         12-2
     12.1.1  Central Analysis Results	12-5
       12.1.1.1  CCh Targets and Achieved Values	12-5
         12.1.1.1.1  Reference Case	12-5
         12.1.1.1.2  Control Case	12-6
         12.1.1.1.3  Off-Cycle, Pickup Incentive and A/C Credits in OMEGA	12-10
         12.1.1.1.4  Projected 2-Cycle CO2	12-12
       12.1.1.2  Cost per Vehicle	12-14
         12.1.1.2.1  Reference & Control Case	12-14
       12.1.1.3  Technology Penetration	12-18
         12.1.1.3.1  Reference Case	12-18
         12.1.1.3.2  Control Case	12-24
       12.1.1.4  Comparisons to the 2012 Final Rule	12-34
     12.1.2  Sensitivity Analysis Results	12-36
       12.1.2.1  Reference Case: CO2 Targets	12-36
       12.1.2.2  Control Case:  CO2 Targets	12-37
       12.1.2.3  Cost per Vehicle and Technology Penetrations	12-37
       12.1.2.4  Observations on Sensitivity Analyses	12-40
     12.1.3  Payback Period & Lifetime Savings	12-41
  12.2   EPA's Projected Impacts on Emissions Inventories & Fuel Consumption	12-47
     12.2.1  Analytical Tools Used	12-47
     12.2.2  Inputs to the Emissions and Fuel Consumption Analysis	12-47
       12.2.2.1  Methods	12-47
       12.2.2.2  Global Warming Potentials	12-48
       12.2.2.3  Years Considered	12-49
       12.2.2.4  Fleet Activity	12-49
         12.2.2.4.1  Vehicle Sales,  Survival  Schedules, and VMT	12-49
       12.2.2.5  Upstream Emission Factors	12-49
         12.2.2.5.1  Gasoline Production and Transport Emission Rates	12-49
         12.2.2.5.2  Electricity Generation Emission Rates	12-50
       12.2.2.6  Reference Case CO2 g/mi & kWh/mi	12-51
       12.2.2.7  Control Case CO2 g/mi & kWh/mi	12-53
       12.2.2.8  Criteria Pollutant and Select Toxic Pollutant Emission Rates	12-55
     12.2.3  Outputs of the Emissions and Fuel Consumption Analysis	12-56
       12.2.3.1  Calendar Year Results	12-57
       12.2.3.2  Model Year Lifetime Results	12-61
     12.2.4  Sensitivity Analysis Results	12-62
       12.2.4.1  Calendar Year Case Comparison Results	12-63
       12.2.4.1  Model Year Lifetime Case Comparison Results	12-63
  12.3   EPA's Benefit-Cost Analysis Results	12-64
     12.3.1  Model Year Analysis	12-64
       12.3.1.1  AEO 2015 Reference Fuel Price Case Using ICMs	12-64
       12.3.1.2  AEO 2015 Reference Fuel Price Case Using RPEs	12-67

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
        12.3.1.3  AEO 2015 High Fuel Price Case Using ICMs	12-69
        12.3.1.4  AEO 2015 Low Fuel Price Case Using ICMs	12-71
        12.3.1.5  Summary of MY Lifetime Benefit-Cost Analysis Results	12-74
     12.3.2  Calendar Year Analysis	12-74
        12.3.2.1  AEO 2015 Reference Fuel Price Case Using ICMs	12-75
        12.3.2.2  AEO 2015 Reference Fuel Price Case Using RPEs	12-76
        12.3.2.3  AEO 2015 High Fuel Price Case Using ICMs	12-76
        12.3.2.4  AEO 2015 Low Fuel Price Case Using ICMs	12-77
        12.3.2.5  Summary of CY Benefit-Cost Analysis Results	12-78
   12.4   Additional OMEGA Cost Analyses	12-79
     12.4.1  Cost per Vehicle Tables - Absolute and Incremental Costs	12-79
     12.4.2  Cost per Percentage Improvement in CO2	12-82

Table of Figures

Figure 12.1  Actual Standard Curves and the Control Case Target Curves Used for Cars in this Draft TAR Analysis
            to Reflect a 5-Year Redesign Cycle	12-7
Figure 12.2  Actual Standard Curves and the Control Case Target Curves Used for Trucks in this Draft TAR
            Analysis to Reflect a 5-Year Redesign Cycle	12-8


Table of Tables

Table 12.1 Reference Case Targets and Achieved CO2 in MY2021 in the Central Analysis (g/mi CO2)	12-6
Table 12.2 Reference Case Targets and Achieved CO2 in MY2025 in the Central Analysis (g/mi CO2)	12-6
Table 12.3 Control Case Targets and Achieved CO2 in MY2021 in the Central Analysis (g/mi CO2)	12-9
Table 12.4 Control Case Targets and Achieved CO2 in MY2025 in the Central Analysis (g/mi CO2)	12-10
Table 12.5 Off-cycle & Pickup Incentive Credits Available for Achieving the CO2 Targets (g/mi CO2)	12-11
Table 12.6 Off-cycle, Pickup Incentive and A/C Credits Used to Achieve the CO2 Targets (g/mi CO2)	12-11
Table 12.7 EPA Projections for Car Tailpipe Emissions  Compliance with CO2 Standards Using ICMs and the AEO
            Reference Fuel Price Case (CO2 g/mi)	12-12
Table 12.8 EPA Projections for Truck Tailpipe Emissions Compliance with CO2 Standards Using ICMs and the
            AEO Reference Fuel Price Case (CO2 g/mi)	12-12
Table 12.9 EPA Projections for Combined Fleet Tailpipe Emissions Compliance with CO2 Standards Using ICMs
            and the AEO Reference Fuel Price Case (CO2 g/mi)	12-12
Table 12.10 EPA Projections for Car Tailpipe Emissions Compliance with CO2 Standards Using ICMs and the
            AEO High Fuel Price Case (CO2 g/mi)	12-13
Table 12.11 EPA Projections for Truck Tailpipe Emissions Compliance with CO2 Standards Using ICMs and the
            AEO High Fuel Price Case (CO2 g/mi)	12-13
Table 12.12 EPA Projections for Combined Fleet Tailpipe Emissions Compliance with CO2 Standards Using ICMs
            and the AEO High Fuel Price Case (CO2 g/mi)	12-13
Table 12.13 EPA Projections for Car Tailpipe Emissions Compliance with CO2 Standards Using ICMs and the
            AEO Low Fuel Price Case (CO2 g/mi)	12-14
Table 12.14 EPA Projections for Truck Tailpipe Emissions Compliance with CO2 Standards Using ICMs and the
            AEO Low Fuel Price Case (CO2 g/mi)	12-14
Table 12.15 EPA Projections for Combined Fleet Tailpipe Emissions Compliance with CO2 Standards Using ICMs
            and the AEO Low Fuel Price Case (CO2 g/mi)	12-14
Table 12.16 MY2021 Control Case Cost/Vehicle Incremental to the Reference Case Cost/Vehicle in the Central
            Analysis Using AEO Reference Case Fuel Prices and Fleet Projections and Using both ICMs and
            RPEs (2013$)	12-15
Table 12.17 MY2025 Control Case Cost/Vehicle Incremental to the Reference Case Cost/Vehicle in the Central
            Analysis Using AEO Reference Case Fuel Prices and Fleet Projections and Using both ICMs and
            RPEs (2013$)	12-16
Table 12.18 MY2021-2025 Control Case Cost/Vehicle Incremental to the Reference Case Cost/Vehicle	12-17

-------
                                          EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.19 MY2021-2025 Control Case Cost/Vehicle Incremental to the Reference Case Cost/Vehicle	12-17
Table 12.20 Technology Code Definitions used in Technology Penetration Tables	12-19
Table 12.21 Absolute Technology Penetrations for Cars in the MY2025 Reference Case Using ICMs	12-20
Table 12.22 Absolute Technology Penetrations for Cars in the MY2025 Reference Case Using RPEs	12-20
Table 12.23 Absolute Technology Penetrations for Trucks in the MY2025 Reference Case Using ICMs	12-21
Table 12.24 Absolute Technology Penetrations for Trucks in the MY2025 Reference Case Using RPEs	12-21
Table 12.25 Absolute Technology Penetrations for the Fleet in the MY2025 Reference Case Using ICMs	12-22
Table 12.26 Absolute Technology Penetrations for the Fleet in the MY2025 Reference Case Using RPEs	12-22
Table 12.27 Summary of Absolute Technology Penetrations in the MY2025 Reference Case	12-23
Table 12.28 Percentage of Vehicles Receiving the Mass Reduction Levels within the Indicated Ranges in the
            MY2025 Reference Case Using ICMs and AEO Reference Case Fuel Prices	12-24
Table 12.29 Absolute Technology Penetrations for Cars in the MY2025 Control Case Using ICMs	12-27
Table 12.30 Absolute Technology Penetrations for Cars in the MY2025 Control Case Using RPEs	12-27
Table 12.31 Absolute Technology Penetrations for Trucks in the MY2025 Control Case Using ICMs	12-28
Table 12.32 Absolute Technology Penetrations for Trucks in the MY2025 Control Case Using RPEs	12-28
Table 12.33 Absolute Technology Penetrations for the Fleet in the MY2025 Control Case Using ICMs	12-29
Table 12.34 Absolute Technology Penetrations for the Fleet in the MY2025 Control Case Using RPEs	12-29
Table 12.35 Incremental Technology Penetrations for Cars in the MY2025 Central Analysis Using ICMs	12-30
Table 12.36 Incremental Technology Penetrations for Cars in the MY2025 Central Analysis Using RPEs	12-30
Table 12.37 Incremental Technology Penetrations for Trucks in the MY2025 Central Analysis Using ICMs.... 12-31
Table 12.38 Incremental Technology Penetrations for Trucks in the MY2025 Central Analysis Using RPEs.... 12-31
Table 12.39 Incremental Technology Penetrations for the Fleet in the MY2025 Central Analysis Using ICMs. 12-32
Table 12.40 Incremental Technology Penetrations for the Fleet in the MY2025 Central Analysis Using RPEs. 12-32
Table 12.41 Summary of Absolute Technology Penetrations in the MY2025 Control Case	12-33
Table 12.42 Summary of Incremental Technology Penetrations in the MY2025 Control Case	12-33
Table 12.43 Percentage of Vehicles Receiving the Mass Reduction Levels within the Indicated Ranges in the
            MY2025 Control Case Using ICMs and AEO Reference Case Fuel Prices	12-34
Table 12.44 Cost per Vehicle Comparison -2012 FRM (2010$) vs Draft TAR (2013$)	12-35
Table 12.45 Final Technology Penetration Comparison - 2012 FRM vs Draft TAR	12-35
Table 12.46 Reference Case CO2 Targets inMY2025 for Each Sensitivity Case (g/mi)	12-36
Table 12.47 Control Case CO2 Targets inMY2025 for Each Sensitivity Case (g/mi)	12-37
Table 12.48 MY2025 Absolute Technology Penetrations & Incremental Costs for Cars in Each OMEGA Run
            (2013$)	12-38
Table 12.49 MY2025 Absolute Technology Penetrations & Incremental Costs for Trucks in Each OMEGA Run
            (2013$)	12-39
Table 12.50 MY2025 Absolute Technology Penetrations & Incremental Costs for the Fleet in Each OMEGA Run
            (2013$)	12-40
Table 12.51 Payback Period for the Sales Weighted Average MY2025 Vehicle in the Central Analysis using ICMs
            Relative to the Reference Case Standards (3% discounting, 2013$)	12-42
Table 12.52 Payback Period for the Sales Weighted Average MY2025 Vehicle in the Central Analysis using RPEs
            Relative to the Reference Case Standards (3% discounting, 2013$)	12-43
Table 12.53 Payback Period for the Sales Weighted Average MY2025 Vehicle in the Central Analysis using ICMs
            Relative to the Reference Case Standards (7% discounting, 2013$)	12-43
Table 12.54 Payback Period for the Sales Weighted Average MY2025 Vehicle in the Central Analysis using RPEs
            Relative to the Reference Case Standards (7% discounting, 2013$)	12-44
Table 12.55 Payback Period for the Sales Weighted Average MY2025 Vehicle using AEO High Fuel Prices and
            ICMs Relative to the Reference Case Standards (3% discounting, 2013$)	12-44
Table 12.56 Payback Period for the Sales Weighted Average MY2025 Vehicle using AEO High Fuel Prices and
            ICMs Relative to the Reference Case Standards (7% discounting, 2013$)	12-45
Table 12.57 Payback Period for the Sales Weighted Average MY2025 Vehicle using AEO Low Fuel Prices and
            ICMs Relative to the Reference Case Standards (3% discounting, 2013$)	12-45
Table 12.58 Payback Period for the Sales Weighted Average MY2025 Vehicle using AEO Low Fuel Prices and
            ICMs Relative to the Reference Case Standards (7% discounting, 2013$)	12-46
Table 12.59 Lifetime Net Savings Associated with the Indicated Control Case Relative to the Reference  Case for
            the Sales-Weighted Average MY2025 Vehicle	12-46
Table 12.60 Global Warming Potentials (GWP) for Inventoried GHGs	12-48

-------
                                          EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.61 Gasoline Production Emission Rates	12-50
Table 12.62 Emission Factors Used in Analysis of Electricity Generation	12-51
Table 12.63 Reference Case Car On-Road CO2 g/mi Used in All OMEGA ICBT Runs	12-52
Table 12.64 Reference Case Truck On-Road CO2 g/mi Used in All OMEGA ICBT Runs	12-52
Table 12.65 Reference Case Car & Truck On-Road kWh/mi Consumption used in the Indicated OMEGA ICBT
            Runs	12-53
Table 12.66 Control Case Car On-Road CO2 g/mi Used in All OMEGA ICBT Runs	12-54
Table 12.67 Control Case Truck On-Road CO2 g/mi Used in All OMEGA ICBT Runs	12-54
Table 12.68 Reference Case Car & Truck On-Road kWh/mi Consumption used in the Indicated OMEGA ICBT
            Runs	12-55
Table 12.69 Annual Emissions Reductions of the MY2022-2025 Standards on GHGs in Select Calendar Years
            (MMTCO2e)	12-57
Table 12.70 Annual Emission Reductions of the MY2022-2025 Standards on GHGs (MMT CO2e)	12-58
Table 12.71 Annual Emission Reductions of the MY2022-2025 Standards on non-GHG Criteria Pollutants in Select
            Years	12-59
Table 12.72 Annual Emission Reductions of the MY2022-2025 Standards on Select Toxic Pollutants in Select
            Years	12-60
Table 12.73 Annual Impacts of the MY2022-2025 Standards on Fuel and Electricity Consumption	12-61
Table 12.74 MY Lifetime Emission Reductions of the MY2022-2025 Standards on GHGs (MMT CO2e)	12-61
Table 12.75 MY Lifetime Emission Reductions of the MY2022-2025 Standards on Select non-GHG Criteria
            Pollutants	12-62
Table 12.76 MY Lifetime Emission Reductions of the MY2022-2025 Standards on Select Toxic Pollutants.... 12-62
Table 12.77 MY Lifetime Impacts of the MY2022-2025 Standards onFuel and Electricity Consumption	12-62
Table 12.78 Annual Emission Reductions of the MY2022-2025 Standards and AEO Fuel Price Cases on Total
            GHGs (MMT CO2e)	12-63
Table 12.79 Annual Impacts of the MY2022-2025 Standards on Fuel Consumption	12-63
Table 12.80 MY Lifetime Emission Reductions of the MY2022-2025 Standards and AEO Fuel Price Cases on Total
            GHGs (MMT CO2e)	12-63
Table 12.81 MY Lifetime Impacts of the MY2022-2025 Standards and AEO Fuel Price Cases on Fuel Consumption
             	12-64
Table 12.82 MY Lifetime Costs & Benefits Using AEO Reference Fuel Prices and ICMs (3 Percent Discount Rate,
            Billions of 2013 $)a'b>c	12-65
Table 12.83 MY Lifetime Costs & Benefits Using AEO Reference Fuel Prices and ICMs, (7 Percent Discount Rate,
            Billions of 2013 $)a>b'c	12-66
Table 12.84 MY Lifetime Costs & Benefits Using AEO Reference Fuel Prices and RPEs (3 Percent Discount Rate,
            Billions of 2013 $)a>b'c	12-67
Table 12.85 MY Lifetime Costs & Benefits Using AEO Reference Fuel Prices and RPEs, (7 Percent Discount Rate,
            Billions of 2013 $)a>b'c	12-68
Table 12.86 MY Lifetime Costs & Benefits Using AEO High Fuel Prices and ICMs (3  Percent Discount Rate,
            Billions of 2013 $)a-b'c	12-70
Table 12.87 MY Lifetime Costs & Benefits Using AEO High Fuel Prices and ICMs (7  Percent Discount Rate,
            Billions of 2013 $)a-b'c	12-71
Table 12.88 MY Lifetime Costs & Benefits Using AEO Low Fuel Prices and ICMs (3 Percent Discount Rate,
            Billions of 2013 $)a-b'c	12-72
Table 12.89 MY Lifetime Costs & Benefits Using AEO Low Fuel Prices and ICMs (7 Percent Discount Rate,
            Billions of 2013 $)a'b'c	12-73
Table 12.90 MY Lifetime Costs & Benefits in the Central & Sensitivity Cases (Billions of 2013$)	12-74
Table 12.91 Annual Costs & Benefits Using AEO Reference Fuel Prices and ICMs (Billions of 2013$) a'b'c.... 12-75
Table 12.92 Annual Costs & Benefits Using AEO Reference Fuel Prices and RPEs (Billions of 2013$)a'b'c.... 12-76
Table 12.93 Annual Costs & Benefits Using AEO High Fuel Prices and ICMs (Billions of 2013$)a- b-c	12-77
Table 12.94 Annual Costs & Benefits Using AEO Low Fuel Prices and RPEs (Billions of 2013$) a'b'c	12-78
Table 12.95 CY Net Present Value Costs & Benefits in the Central & Sensitivity Cases (Billions of 2013$).... 12-79
Table 12.96 MY2021 Absolute and Incremental Costs per Vehicle in the Central Analysis Using AEO Reference
            Case Fuel Prices and Fleet Projections and Using both ICMs and RPEs (2013$)	12-80
Table 12.97 MY2025 Absolute and Incremental Costs per Vehicle in the Central Analysis Using AEO Reference
            Case Fuel Prices and Fleet Projections and Using both ICMs and RPEs (2013$)	12-81

-------
                                         EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.98 Reference Case Absolute Cost/Vehicle Used as Inputs to the OMEGA Inventory, Cost and Benefit
            Tool (2013$)	12-82
Table 12.99 Control Case Absolute Cost/Vehicle Used as Inputs to the OMEGA Inventory, Cost and Benefit Tool
            (2013$)	12-82
Table 12.100 CO2 and Cost Changes in MY2025 using the 2014 Standards as the Reference Case and the 2025
            Standards as the Control Case for Cars (CO2 in g/mi, dollar values in 2013$)	12-83
Table 12.101 CO2 and Cost Changes in MY2025 using the 2014 Standards as the Reference Case and the 2025
            Standards as the Control Case for Trucks (CO2ing/mi, dollar values in 2013$)	12-83
Table 12.102 CO2 and Cost Changes in MY2025 using the 2014 Standards as the Reference Case and the 2025
            Standards as the Control Case for the Combined Fleet (CO2 in g/mi, dollar values in 2013$)	12-84

-------
                                    EPA's Analysis of the MY2022-2025 GHG Standards
Chapter 12:EPA's Analysis of the MY2022-2025 GHG Standards

   This chapter documents EPA's initial analysis of the impacts of the MY2022 through 2025
GHG emission standards for light duty vehicles. While the Draft TAR is not a policy or decision
document, EPA believes it is important to present our updated assessment of the potential effects
of the changes that have been observed since the 2012 FRM to the light-duty automobile market
on the MY2022 to 2025 greenhouse gas program. In Section 12.1, EPA presents the inputs and
the outputs of our OMEGA analysis. This includes the CCh targets and achieved levels in
meeting the MY2022-2025 standards, along with the associated costs per vehicle and technology
penetrations for a central set of input values and several sensitivity cases.  This section also
includes payback metrics associated with increased vehicle purchase costs countered by
increased fuel savings to illustrate how long it takes for those fuel savings to "pay back" the
higher upfront costs.  In Section 12.2, EPA presents our estimates of emission inventory impacts,
including CCh and other GHGs and criteria pollutants, and impacts on fuel consumption. In
Section 12.3,  EPA presents our draft benefit cost analysis (BCA) for both our model year
lifetime analysis  (BCA considering the full lifetimes of MY2021-2025 vehicles) and our
calendar year analysis (BCA considering the calendar years 2021 through 2050).

   The MY2022  through 2025 GHG standards will significantly reduce harmful GHG emissions.
CCh emissions from automobiles are the product of fuel combustion  and, consequently, reducing
CCh emissions will also achieve a significant reduction in projected fuel consumption. EPA's
projections of these impacts are also shown in this chapter.  Because of anticipated changes to
driving behavior  and fuel production, co-pollutant emissions would also be affected by the
standards. This analysis quantifies the impacts on GHGs, including carbon dioxide (CCh),
methane (CH4), nitrous oxide (N2O) and hydrofluorocarbons (HFC-134a); impacts on "criteria"
air pollutants, including carbon monoxide (CO), fine particulate matter (PIVh.s), sulfur dioxide
(SO2), and the ozone precursors hydrocarbons (VOC) and oxides of nitrogen (NOx); and impacts
on several air toxics, including benzene, 1,3-butadiene, formaldehyde, acetaldehyde, and
acrolein.

   This chapter describes the methods used by EPA in its analysis. Detailed discussion of the
inputs to this analysis are found elsewhere in this Draft TAR (e.g., baseline fleet development is
in Chapter 4, technology costs and effectiveness are in Chapter 5, VMT, rebound effect, and
other economic inputs are in Chapter 10). Chapter 4 also includes a discussion of how the ZEV
program is characterized in our analysis fleet which includes over 400,000 ZEV program
vehicles by MY2025. Note that if the GHG assessment did not consider the  California ZEV
program and the  adoption of that program by several states across the country, then our
assessment of the technology pathways for meeting the 2022-2025 standards would likely show
higher penetrations of other more advanced technologies, such as mild and strong hybrids.

   All OMEGA input and output files for runs presented in this Chapter, and all input and output
files supporting the inventories, benefits and costs presented here are in the EPA docket and are
available on EPA's website at httMI//www3,!e^
                                             12-1

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
12.1  EPA's Estimates of Costs per Vehicle & Technology Penetrations Based
on OMEGA

   As in the analysis of the MYs 2017-2025 rulemaking (the 2012 FRM), our evaluation here
includes identifying potentially available technologies and assessing their effectiveness, cost, and
impacts. The wide number of technologies that are available, and likely to be used in
combination, requires a method to account for their combined cost and effectiveness, as well as
estimates of their availability to be applied to vehicles. The methodologies and tools applied in
this Draft TAR are largely unchanged since the 2012 FRM. The inputs to the process have
changed significantly to reflect all of the research and analysis that EPA has performed as part of
the development of this Draft TAR.

   As done in establishing the GHG standards for MY2012-2016 and 2017-2025, EPA is using a
computerized program called the Optimization Model for reducing Emissions of Greenhouse
gases from Automobiles (OMEGA). Broadly, OMEGA starts with a description of the future
vehicle fleet, including manufacturer, sales, base CO2 emissions, vehicle footprint,  and an
assessment of which GHG emissions-reducing technologies are already employed on the
vehicles. For the purpose of this analysis, EPA uses OMEGA to analyze roughly 200 vehicle
platforms which encompass approximately 1,300 vehicle models to capture the important
differences in vehicle and engine design and utility of future vehicle sales of roughly 15-17
million units annually in the 2021-2025 timeframe.A  The model is then provided with a list of
technologies applicable to various types of vehicles, along with the technologies' cost and
effectiveness and the percentage of vehicle sales that we estimate can be applied to each
technology during the redesign period. The model combines this information with  economic
parameters, such as fuel prices and discount rates, to project how various manufacturers could
apply the available technology in  order to meet increasing levels of GHG emissions control.  The
result is a description of which technologies could be added to each vehicle and vehicle platform,
along with the resulting costs and achieved CO2 levels. The model can also be set to account for
some types of compliance flexibilities.6

   EPA has described OMEGA's specific methodologies and algorithms previously in the model
documentation.2 The model is publicly available on the EPA website,3 and it has been peer
reviewed.4 Emission control technology can be applied individually or in groups, often called
technology "packages." The OMEGA user specifies the cost and effectiveness of each
technology or package for a specific "vehicle type," such as midsize cars with V6 engines or
large trucks with V8 engines. The user can limit the application of a specific technology to a
A The MY2014 baseline fleet used in this analysis actually consists of over 2000 vehicle models, but many of those
  are only minor variations of others (generally a minor footprint~a vehicle's footprint is the product of its track
  width and wheelbase, usually specified in terms of square feet-variation of 0.1 square feet due to, for example,
  different wheel and/or tire applications). For simplicity here, we do not focus on those minor variations although
  our modeling does indeed make use of those variations since a different footprint results in a different target for
  any given vehicle.
B While OMEGA can apply technologies which reduce CCh efficiency related emissions and refrigerant leakage
  emissions associated with air conditioner use, this task is currently handled outside of the OMEGA core model.
  A/C improvements are highly cost-effective, and would always be added to vehicles by the model, thus they are
  simply added into the OMEGA results at the projected penetration levels (see Table 12-6) for each manufacturer.
                                               12-2

-------
                                       EPA's Analysis of the MY2022-2025 GHG Standards
specified percentage of each vehicle's sales (i.e., a "maximum penetration cap").  The
effectiveness, cost, and any application limits of each technology package can also vary over
time.c A list of technologies or packages is provided to OMEGA for each vehicle type,
providing the connection to the specific vehicles being modeled. Appendix C includes more
details on the OMEGA model and approaches used in OMEGA, such as the building of
technology packages, a detailed description of the technology packages, and the mapping of the
fleet into vehicle types and classes, etc.

   For each manufacturer, OMEGA applies technology (subject to any appropriate penetration
caps, as discussed in Appendix C) to vehicles until the sales and VMT-weighted emission
average complies with a given standard or until all the available technologies have been applied.
OMEGA allows the input of a standard which can be in the form of a flat standard applicable to
all vehicles within a vehicle  class (e.g., cars, trucks or both cars and trucks).  Alternatively, the
GHG standard can be in the form of a linear or constrained logistic function, which sets each
vehicle's CO2 target as a function of a vehicle attribute, such as footprint (vehicle track width
times wheelbase). When the linear form of footprint-based standard is used, the "line" can be
converted to a flat standard for footprints either above or below specified levels. This is referred
to as a piece-wise linear standard, and was used  in modeling the footprint-based standards in this
analysis.

   The OMEGA model is designed to estimate the cost of complying with a standard (or target)
in a given future year. While the OMEGA design assumes that a manufacturer's entire fleet of
vehicles can be redesigned within one redesign cycle, it is unlikely that a manufacturer will
redesign the exact same percentage of its vehicle sales in each and every model year.  The base
emissions and emission reductions of the vehicles being redesigned will vary.  Thus, OMEGA
inherently  assumes the averaging and banking of credits—such credits differ from off-cycle
credits—to enable compliance with standards in the intermediate years of a redesign cycle using
the technology projected for the final year of the cycle, assuming that the intermediate standards
require gradual improvement each year.D'E  This assumption has been confirmed by compliance
data from the 2012-2016 MY light duty vehicle  standards, which reflect robust use of averaging
by the manufacturers. We also allow for  transfer of credits between cars and trucks within each
c "Learning," as discussed in Chapter 5.3, is the process whereby the cost of manufacturing a certain item tends to
  decrease with increased production volumes. While OMEGA does not explicitly incorporate "learning" into the
  technology cost estimation procedure, the user can currently simulate learning by inputting lower technology
  costs in each subsequent redesign cycle.
D ABT credits have to do with averaging under- and over-compliance with the standards. Over-compliance
  somewhere allows for under-compliance somewhere else provided "on-average" a fleet complies. If over-
  compliance exceeds under-compliance in any given year, those over-compliance credits can be banked for future
  use within the framework and restrictions of the given program. Trading allows for trading of credits between
  entities, presumably at a cost to the recipient and a financial gain to the provider. Off-cycle credits are real CO2
  reductions that would occur in-use, or the real world, but that are not measured on the 2-cycle test upon which
  fuel economy regulations have long been based.
E EPA considered modeling credit banking as part of this analysis, but decided that the central analysis would not
  analyze the program using this approach for two reasons. First, since the GHG standards continue indefinitely,
  rather than expiring in 2025, EPA wants to represent the cost of bringing vehicles into compliance with the
  standards in MY2025. Second, consistent with the design of the OMEGA model, EPA is not using the OMEGA
  model to project changes on a year-by-year basis, which could be an important element of explicitly modeling
  credit banking.
                                                 12-3

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
manufacturer's fleet allowing the more cost effective of the car/truck fleets to "assist" the other in
compliance.

   EPA has typically used a 5 year redesign cycle in OMEGA.  As such, in the control case for
this analysis, some portion of the fleet is estimated for redesign to the MY2025 standards in
MY2021. This in turn results in the achieved CCh level in the control case in MY2021 being
lower than the target level for that model year.  We explain in section 12.1.1.1.2 the process used
to generate the control case standards in MY2021.

   Once technology has been added so that every manufacturer meets the specified targets (or
exhausts all of the available technologies), the model produces a variety of output files.  The files
include information about the specific technology added to each vehicle and the resulting costs
and emissions levels.  Average costs and emissions per vehicle by manufacturer and industry-
wide are also determined for each vehicle fleet (car and truck).

   Throughout the discussion of EPA's  analysis results is mention of a "reference case" and a
"control case." Since the purpose of this Draft TAR is to assess issues relevant to the MY2022-
2025 standards, the reference case refers to a situation where the future fleet continues to comply
with the MY2021 standards indefinitely. Note that EPA's "baseline fleet" (as described  in
Chapter 4.1) is based on the MY2014 fleet with sales projections going forward through the year
2030. That fleet, by definition, complies with the 2014 standards in MY2014 but not  necessarily
in MY2025.F That "baseline fleet" is contrasted by the "reference case fleet" which adds
additional technology to bring the "baseline fleet" into compliance with the reference case, or
2021 standards.  That "reference case fleet" would then continue meeting the reference case
standards (i.e., the MY 2021 standards) indefinitely. The  "control case" refers to any situation
where the future fleet complies with the MY2022 through MY2025 standards, and then with the
MY2025 standards indefinitely thereafter.  The difference between these two cases is the
incremental effect of the standards (or "delta"). We use "central analysis" control case to
specifically refer to the MY2022-2025 standards established in the 2012 FRM and as analyzed
using what EPA considers to be the central set  of input values (e.g., AEO 2015 reference case
fuel prices are considered to be part of the central analysis).0 The general term "control  case"
can be used for any control case whether it be the central case or a sensitivity case (e.g., AEO
2015 high or low fuel prices are used in sensitivities). As such, while there are several control
cases, one control case is actually considered to be the central control case.  Sensitivity analyses
use different inputs that can vary the analytical outcomes.

   Finally, EPA decided to complete three analysis scenarios built around the AEO 2015
estimates for future fuel prices (see Chapter 10.2).  These  future fuel price scenarios include a
low, reference and high fuel price forecast. EPA is treating the reference fuel price forecast as its
central analysis case.  These fuel price scenarios are also reflected in the development of the
baseline fleet as described in Chapter 4.
F Given the fleet changes projected by the year 2025, that fleet in fact does not comply with the MY2014 standards
  inMY2025.
G Throughout the discussion presented here in Chapter 12, any reference to "AEO" is meant to refer to "AEO2015."
                                               12-4

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
12.1.1  Central Analysis Results

   The central analysis uses the AEO 2015 reference fuel price case and, thus, the AEO 2015
reference fuel price based fleet. The central case also uses both indirect cost multipliers (ICMs)
and retail price equivalents (RPEs) as a means of estimating the indirect costs of technologies.
The central analysis consists of a reference case representing a future fleet complying with the
MY2021 standards indefinitely, and a control case representing a future fleet complying with the
MY2022 to 2025 standards in those respective model years, and then with the MY2025 standard
indefinitely.

12.1.1.1       CO2 Targets and Achieved Values

   The central analysis uses two approaches for reflecting indirect costs, both ICMs and RPEs as
discussed in Chapter 5. Because there are differences in the technology costs for the ICM and
RPE cases, which result in slightly different technology penetrations and car/truck credit
transfers, these differences lead to differences in the CCh Achieved levels in the ICM compared
to RPE cases, as shown below. Technology costs are presented in Section 12.1.1.2, and
technology penetration rates are presented in Section 12.1.1.3.

   Note that the GHG standards (i.e.,  the standard curves) apply to individual vehicles.
Depending on the footprint and model year of that individual vehicle, its target value can be
determined by selecting the appropriate standard curve. A fleet of vehicles—whether a car/truck
fleet, a given manufacturer's fleet, or the entire fleet—complying with its individual targets
(determined by the standard curves) while giving consideration to the sales,  or sales weighting,
of each would result in a target value  for that given fleet.  We present here the fleetwide target
values for each manufacturer's car fleet, the entire car fleet, each manufacturer's truck fleet, and
the entire truck fleet. These target values are not the standards but rather the sales-weighted CCh
emissions of each particular fleet assuming that individual vehicles comply with their respective
footprint targets.

12.1.1.1.1     Reference  Case

   The reference case represents the fleet meeting the MY2021 standards  in MYs 2021 and
thereafter. We present the reference case CCh targets and projected achieved levels in MY2021
in Table 12.1.  We present the reference case CCh targets and projected achieved levels in
MY2025 in Table 12.2. While both tables represent the same set of reference case standards, the
target and achieved CCh levels reflect differences,  which are attributed to fleet changes between
MYs 2021 and 2025.
                                              12-5

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
    Table 12.1  Reference Case Targets and Achieved CCh in MY2021 in the Central Analysis (g/mi CCh)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Car
Target
178.6
182.4
179.8
178.7
172.5
177.0
189.7
175.5
180.3
164.7
173.6
170.0
205.7
174.4
174.0
182.0
177.0
Truck
Target
237.5
246.0
280.3
277.2
231.0
227.2
235.4
223.1
237.5
208.4
241.9
210.6
0.0
247.3
231.1
227.7
251.5
Fleet
Target
194.4
227.1
239.3
230.1
201.0
183.3
226.8
189.9
204.0
181.5
202.4
201.6
205.7
209.1
196.6
206.4
213.8
Car
Achieved,
ICM
182.7
194.1
198.3
197.8
176.5
178.9
169.7
181.3
179.6
180.1
179.1
206.5
0.0
173.1
170.4
193.4
182.2
Truck
Achieved,
ICM
227.7
241.5
268.7
262.2
226.5
215.4
239.4
211.6
237.8
185.2
235.1
201.8
0.0
248.3
235.7
219.2
244.5
Fleet
Achieved,
ICM
194.8
227.5
240.0
231.4
200.8
183.5
226.2
190.5
203.7
182.0
202.7
202.8
0.0
208.9
196.2
207.1
213.0
Car
Achieved,
RPE
183.3
196.4
203.2
197.2
176.7
179.8
172.8
177.1
177.9
178.8
179.8
215.2
0.0
173.0
169.1
191.6
183.0
Truck
Achieved,
RPE
226.5
240.9
266.5
262.7
226.5
210.7
238.8
219.1
240.3
189.1
234.1
198.9
0.0
247.7
237.1
220.5
243.8
Fleet
Achieved,
RPE
194.9
227.7
240.7
231.4
200.9
183.7
226.3
189.8
203.7
182.7
202.7
202.5
0.0
208.5
196.0
207.0
213.1
       Note:  Fleet values are sales weighted but not VMT weighted.
    Table 12.2 Reference Case Targets and Achieved CCh in MY2025 in the Central Analysis (g/mi CCh)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Car
Target
177.5
182.3
179.6
178.8
172.8
177.1
189.7
175.2
180.0
164.8
173.3
170.0
205.7
174.5
174.6
182.0
176.9
Truck
Target
237.0
247.2
280.0
277.3
232.9
227.9
235.0
223.4
237.0
208.4
243.0
210.5
0.0
246.3
230.4
227.7
251.3
Fleet
Target
191.7
227.9
237.9
227.9
200.8
183.1
225.5
190.0
201.7
180.4
200.9
201.4
205.7
207.0
195.7
205.8
212.4
Car
Achieved,
ICM
180.3
189.6
196.4
197.1
175.9
179.0
170.8
177.6
180.4
179.2
177.2
210.6
0.0
174.6
170.1
188.3
181.0
Truck
Achieved,
ICM
229.2
244.6
269.5
261.6
229.9
214.5
239.0
218.1
236.4
186.4
237.0
200.3
0.0
246.2
236.3
222.7
244.8
Fleet
Achieved,
ICM
192.0
228.2
238.9
229.3
201.1
183.2
224.7
190.0
201.7
181.8
200.9
202.6
0.0
207.0
195.1
206.2
211.4
Car
Achieved,
RPE
180.1
192.0
198.5
197.5
178.6
178.3
167.4
177.9
178.3
178.4
177.8
213.0
0.0
170.2
168.9
187.5
180.9
Truck
Achieved,
RPE
229.8
243.7
268.4
260.8
227.2
213.5
239.0
218.1
239.3
187.6
237.1
199.5
0.0
250.5
237.6
222.6
244.7
Fleet
Achieved,
RPE
191.9
228.3
239.1
229.1
201.3
182.5
224.1
190.2
201.5
181.7
201.2
202.6
0.0
206.6
194.9
205.8
211.3
       Note:  Fleet values are sales weighted but not VMT weighted.
12.1.1.1.2
Control Case
                                                 12-6

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
   The central analysis control case represents the fleet meeting the MY2022 through MY2025
standards in their respective model years, and the fleet meeting the MY2025 standards
indefinitely thereafter. We continue to estimate a 5 year redesign cycle.  This cycle is consistent
with our understanding of industry practice (although there are indications that cycles are
becoming shorter due to competitive pressures, especially on cars).  This is how EPA's modeling
has always been done. We know that industry plans ahead for compliance with future standards
and carefully considers their redesign cycles when developing their compliance plans. To
accommodate a 5 year redesign cycle, we have estimated that 20 percent of the MY2021 fleet
will be redesigned to meet the MY2025 standards, and so on through MY2024. As noted above,
this effectively results in the MY2021 through MY2024 control case targets and achieved CCh
levels being below (i.e., better than) the reference case target (i.e., the MY2021  target) since 20
percent of each fleet will be redesigned to meet the MY2025 standards.  The actual standards and
the control case targets used in this analysis are shown graphically in Figure 12.1 for cars and
Figure 12.2 for trucks.
                     Actual Standard and Control Case Target Curves for Cars
       240
       220
       200
       180
     O
     u
       160
       140
       120
       100
•2021 Actual Std

-2021 Control Target

•2022 Actual Std

-2022 Control Target

•2023 Actual Std

-2023 Control Target

 2024 Actual Std

 2024 Control Target

-2025 Control & Actual
          35
                          40
                                          45               50
                                            Footprint (square feet)


 Figure 12.1 Actual Standard Curves and the Control Case Target Curves Used for Cars in this Draft TAR
                           Analysis to Reflect a 5-Year Redesign Cycle
                                                12-7

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
                    Actual Standard and Control Case Target Curves for Trucks
       380
       340
       300

     O
     u

     I
       220
       180
       140
       100
2021 Actual Std

2021 Control Target

2022 Actual Std

2022 Control Target

2023 Actual Std

2023 Control Target

2024 Actual Std

2024 Control Target

2025 Control & Actual
          35
                   40
                            45
                                     50        55        60
                                            Footprint (square feet)

                                                                         70
                                                                                  75
                                                                                           80
Figure 12.2 Actual Standard Curves and the Control Case Target Curves Used for Trucks in this Draft TAR
                           Analysis to Reflect a 5-Year Redesign Cycle
   Shown in these figures are the "actual," or promulgated greenhouse gas standard curves for
the years 2021 through 2024 (dashed lines) and the control case target curves used in this
analysis (solid lines). The control case target curves reflect greater stringency (lower CCh) to
reflect the 5 year redesign cycle discussed above. In effect, the target curves represent over-
compliance with the actual standard curves in each year leading up to 2025.  Just one curve is
shown for 2025  since the actual standard and control case target curves are the same by then.

   Importantly, the control case "standards" being used here are not new standard curves.
Instead, they are an OMEGA modeling artifact used to simulate over-compliance with the actual
standards.  This  over-compliance is being projected by EPA only to accommodate  the 5-year
redesign cycle stance, reflecting industry practice, and which we have used in the analyses for
both of the LDV GHG rules.

   Nonetheless,  these standard curves, whether actual or the control case curves are being used,
are used for determining the OMEGA target values for individual vehicles depending  on the MY
and their unique footprints.  By determining those target values for each vehicle in the fleet and
sales-weighting  those, a fleet target can be determined for each manufacturer and for the entire
fleet. Running that fleet through OMEGA and determining the most cost-effective path toward
                                               12-8

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
compliance (while also considering any appropriate technology penetration caps (see Appendix
C) and other limitations on the application of technology), and considering credits and transfers
as allowed under the program, we can estimate the achieved CCh level for each manufacturer and
for the entire fleet.

  We present the CCh targets and projected achieved levels in MY2021 in Table 12.3 and in
MY2025 in Table 12.4. Note that the targets and achieved values shown in Table 12.3 include
over-compliance with the actual standards, as explained above. For the 2012 FRM, EPA
predicted an overall fleet average CCh performance of 163 g/mi.  As shown in Table 12.4, the
overall fleet performance is predicted to achieve 174.1 g/mi. This increase in CCh emissions can
be largely attributed to the increase in sales of trucks.
    Table 12.3 Control Case Targets and Achieved CCh in MY2021 in the Central Analysis (g/mi CCh)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Car
Target
172.6
176.3
173.8
172.7
166.7
171.1
183.3
169.6
174.2
159.2
167.7
164.3
198.9
168.6
168.1
175.9
171.0
Truck
Target
228.9
237.2
270.4
267.4
222.7
218.9
226.9
215.0
229.0
200.7
233.2
202.9
0.0
238.4
222.8
219.5
242.5
Fleet
Target
187.8
219.1
231.0
222.2
193.9
177.1
218.7
183.3
196.9
175.1
195.3
194.3
198.9
201.8
189.7
199.1
206.4
Car
Achieved,
ICM
173.4
192.1
191.1
187.9
171.0
173.7
157.3
171.0
170.3
175.6
174.9
202.8
-18.8
169.1
161.8
179.2
176.0
Truck
Achieved,
ICM
225.7
230.3
260.2
255.4
217.9
201.3
231.3
211.6
233.2
176.9
224.7
193.5
0.0
237.8
229.4
217.0
235.7
Fleet
Achieved,
ICM
187.5
219.0
232.1
223.1
193.8
177.2
217.3
183.2
196.3
176.1
195.9
195.6
-18.8
201.8
188.6
199.4
205.5
Car
Achieved,
RPE
173.7
192.7
191.2
190.5
171.2
174.5
155.2
171.1
170.7
176.3
174.7
209.0
0.0
170.2
160.8
181.8
176.8
Truck
Achieved,
RPE
226.2
231.1
260.2
253.5
217.1
197.9
231.8
211.9
233.3
177.3
224.9
191.0
0.0
235.9
231.3
214.2
235.0
Fleet
Achieved,
RPE
187.8
219.7
232.1
223.4
193.5
177.5
217.3
183.5
196.6
176.7
195.9
195.0
0.0
201.5
188.7
199.1
205.6
Note: Fleet values are sales weighted but not VMT weighted; targets include 20% over-compliance to the MY2025
standards.
                                              12-9

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
    Table 12.4 Control Case Targets and Achieved CCh in MY2025 in the Central Analysis (g/mi CCh)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Car
Target
147.7
151.8
149.5
148.8
143.7
147.3
158.1
145.8
149.8
137.0
144.1
141.3
171.6
145.2
145.2
151.5
147.2
Truck
Target
193.8
202.3
229.7
227.4
190.4
186.2
192.1
182.4
193.8
169.9
198.8
171.7
0.0
201.5
188.3
186.1
205.7
Fleet
Target
158.7
187.3
196.1
188.0
165.5
152.0
185.0
157.0
166.6
148.8
165.8
164.9
171.6
170.7
161.5
169.5
175.1
Car
Achieved,
ICM
146.0
159.9
165.7
168.6
142.2
149.5
102.4
148.6
141.8
148.2
145.6
173.6
-18.8
143.4
133.7
154.3
149.8
Truck
Achieved,
ICM
197.5
199.4
219.0
209.3
191.9
171.0
204.6
174.9
203.2
150.5
196.9
163.7
0.0
203.5
204.5
183.3
200.7
Fleet
Achieved,
ICM
158.3
187.6
196.7
188.9
165.4
152.0
183.2
156.6
165.2
149.0
165.9
165.9
-18.8
170.5
160.5
169.4
174.1
Car
Achieved,
RPE
146.3
159.4
168.5
168.9
147.5
148.3
104.4
149.0
142.5
149.5
150.9
179.8
0.0
144.8
131.3
147.2
151.2
Truck
Achieved,
RPE
197.6
199.5
218.0
210.1
186.3
173.1
204.2
176.2
203.9
150.5
188.9
162.2
0.0
201.3
205.9
189.5
199.4
Fleet
Achieved,
RPE
158.5
187.6
197.2
189.5
165.6
151.3
183.3
157.3
165.9
149.8
165.9
166.1
0.0
170.4
159.5
169.2
174.2
Note: Fleet values are sales weighted but not VMT weighted; targets include 20% over-compliance to the MY2025
standards.

12.1.1.1.3     Off-Cycle. Pickup Incentive andA/C Credits in OMEGA

   In achieving the targets as shown in the tables above, manufacturers have available to them
off-cycle credits for technologies, such as active aero and stop-start, that achieve real world CCh
reductions although their impact is not adequately captured on the 2-cycle test (see II.F.2 of the
2012 FRM, 77 FR 62726).  There are also incentive credits available for certain advanced
technologies, such as strong hybrids on pickup trucks (see II.F.3 of the 2012 FRM, 77 FR
62738). Lastly, there are A/C credits which EPA assumes that all manufacturers will use in
meeting the targets shown above (see II.F. 1 of the 2012 FRM, 77 FR 62721).  While
manufacturers have available to them broader options for utilizing off-cycle technologies,
including a fuller list of pre-approved off-cycle credits  (see 40 CFR 86.1869-12), EPA is making
a very conservative assumption for purposes of this Draft TAR analysis and is only making
available within the OMEGA model two of those off-cycle technologies, active aero and stop-
start, as shown in the table below. EPA will consider expanding the off-cycle technology
included in our modeling assessment for future steps in the MTE process. The credits shown
below are available within the model in both the reference and control cases.
                                              12-10

-------
                                       EPA's Analysis of the MY2022-2025 GHG Standards
    Table 12.5 Off-cycle & Pickup Incentive Credits Available for Achieving the CCh Targets (g/mi CCh)
MY
2021
2022
2023
2024
2025
2021
2022
2023
2024
2025
2021
2022
2023
2024
2025
Vehicle
Car
Car
Car
Car
Car
Truck, non-pickup
Truck, non-pickup
Truck, non-pickup
Truck, non-pickup
Truck, non-pickup
Pickup
Pickup
Pickup
Pickup
Pickup
Active Aero
0.6
0.6
0.6
0.6
0.6
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
Stop-start
2.5
2.5
2.5
2.5
2.5
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
4.4
Mild HEV Incentive
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
10.0
0.0
0.0
0.0
0.0
Strong HEV Incentive
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
20.0
20.0
20.0
20.0
20.0
   The magnitude of the credits used within OMEGA, and reflected in the achieved CCh values
presented in the "Target and Achieved CCh" tables above are shown in the table below. The A/C
credits used within OMEGA and reflected in both the targets and the achieved CO2 values
presented in the "Target and Achieved CO2" tables above are also shown in the tables below.
    Table 12.6  Off-cycle, Pickup Incentive and A/C Credits Used to Achieve the COi Targets (g/mi COi)

Case
AEO
Ref,
ICM
AEO
High,
ICM
AEO
Low,
ICM
AEO
Ref,
RPE

Standard
Reference
Control
Reference
Control
Reference
Control
Reference
Control
Reference
Control
Reference
Control
Reference
Control
Reference
Control

MY
2021
2021
2025
2025
2021
2021
2025
2025
2021
2021
2025
2025
2021
2021
2025
2025
Off-cycle Credits
Car
0.499
0.612
0.458
1.089
0.533
0.630
0.376
1.190
0.470
0.597
0.330
1.105
0.515
0.629
0.480
1.233
Truck
1.367
2.256
1.083
3.481
1.664
2.624
1.199
3.482
1.364
2.065
1.057
3.356
1.350
2.329
1.225
3.858
Combined
0.960
1.485
0.779
2.317
1.030
1.506
0.721
2.151
0.979
1.434
0.733
2.354
0.959
1.531
0.863
2.579
Incentive
Credits
Pickups
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
A/C Credits
Car
18.8
18.8
18.8
18.8
18.8
18.8
18.8
18.8
18.8
18.8
18.8
18.8
18.8
18.8
18.8
18.8
Truck
24.4
24.4
24.4
24.4
24.4
24.4
24.4
24.4
24.4
24.4
24.4
24.4
24.4
24.4
24.4
24.4
Combined
21.6
21.6
21.5
21.5
21.6
21.6
21.5
21.5
21.6
21.6
21.5
21.5
21.6
21.6
21.5
21.5
Note: The car A/C credit is composed of an indirect (or efficiency) credit of 5.0 g/mi CCh and a direct (or leakage)
credit of 13.8 g/mi CO2eq; the truck credit is composed of an indirect credit of 7.2 g/mi CO2 and a direct credit
(leakage credit) of 17.2 g/mi CO2eq.
                                                12-11

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
12.1.1.1.4     Projected 2-Cvcle CO2

   The compliance targets presented above include use of A/C and the specified off-cycle
credits.  The actual tailpipe CCh as tested over the 2-cycle test procedure are higher than the
actual targets since the A/C portion of the standards are not included as part of the test results.
The tables below show the projected 2-cycle tailpipe CCh values for cars, trucks and the fleet
using AEO 2015 reference fuel price case and ICMs.
  Table 12.7 EPA Projections for Car Tailpipe Emissions Compliance with CCh Standards Using ICMs and
                          the AEO Reference Fuel Price Case (CCh g/mi)
MY
2021
2022
2023
2024
2025
Compliance
Target, (CO2
Standard)
171
165
159
153
147
Compliance
Target as MPG
51.9
53.9
56.0
58.2
60.3
Incentive
Credits
0
0
0
0
0
Off-cycle
Credits
0.612
0.731
0.851
0.970
1.089
Leakage
A/C
Credits
13.8
13.8
13.8
13.8
13.8
Efficiency
A/C Credits
5.0
5.0
5.0
5.0
5.0
Projected
2-cycle
Tailpipe
CO2
190
184
178
173
167
 Table 12.8 EPA Projections for Truck Tailpipe Emissions Compliance with CCh Standards Using ICMs and
                          the AEO Reference Fuel Price Case (CO2 g/mi)
MY
2021
2022
2023
2024
2025
Compliance
Target, (CO2
Standard)
242
232
223
214
206
Compliance
Target as MPG
36.7
38.3
39.9
41.6
43.2
Incentive
Credits
0
0
0
0
0
Off-cycle
Credits
2.256
2.562
2.869
3.175
3.481
Leakage
A/C
Credits
17.2
17.2
17.2
17.2
17.2
Efficiency
A/C Credits
7.2
7.2
7.2
7.2
7.2
Projected
2-cycle
Tailpipe
CO2
269
259
250
241
233
 Table 12.9 EPA Projections for Combined Fleet Tailpipe Emissions Compliance with COi Standards Using
                     ICMs and the AEO Reference Fuel Price Case (CO2 g/mi)
MY
2021
2022
2023
2024
2025
Compliance
Target, (CO2
Standard)
206
198
190
182
175
Compliance
Target as MPG
43.1
44.9
46.8
48.8
50.8
Incentive
Credits
0
0
0
0
0
Off-cycle
Credits
1.485
1.693
1.901
2.109
2.317
Leakage
A/C Credits
15.5
15.5
15.5
15.4
15.4
Efficiency
A/C Credits
6.1
6.1
6.1
6.1
6.0
Projected
2-cycle
Tailpipe CO2
229
221
213
206
199
   The tables below show the projected 2-cycle tailpipe CCh values for cars, trucks and the fleet
using AEO 2015 high fuel price case and ICMs.
                                              12-12

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.10  EPA Projections for Car Tailpipe Emissions Compliance with CCh Standards Using ICMs and
                            the AEO High Fuel Price Case (CCh g/mi)
MY
2021
2022
2023
2024
2025
Compliance
Target, (CO2
Standard)
171
165
158
153
147
Compliance
Target as MPG
52.0
54.0
56.1
58.2
60.4
Incentive
Credits
0
0
0
0
0
Off-cycle
Credits
0.630
0.770
0.910
1.050
1.190
Leakage
A/C
Credits
13.8
13.8
13.8
13.8
13.8
Efficiency
A/C Credits
5.0
5.0
5.0
5.0
5.0
Projected
2-cycle
Tailpipe
CO2
190
184
178
173
167
Table 12.11  EPA Projections for Truck Tailpipe Emissions Compliance with CCh Standards Using ICMs and
                            the AEO High Fuel Price Case (CCh g/mi)
MY
2021
2022
2023
2024
2025
Compliance
Target, (CO2
Standard)
240
231
222
213
204
Compliance
Target as MPG
36.7
38.3
39.9
41.6
43.2
Incentive
Credits
0
0
0
0
0
Off-cycle
Credits
2.624
2.838
3.053
3.267
3.482
Leakage
A/C
Credits
17.2
17.2
17.2
17.2
17.2
Efficiency
A/C Credits
7.2
7.2
7.2
7.2
7.2
Projected
2-cycle
Tailpipe
CO2
268
258
249
240
232
Table 12.12 EPA Projections for Combined Fleet Tailpipe Emissions Compliance with CCh Standards Using
                        ICMs and the AEO High Fuel Price Case (CO2 g/mi)
MY
2021
2022
2023
2024
2025
Compliance
Target, (CO2
Standard)
199
191
183
176
169
Compliance
Target as MPG
43.1
44.9
46.8
48.8
50.8
Incentive
Credits
0
0
0
0
0
Off-cycle
Credits
1.506
1.667
1.829
1.990
2.151
Leakage
A/C Credits
15.2
15.2
15.1
15.1
15.1
Efficiency
A/C Credits
5.9
5.9
5.9
5.9
5.8
Projected
2-cycle
Tailpipe CO2
222
214
206
199
192
   The tables below show the projected 2-cycle tailpipe CCh values for cars, trucks and the fleet
using AEO 2015 low fuel price case and ICMs.
                                               12-13

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.13 EPA Projections for Car Tailpipe Emissions Compliance with CCh Standards Using ICMs and
                            the AEO Low Fuel Price Case (CCh g/mi)
MY
2021
2022
2023
2024
2025
Compliance
Target, (CO2
Standard)
171
165
159
153
147
Compliance
Target as MPG
51.9
53.9
56.0
58.1
60.3
Incentive
Credits
0
0
0
0
0
Off-cycle
Credits
0.597
0.724
0.851
0.978
1.105
Leakage
A/C
Credits
13.8
13.8
13.8
13.8
13.8
Efficiency
A/C Credits
5.0
5.0
5.0
5.0
5.0
Projected
2-cycle
Tailpipe
CO2
191
184
178
173
167
Table 12.14 EPA Projections for Truck Tailpipe Emissions Compliance with CCh Standards Using ICMs and
                            the AEO Low Fuel Price Case (CCh g/mi)
MY
2021
2022
2023
2024
2025
Compliance
Target, (CO2
Standard)
242
233
223
214
206
Compliance
Target as MPG
36.7
38.2
39.8
41.5
43.2
Incentive
Credits
0
0
0
0
0
Off-cycle
Credits
2.065
2.388
2.711
3.034
3.356
Leakage
A/C
Credits
17.2
17.2
17.2
17.2
17.2
Efficiency
A/C Credits
7.2
7.2
7.2
7.2
7.2
Projected
2-cycle
Tailpipe
CO2
269
259
250
242
234
Table 12.15 EPA Projections for Combined Fleet Tailpipe Emissions Compliance with CCh Standards Using
                       ICMs and the AEO Low Fuel Price Case (COi g/mi)
MY
2021
2022
2023
2024
2025
Compliance
Target, (CO2
Standard)
209
201
193
185
178
Compliance
Target as MPG
42.5
44.2
46.2
48.1
50.0
Incentive
Credits
0
0
0
0
0
Off-cycle
Credits
1.434
1.664
1.894
2.124
2.354
Leakage
A/C Credits
15.6
15.6
15.6
15.6
15.6
Efficiency
A/C Credits
6.2
6.2
6.2
6.1
6.1
Projected
2-cycle
Tailpipe CO2
232
224
216
209
202
12.1.1.2       Cost per Vehicle

12.1.1.2.1     Reference & Control Case

   EPA presents the incremental costs of meeting the control case standards in MY2021 in Table
12.16 and in MY2025 in Table 12.17.  We present the estimated progression of these
incremental, control case costs relative to the reference case costs for cars in Table 12.18, and for
trucks in Table 12.19.

   As shown in Table 12.17, the average per vehicle costs to meet the MY2025 standards in
MY2025 (compared to meeting  the MY2021 standards in MY2025) is between $894 and $1,017.
These costs are less than those estimated in the 2012 FRM, as discussed below in section
12.1.1.2.2.
                                              12-14

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
   EPA presents absolute costs for MY2025 vehicles meeting the 2021 standards (i.e., the
reference case) and for MY2025 vehicles meeting the 2025 standards (i.e., the central analysis
control case), for cars, trucks, and the fleet in section 12.4. The costs presented there are the costs
used as inputs to the OMEGA Inventory, Cost and Benefit Tool discussed in more detail in
Chapter 12.2
   Table 12.16 MY2021 Control Case Cost/Vehicle Incremental to the Reference Case Cost/Vehicle in the
  Central Analysis Using AEO Reference Case Fuel Prices and Fleet Projections and Using both ICMs and
                                        RPEs (2013$)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Car
$402-$423
$64-$126
$105-$ 188
$185-$215
$68-$75
$144-$ 155
$941-$1264
$80-$126
$351-$453
$60-$98
$90-$92
$33-$66
$0-$0
$45-$50
$432-$433
$445-$624
$154-$162
Truck
$21-$123
$382-$394
$128-$159
$137-$163
$150-$169
$491-$526
$578-$677
$0-$115
$332-$501
$177-$273
$191-$196
$84-$85
$0-$0
$189-$238
$437-$468
$194-$455
$225-$234
Combined
$299-$342
$296-$306
$137-$153
$173-$174
$108-$ 120
$187-$202
$708-$727
$88-$90
$403-$413
$128-$142
$134-$135
$73-$81
$0-$0
$113-$140
$434-$447
$395-$450
$189-$197
                                               12-15

-------
                                    EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.17 MY2025 Control Case Cost/Vehicle Incremental to the Reference Case Cost/Vehicle in the
Central Analysis Using AEO Reference Case Fuel Prices and Fleet Projections and Using both ICMs and
                                      RPEs (2013$)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Car
$1080-$1181
$879-$ 1063
$535-$606
$593-$710
$544-$569
$731-$901
$3363-$3366
$469-$539
$1383-$ 1401
$673-$724
$635-$680
$451-$461
$0-$0
$548-$555
$1333-$1544
$1247-$1575
$707-$789
Truck
$1070-$1188
$1400-$1501
$1147-$1273
$1520-$1633
$493-$771
$1279-$1284
$1391-$1592
$652-$748
$1253-$1528
$719-$866
$816-$1218
$531-$647
$0-$0
$871-$ 1140
$1202-$1316
$1257-$1575
$1099-$1267
Combined
$1078-$1183
$1245-$1371
$890-$993
$1055-$1170
$520-$663
$797-$946
$1804-$ 1963
$525-$603
$1334-$ 1449
$689-$775
$734-$866
$515-$603
$0-$0
$694-$820
$1284-$ 1458
$1410-$1417
$894-$1017
                                             12-16

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
   Table 12.18 MY2021-2025 Control Case Cost/Vehicle Incremental to the Reference Case Cost/Vehicle
   Year-by-Year Costs per Car in the Central Analysis Using AEO Reference Case Fuel Prices and Fleet
                                      Projections (2013$)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
MY2021
$402-$423
$64-$126
$105-$ 188
$185-$215
$68-$75
$144-$ 155
$941-$1264
$80-$126
$351-$453
$60-$98
$90-$92
$33-$66
$0-$0
$45-$50
$432-$433
$445-$624
$154-$162
MY2022
$588-$597
$268-$360
$213-$293
$310-$316
$187-$198
$291-$342
$1546-$1790
$195-$212
$614-$685
$226-$242
$228-$237
$140-$ 162
$0-$0
$170-$177
$658-$710
$727-$780
$292-$319
MY2023
$752-$791
$472-$594
$320-$397
$404-$448
$306-$322
$437-$528
$2152-$2315
$298-$310
$876-$918
$386-$392
$364-$385
$247-$258
$0-$0
$296-$303
$883-$988
$935-$1010
$430-$475
MY2024
$916-$986
$675-$829
$427-$501
$499-$579
$425-$445
$584-$715
$2757-$2840
$384-$424
$1138-$1151
$529-$558
$500-$533
$354-$355
$0-$0
$422-$429
$1108-$1266
$1091-$1292
$569-$632
MY2025
$1080-$1181
$879-$1063
$535-$606
$593-$710
$544-$569
$731-$901
$3363-$3366
$469-$539
$1383-$1401
$673-$724
$635-$680
$451-$461
$0-$0
$548-$555
$1333-$1544
$1247-$1575
$707-$789
   Table 12.19 MY2021-2025 Control Case Cost/Vehicle Incremental to the Reference Case Cost/Vehicle
  Year-by-Year Costs per Truck in the Central Analysis Using AEO Reference Case Fuel Prices and Fleet
                                      Projections (2013$)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
MY2021
$21-$123
$382-$394
$128-$159
$137-$163
$150-$169
$491-$526
$578-$677
$0-$115
$332-$501
$177-$273
$191-$196
$84-$85
$0-$0
$189-$238
$437-$468
$194-$455
$225-$234
MY2022
$313-$360
$645-$662
$406-$414
$483-$530
$235-$319
$689-$714
$831-$855
$163-$273
$563-$758
$312-$422
$351-$448
$196-$225
$0-$0
$359-$464
$651-$656
$539-$656
$444-$492
MY2023
$597-$605
$897-$942
$653-$700
$829-$898
$321-$470
$887-$903
$1034-$ 1085
$326-$432
$793-$1014
$448-$570
$506-$705
$308-$366
$0-$0
$530-$689
$835-$876
$856-$884
$662-$750
MY2024
$834-$897
$1148-$1221
$900-$987
$1174-$1266
$407-$621
$1085-$ 1091
$1213-$1339
$489-$590
$1023-$1271
$583-$718
$661-$961
$419-$506
$0-$0
$700-$915
$1018-$1096
$1057-$1230
$881-$1008
MY2025
$1070-$1188
$1400-$1501
$1147-$1273
$1520-$1633
$493-$771
$1279-$1284
$1391-$1592
$652-$748
$1253-$1528
$719-$866
$816-$1218
$531-$647
$0-$0
$871-$ 1140
$1202-$1316
$1257-$1575
$1099-$1267
   Note that the costs shown in Table 12.18 and Table 12.19 are based on interpolations between
the incremental costs of the control case standards in MY2021 and the control case standards in
MY2025 (both based on actual OMEGA output), using the control case CCh targets for each
fleet (car and truck) for each individual OEM.
                                               12-17

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
12.1.1.3       Technology Penetration

12.1.1.3.1     Reference Case

   EPA presents technology penetration rates in the MY2025 reference case (that is, the case
where MY2021 standards remain in place in MY2025), in absolute terms, for cars and trucks and
the fleet,  using both ICMs  and RPEs, in the tables below. First we present a table with the
technology codes and their definitions as used in the following technology penetration tables.
For detailed descriptions of each technology, refer to Chapter 5. In the interests of space, we do
not present the technology penetrations for all technologies considered in this analysis. We
present here only those technologies that we believe to be of most interest to the reader.
Therefore, technologies like the accommodation of low friction lubes and lower rolling
resistance tires are not presented here largely because those technologies are very cost effective
and, therefore, have very high penetrations and, while important in achieving the standards, are
not the primary drivers behind the feasibility of the standards.  Note that the OMEGA output
files include technology penetrations for  all technologies considered; those output files are
contained in the docket and on EPA's website at https://www3.epa.gov/otaq/climate/models.htm.

   All technology penetration rate tables  use the AEO 2015 reference fuel price case.  One note
of interest regarding the weight reduction technologies shown in the following tables:  The
"WRtech" is the weight reduction technology applied to the vehicle.  This is the technology used
to determine the costs associated with weight reduction.  If 10 percent WRtech is applied,  then
the costs  associated with that are those costs for a 10  percent weight reduction. The "WRnet" is
the net weight reduction, or the WRtech less the added weight of any added batteries for
electrification (i.e., HEVs,  EVs, and PHEVs). The WRnet value determines effectiveness values
and is used in the safety analysis (Chapter 8). As shown in the technology penetration tables that
follow, there is not much difference between "WRtech" and "WRnet" because our  modeling
does projects very little increased electrification of the fleet to meet either the reference or
control case standards.  Nonetheless, the  distinction between these two technologies is important
and is tracked for that reason.

   Note that the electrified vehicle technology penetrations—EV and PHEV, in particular--
include the penetration of ZEV program vehicles as discussed in detail in Chapter 4 of this Draft
TAR. Importantly, the ZEV program vehicles were "built" into the fleet with the projection that
they would apply 20 percent mass reduction technology (WRtech) and 20 percent net mass
reduction (WRnet).  The result being that the mass reduction technology penetrations include a
20 percent mass reduction  on roughly 2.5 percent of the fleet due to the way we have assessed
the ZEV  program vehicles.
                                              12-18

-------
                                 EPA's Analysis of the MY2022-2025 GHG Standards
         Table 12.20  Technology Code Definitions used in Technology Penetration Tables
Code
WRTech
WRNet
TDS18
TDS24
TRX11
TRX12
TRX21
TRX22
Deac
VVLT
VVT
CEGR
Strong HEV
EV
PHEV
ATK1
ATK2
Miller
Stop-Start
Mild Hybrid
DSL
Definition
Weight reduction technology applied
Weight reduction net (includes added weight from batteries on electrified vehicles)
Turbocharged and downsized engine - 18 bar BMEP
Turbocharged and downsized engine- 24 bar BMEP
Transmission level 1 (i.e., 6 speed auto, 6 speed DCT or CVT today)
Transmission level 1 (i.e., TRX11 with efficiency improvements)
Transmission level 2 (i.e., TRX11 with a wider gear ratio spread)
Transmission level 2 (i.e., TRX21 with efficiency improvements)
Cylinder deactivation
Variable valve lift
Variable valve timing
Cooled Exhaust Gas Recirculation
Strong hybrid
Full battery electric vehicle
Plug-in hybrid electric vehicle)
Atkinson cycle engine used in Full Hybrid & REEV
Atkinson cycle engine used in naturally aspirated, non-hybrid engines
Miller cycle, or turbocharged ATK2
Stop-start, but without also being hybridized
Mild hybrid 48 Volt
Diesel
The tables that follow for reference case technology penetrations are:

    •  Table 12.21  Absolute Technology Penetrations for Cars in the MY2025 Reference
       Case Using ICMs
    •  Table 12.22  Absolute Technology Penetrations for Cars in the MY2025 Reference
       Case Using RPEs
    •  Table 12.23  Absolute Technology Penetrations for Trucks in the MY2025 Reference
       Case Using ICMs
    •  Table 12.24  Absolute Technology Penetrations for Trucks in the MY2025 Reference
       Case Using RPEs
    •  Table 12.25  Absolute Technology Penetrations for the Fleet in the MY2025
       Reference Case Using ICMs
    •  Table 12.26  Absolute Technology Penetrations for the Fleet in the MY2025
       Reference Case Using RPEs
    •  Table 12.27  Summary of Absolute Technology Penetrations in the MY2025
       Reference Case
                                          12-19

-------
                                                    EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.21  Absolute Technology Penetrations for Cars in the MY2025 Reference Case Using ICMs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
7.3%
7.3%
4.5%
5.2%
3.1%
3.8%
16.0%
6.8%
10.5%
4.2%
6.0%
2.9%
0.0%
3.4%
9.1%
9.5%
5.2%
WR
Net
5.2%
7.3%
4.2%
4.5%
3.1%
3.7%
15.0%
6.8%
10.1%
4.2%
6.0%
2.9%
0.0%
3.0%
8.3%
9.5%
4.8%
TDS
18
47.6%
35.9%
39.0%
48.7%
0.0%
13.6%
51.4%
0.0%
62.1%
9.6%
19.0%
9.0%
0.0%
0.0%
51.4%
84.5%
25.2%
TDS
24
10.9%
0.0%
0.0%
0.0%
0.0%
0.0%
3.6%
0.0%
15.7%
0.0%
0.0%
0.0%
0.0%
0.0%
12.0%
4.3%
1.6%
TRX
11
1.2%
0.4%
23.7%
9.4%
75.4%
3.6%
0.0%
75.2%
0.0%
10.4%
26.2%
77.8%
0.0%
68.5%
0.0%
0.0%
27.7%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
62.1%
71.4%
66.8%
73.6%
5.8%
73.7%
51.4%
0.0%
64.3%
43.8%
64.3%
0.0%
0.0%
25.2%
60.8%
73.4%
52.3%
TRX
22
23.0%
17.4%
0.0%
9.6%
0.0%
17.0%
24.2%
0.0%
27.3%
0.6%
0.0%
0.0%
0.0%
0.0%
25.8%
18.3%
8.3%
Deac
29.2%
51.3%
0.0%
6.6%
9.2%
45.1%
20.6%
0.0%
12.9%
20.0%
0.0%
0.0%
0.0%
0.0%
17.6%
2.9%
14.4%
VVLT
49.0%
10.9%
0.0%
12.6%
94.5%
0.0%
51.4%
0.0%
61.7%
0.0%
5.1%
0.0%
0.0%
1.4%
44.2%
0.0%
19.0%
VVT
89.9%
94.4%
95.6%
97.0%
94.5%
97.1%
75.6%
96.1%
90.9%
95.5%
95.0%
96.1%
0.0%
95.5%
80.7%
91.7%
93.3%
CEGR
29.2%
13.1%
0.0%
0.5%
1.1%
10.2%
24.2%
0.0%
28.1%
0.0%
0.0%
35.8%
0.0%
18.8%
34.9%
7.2%
9.8%
Strong
HEV
0.1%
0.0%
3.7%
0.0%
12.0%
2.2%
0.0%
0.0%
0.0%
0.0%
0.6%
0.0%
0.0%
17.9%
0.4%
0.0%
4.5%
EV
3.4%
2.8%
2.3%
1.6%
2.6%
1.3%
14.5%
1.8%
4.1%
3.1%
2.7%
1.8%
100.0%
2.4%
4.2%
3.8%
3.6%
PHEV
4.6%
2.9%
2.9%
2.7%
3.0%
1.6%
10.5%
2.1%
4.3%
1.4%
2.3%
2.1%
0.0%
3.0%
3.2%
4.5%
2.7%
ATK1
0.0%
0.0%
3.7%
1.1%
6.5%
1.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
18.7%
0.0%
0.0%
4.0%
ATK2
16.9%
13.1%
0.0%
0.0%
0.0%
10.2%
20.6%
96.1%
12.2%
0.0%
0.0%
0.0%
0.0%
0.0%
17.3%
2.9%
7.3%
Miller
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
1.4%
0.0%
0.1%
Stop-
Start
67.6%
1.6%
0.0%
11.2%
0.0%
1.5%
35.3%
0.0%
76.7%
0.0%
0.0%
0.0%
0.0%
0.0%
62.6%
5.4%
10.6%
Mild
HEV
12.3%
0.1%
0.0%
0.6%
0.0%
0.0%
40.3%
0.0%
12.8%
0.0%
0.0%
0.0%
0.0%
0.0%
29.1%
0.0%
2.8%
DSL
3.3%
0.0%
0.0%
0.5%
0.0%
0.0%
0.0%
0.0%
0.7%
0.0%
0.0%
0.0%
0.0%
0.0%
11.9%
0.0%
0.9%
Table 12.22 Absolute Technology Penetrations for Cars in the MY2025 Reference Case Using RPEs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
5.0%
5.6%
3.1%
4.1%
2.2%
3.3%
12.8%
6.7%
7.2%
3.5%
5.7%
0.8%
0.0%
2.9%
7.4%
8.9%
4.1%
WR
Net
4.6%
5.6%
2.9%
3.9%
2.2%
3.3%
11.8%
6.7%
6.0%
3.5%
5.7%
0.8%
0.0%
2.7%
6.7%
8.9%
4.0%
TDS
18
48.7%
35.9%
31.9%
48.5%
0.0%
13.6%
48.6%
0.0%
61.6%
9.6%
9.2%
9.0%
0.0%
0.0%
45.8%
83.4%
23.1%
TDS
24
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
11
1.2%
0.0%
18.7%
1.8%
59.7%
2.2%
0.0%
52.9%
0.0%
3.5%
14.3%
76.9%
0.0%
36.3%
0.0%
0.0%
17.9%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
61.4%
71.4%
68.9%
73.6%
18.7%
73.7%
48.6%
22.3%
63.5%
43.8%
72.5%
0.9%
0.0%
57.1%
59.2%
73.4%
59.6%
TRX
22
23.7%
17.8%
2.9%
17.3%
0.0%
18.4%
24.2%
0.0%
27.3%
7.4%
3.7%
0.0%
0.0%
0.2%
25.8%
18.3%
10.5%
Deac
38.6%
46.3%
0.0%
8.1%
9.2%
73.2%
24.2%
0.0%
28.1%
27.2%
0.0%
0.0%
0.0%
0.0%
33.8%
8.3%
19.9%
WLT
50.9%
15.6%
0.0%
12.6%
94.5%
0.0%
46.3%
0.0%
56.2%
0.0%
9.7%
0.0%
0.0%
1.4%
33.4%
0.0%
19.1%
WT
89.9%
94.4%
95.6%
97.0%
94.5%
97.1%
72.8%
96.1%
90.5%
95.5%
95.0%
96.1%
0.0%
95.5%
79.3%
91.7%
93.2%
CEGR
29.2%
13.1%
0.0%
0.5%
1.1%
10.0%
24.2%
0.0%
27.7%
0.0%
0.0%
35.8%
0.0%
18.8%
34.7%
8.3%
9.8%
Strong
HEV
0.1%
0.0%
3.7%
0.0%
12.0%
2.2%
0.0%
0.0%
0.0%
0.0%
0.6%
0.0%
0.0%
17.9%
0.4%
0.0%
4.5%
EV
3.4%
2.8%
2.3%
1.6%
2.6%
1.3%
17.7%
1.8%
4.9%
3.1%
2.7%
1.8%
100.0%
2.4%
6.0%
3.8%
3.7%
PHEV
4.6%
2.9%
2.9%
2.7%
3.0%
1.6%
10.5%
2.1%
4.3%
1.4%
2.3%
2.1%
0.0%
3.0%
3.2%
4.5%
2.7%
ATK1
0.0%
0.0%
3.7%
1.1%
6.5%
1.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
18.7%
0.0%
0.0%
4.0%
ATK2
27.8%
13.1%
0.0%
0.0%
0.0%
10.0%
24.2%
96.1%
28.0%
0.0%
0.0%
0.0%
0.0%
0.0%
29.4%
8.3%
8.9%
Miller
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.9%
0.0%
0.1%
Stop-
Start
73.8%
1.7%
0.0%
11.2%
0.0%
1.5%
34.5%
0.0%
43.9%
0.0%
0.0%
0.0%
0.0%
0.0%
61.1%
17.2%
9.9%
Mild
HEV
12.3%
0.1%
0.0%
0.6%
0.0%
0.0%
38.3%
0.0%
44.8%
0.0%
0.0%
0.0%
0.0%
0.0%
29.4%
0.0%
3.7%
DSL
3.3%
0.0%
0.0%
0.5%
0.0%
0.0%
0.0%
0.0%
0.4%
0.0%
0.0%
0.0%
0.0%
0.0%
11.6%
0.0%
0.9%
                                           12-20

-------
                                                     EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.23  Absolute Technology Penetrations for Trucks in the MY2025 Reference Case Using ICMs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
5.6%
5.6%
5.0%
5.3%
5.7%
7.1%
12.5%
5.4%
6.9%
7.9%
6.7%
5.6%
0.0%
5.2%
8.7%
5.0%
5.8%
WR
Net
4.4%
5.6%
5.0%
5.3%
5.7%
7.1%
11.3%
5.4%
5.7%
7.9%
6.7%
5.5%
0.0%
5.2%
7.4%
4.5%
5.6%
TDS
18
69.4%
49.6%
66.9%
28.9%
0.0%
49.9%
70.0%
0.0%
66.3%
0.0%
57.3%
2.9%
0.0%
51.6%
66.1%
73.1%
42.9%
TDS
24
9.8%
12.4%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
12.2%
0.0%
0.0%
0.0%
0.0%
0.0%
8.1%
11.7%
3.1%
TRX
11
0.0%
0.0%
8.5%
0.0%
35.6%
0.0%
0.0%
19.4%
0.0%
0.0%
17.3%
92.4%
0.0%
16.2%
0.0%
0.0%
14.2%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
71.0%
78.4%
79.8%
79.8%
62.3%
78.8%
70.0%
77.8%
70.0%
78.5%
76.6%
0.2%
0.0%
77.1%
70.3%
76.1%
71.4%
TRX
22
29.0%
19.6%
11.5%
20.0%
0.0%
19.7%
30.0%
0.0%
30.0%
19.6%
3.2%
0.0%
0.0%
3.7%
29.7%
23.9%
12.6%
Deac
18.1%
34.0%
0.0%
64.9%
55.7%
48.5%
30.0%
0.0%
13.9%
0.0%
0.0%
0.0%
0.0%
0.0%
15.4%
15.3%
24.8%
WLT
75.5%
2.2%
0.0%
0.0%
97.9%
0.0%
70.0%
0.0%
78.5%
53.0%
1.9%
0.0%
0.0%
0.0%
74.1%
15.6%
16.4%
VVT
97.2%
97.2%
99.8%
99.8%
97.9%
98.5%
100.0%
97.2%
92.4%
88.4%
98.8%
96.3%
0.0%
98.7%
89.6%
100.0%
98.0%
CEGR
30.6%
19.6%
0.0%
1.9%
0.0%
12.1%
30.0%
0.0%
33.7%
0.0%
0.0%
3.9%
0.0%
6.1%
33.9%
26.9%
8.2%
Strong
HEV
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
11.5%
0.0%
0.0%
0.0%
1.4%
2.1%
0.0%
0.7%
0.9%
0.0%
0.5%
EV
0.0%
0.3%
0.1%
0.1%
1.0%
0.7%
0.0%
1.3%
0.0%
0.6%
0.5%
1.7%
0.0%
0.7%
0.0%
0.0%
0.4%
PHEV
0.0%
0.4%
0.1%
0.1%
1.1%
0.8%
0.0%
1.5%
0.0%
1.3%
0.7%
2.0%
0.0%
0.6%
0.0%
0.0%
0.5%
ATK1
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
ATK2
18.1%
12.5%
0.0%
1.9%
0.0%
12.1%
30.0%
68.9%
13.9%
0.0%
0.0%
0.0%
0.0%
0.0%
15.4%
15.3%
5.5%
Miller
2.6%
0.0%
0.0%
0.0%
0.0%
0.0%
30.0%
0.0%
13.9%
0.0%
0.0%
0.0%
0.0%
0.0%
15.4%
0.0%
1.3%
Stop-
Start
54.8%
29.6%
0.0%
0.0%
0.0%
12.5%
38.5%
0.0%
50.0%
0.0%
0.0%
0.0%
0.0%
0.0%
49.5%
43.3%
10.2%
Mild
HEV
45.2%
0.0%
0.0%
0.0%
0.0%
0.0%
50.0%
0.0%
50.0%
0.0%
0.0%
0.0%
0.0%
0.0%
49.5%
19.5%
4.2%
DSL
2.8%
2.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
7.6%
0.0%
0.0%
0.0%
0.0%
0.0%
10.4%
0.0%
1.0%
Table 12.24 Absolute Technology Penetrations for Trucks in the MY2025 Reference Case Using RPEs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
5.7%
5.6%
5.0%
5.2%
5.7%
7.1%
12.3%
5.4%
5.8%
7.9%
6.7%
5.6%
0.0%
5.2%
8.0%
5.2%
5.7%
WR
Net
4.8%
5.6%
5.0%
5.2%
5.7%
7.1%
11.0%
5.4%
4.6%
7.9%
6.7%
5.5%
0.0%
5.2%
6.7%
4.7%
5.6%
TDS
18
66.1%
46.5%
59.1%
28.9%
0.0%
37.4%
70.0%
0.0%
67.0%
0.0%
26.1%
2.9%
0.0%
5.9%
66.1%
71.9%
32.8%
TDS
24
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
11
0.0%
0.0%
3.2%
0.0%
19.6%
0.0%
0.0%
19.4%
0.0%
0.0%
5.2%
87.9%
0.0%
5.9%
0.0%
0.0%
9.4%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
76.7%
77.3%
79.8%
79.8%
78.3%
78.8%
70.0%
77.8%
70.0%
78.5%
76.5%
4.7%
0.0%
77.0%
70.3%
80.0%
73.1%
TRX
22
23.3%
20.7%
16.7%
20.0%
0.0%
19.7%
30.0%
0.0%
30.0%
19.6%
15.3%
0.0%
0.0%
14.0%
29.7%
20.0%
15.8%
Deac
28.3%
46.3%
0.0%
65.2%
55.7%
61.0%
30.0%
0.0%
25.4%
2.4%
0.7%
0.0%
0.0%
0.0%
23.5%
28.1%
28.2%
WLT
61.1%
4.4%
0.0%
0.0%
97.9%
0.0%
70.0%
0.0%
67.0%
50.5%
2.0%
0.0%
0.0%
0.0%
66.0%
23.8%
16.1%
WT
94.5%
97.2%
99.8%
99.8%
97.9%
98.5%
100.0%
97.2%
92.4%
90.8%
98.8%
96.3%
0.0%
98.7%
89.6%
100.0%
98.0%
CEGR
33.9%
24.4%
0.0%
1.9%
0.0%
24.6%
30.0%
0.0%
33.0%
0.0%
0.7%
3.9%
0.0%
6.1%
33.9%
28.1%
9.4%
Strong
HEV
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
12.1%
0.0%
0.0%
0.0%
1.4%
2.1%
0.0%
0.7%
0.9%
0.0%
0.5%
EV
0.0%
0.3%
0.1%
0.1%
1.0%
0.7%
0.0%
1.3%
0.0%
0.6%
0.5%
1.7%
0.0%
0.7%
0.0%
0.0%
0.4%
PHEV
0.0%
0.4%
0.1%
0.1%
1.1%
0.8%
0.0%
1.5%
0.0%
1.3%
0.7%
2.0%
0.0%
0.6%
0.0%
0.0%
0.5%
ATK1
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
ATK2
28.3%
22.3%
0.0%
1.9%
0.0%
24.6%
30.0%
68.9%
25.4%
0.0%
0.7%
0.0%
0.0%
0.0%
23.5%
28.1%
8.4%
Miller
12.4%
0.0%
0.0%
0.0%
0.0%
0.0%
30.0%
0.0%
7.4%
0.0%
0.0%
0.0%
0.0%
0.0%
11.7%
0.0%
1.1%
Stop-
Start
82.4%
30.7%
0.0%
6.5%
0.0%
0.0%
38.5%
0.0%
50.0%
0.0%
0.0%
0.0%
0.0%
0.0%
49.5%
45.9%
11.6%
Mild
HEV
16.3%
1.1%
0.0%
0.0%
0.0%
0.0%
49.4%
0.0%
50.0%
0.0%
0.0%
0.0%
0.0%
0.0%
49.5%
19.5%
4.1%
DSL
5.5%
2.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
7.6%
0.0%
0.0%
0.0%
0.0%
0.0%
10.4%
0.0%
1.0%
                                            12-21

-------
                                                      EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.25 Absolute Technology Penetrations for the Fleet in the MY2025 Reference Case Using ICMs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
6.9%
6.1%
4.8%
5.2%
4.4%
4.2%
13.3%
6.4%
9.1%
5.5%
6.3%
5.0%
0.0%
4.2%
8.9%
7.1%
5.5%
WR
Net
5.0%
6.1%
4.7%
4.9%
4.3%
4.1%
12.1%
6.4%
8.4%
5.5%
6.3%
4.9%
0.0%
4.0%
8.0%
6.9%
5.2%
TDS
18
52.8%
45.6%
55.2%
38.8%
0.0%
18.0%
66.1%
0.0%
63.7%
6.1%
34.2%
4.2%
0.0%
23.3%
56.9%
78.5%
33.6%
TDS
24
10.6%
8.7%
0.0%
0.0%
0.0%
0.0%
0.7%
0.0%
14.4%
0.0%
0.0%
0.0%
0.0%
0.0%
10.5%
8.1%
2.3%
TRX
11
0.9%
0.1%
14.9%
4.7%
56.8%
3.2%
0.0%
58.1%
0.0%
6.6%
22.7%
89.1%
0.0%
44.8%
0.0%
0.0%
21.3%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
64.2%
76.3%
74.4%
76.7%
32.1%
74.3%
66.1%
23.8%
66.5%
56.2%
69.1%
0.2%
0.0%
48.7%
64.4%
74.8%
61.4%
TRX
22
24.4%
19.0%
6.7%
14.8%
0.0%
17.3%
28.8%
0.0%
28.3%
7.4%
1.3%
0.0%
0.0%
1.7%
27.3%
21.2%
10.4%
Deac
26.5%
39.1%
0.0%
35.7%
30.9%
45.5%
28.0%
0.0%
13.3%
12.8%
0.0%
0.0%
0.0%
0.0%
16.7%
9.3%
19.4%
VVLT
55.3%
4.8%
0.0%
6.3%
96.1%
0.0%
66.1%
0.0%
68.1%
19.0%
3.8%
0.0%
0.0%
0.8%
55.5%
8.1%
17.8%
VVT
91.6%
96.4%
98.0%
98.4%
96.1%
97.3%
94.9%
96.5%
91.5%
93.0%
96.5%
96.2%
0.0%
96.9%
84.1%
96.0%
95.5%
CEGR
29.5%
17.7%
0.0%
1.2%
0.6%
10.4%
28.8%
0.0%
30.2%
0.0%
0.0%
11.0%
0.0%
13.0%
34.5%
17.5%
9.0%
Strong
HEV
0.1%
0.0%
1.5%
0.0%
6.4%
2.0%
9.1%
0.0%
0.0%
0.0%
0.9%
1.6%
0.0%
10.1%
0.6%
0.0%
2.6%
EV
2.6%
1.0%
1.0%
0.9%
1.8%
1.3%
3.0%
1.6%
2.5%
2.2%
1.8%
1.7%
100.0%
1.6%
2.6%
1.8%
2.1%
PHEV
3.5%
1.2%
1.3%
1.4%
2.1%
1.5%
2.2%
1.9%
2.7%
1.4%
1.7%
2.0%
0.0%
1.9%
2.0%
2.1%
1.7%
ATK1
0.0%
0.0%
1.5%
0.5%
3.4%
1.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
10.2%
0.0%
0.0%
2.1%
ATK2
17.2%
12.7%
0.0%
1.0%
0.0%
10.4%
28.0%
87.8%
12.8%
0.0%
0.0%
0.0%
0.0%
0.0%
16.6%
9.3%
6.4%
Miller
0.6%
0.0%
0.0%
0.0%
0.0%
0.0%
23.7%
0.0%
5.3%
0.0%
0.0%
0.0%
0.0%
0.0%
6.7%
0.0%
0.7%
Stop-
Start
64.5%
21.2%
0.0%
5.6%
0.0%
2.8%
37.8%
0.0%
66.5%
0.0%
0.0%
0.0%
0.0%
0.0%
57.6%
25.1%
10.4%
Mild
HEV
20.1%
0.0%
0.0%
0.3%
0.0%
0.0%
48.0%
0.0%
27.0%
0.0%
0.0%
0.0%
0.0%
0.0%
36.8%
10.1%
3.5%
DSL
3.2%
1.5%
0.0%
0.3%
0.0%
0.0%
0.0%
0.0%
3.4%
0.0%
0.0%
0.0%
0.0%
0.0%
11.3%
0.0%
1.0%
Table 12.26 Absolute Technology Penetrations for the Fleet in the MY2025 Reference Case Using RPEs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
5.2%
5.6%
4.2%
4.6%
3.9%
3.8%
12.4%
6.3%
6.7%
5.1%
6.1%
4.5%
0.0%
3.9%
7.7%
7.0%
4.9%
WR
Net
4.7%
5.6%
4.1%
4.5%
3.9%
3.8%
11.2%
6.3%
5.5%
5.1%
6.1%
4.4%
0.0%
3.8%
6.7%
6.7%
4.7%
TDS
18
52.9%
43.4%
47.7%
38.7%
0.0%
16.5%
65.5%
0.0%
63.6%
6.1%
15.9%
4.2%
0.0%
2.7%
53.5%
77.4%
27.7%
TDS
24
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
11
0.9%
0.0%
9.7%
0.9%
41.0%
2.0%
0.0%
42.6%
0.0%
2.3%
10.7%
85.4%
0.0%
22.6%
0.0%
0.0%
13.8%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
65.0%
75.6%
75.2%
76.7%
46.5%
74.3%
65.5%
39.3%
66.0%
56.2%
74.1%
3.9%
0.0%
66.1%
63.4%
76.8%
66.0%
TRX
22
23.6%
19.8%
10.9%
18.6%
0.0%
18.6%
28.8%
0.0%
28.3%
11.8%
8.3%
0.0%
0.0%
6.5%
27.3%
19.2%
13.0%
Deac
36.1%
46.3%
0.0%
36.6%
30.9%
71.7%
28.8%
0.0%
27.1%
18.3%
0.3%
0.0%
0.0%
0.0%
29.9%
18.6%
23.9%
VVLT
53.3%
7.7%
0.0%
6.3%
96.1%
0.0%
65.0%
0.0%
60.3%
18.1%
6.6%
0.0%
0.0%
0.8%
45.7%
12.4%
17.7%
VVT
91.0%
96.4%
98.0%
98.4%
96.1%
97.3%
94.3%
96.5%
91.2%
93.8%
96.5%
96.2%
0.0%
96.9%
83.2%
96.0%
95.5%
CEGR
30.3%
21.0%
0.0%
1.2%
0.6%
11.7%
28.8%
0.0%
29.7%
0.0%
0.3%
11.0%
0.0%
13.0%
34.4%
18.6%
9.6%
Strong
HEV
0.1%
0.0%
1.5%
0.0%
6.4%
2.0%
9.5%
0.0%
0.0%
0.0%
0.9%
1.6%
0.0%
10.1%
0.6%
0.0%
2.6%
EV
2.6%
1.0%
1.0%
0.9%
1.8%
1.3%
3.7%
1.6%
3.0%
2.2%
1.8%
1.7%
100.0%
1.6%
3.7%
1.8%
2.2%
PHEV
3.5%
1.2%
1.3%
1.4%
2.1%
1.5%
2.2%
1.9%
2.7%
1.4%
1.7%
2.0%
0.0%
1.9%
2.0%
2.1%
1.7%
ATK1
0.0%
0.0%
1.5%
0.5%
3.4%
1.1%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
10.2%
0.0%
0.0%
2.1%
ATK2
27.9%
19.5%
0.0%
1.0%
0.0%
11.7%
28.8%
87.8%
27.0%
0.0%
0.3%
0.0%
0.0%
0.0%
27.1%
18.6%
8.7%
Miller
3.0%
0.0%
0.0%
0.0%
0.0%
0.0%
23.7%
0.0%
2.8%
0.0%
0.0%
0.0%
0.0%
0.0%
5.0%
0.0%
0.6%
Stop-
Start
75.9%
22.1%
0.0%
8.8%
0.0%
1.3%
37.7%
0.0%
46.2%
0.0%
0.0%
0.0%
0.0%
0.0%
56.8%
32.1%
10.7%
Mild
HEV
13.2%
0.8%
0.0%
0.3%
0.0%
0.0%
47.1%
0.0%
46.8%
0.0%
0.0%
0.0%
0.0%
0.0%
37.0%
10.1%
3.9%
DSL
3.9%
1.5%
0.0%
0.3%
0.0%
0.0%
0.0%
0.0%
3.1%
0.0%
0.0%
0.0%
0.0%
0.0%
11.2%
0.0%
1.0%
                                             12-22

-------
                                              EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.27 Summary of Absolute Technology Penetrations in the MY2025 Reference Case
Indirect Cost
Approach
ICM
ICM
ICM
RPE
RPE
RPE
C/T/Fleet
C
T
Fleet
C
T
Fleet
WR
Tech
5.2%
5.8%
5.5%
4.1%
5.7%
4.9%
WR
Net
4.8%
5.6%
5.2%
4.0%
5.6%
4.7%
IDS
18
25.2%
42.9%
33.6%
23.1%
32.8%
27.7%
IDS
24
1.6%
3.1%
2.3%
0.0%
0.0%
0.0%
TRX
11
27.7%
14.2%
21.3%
17.9%
9.4%
13.8%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
52.3%
71.4%
61.4%
59.6%
73.1%
66.0%
TRX
22
8.3%
12.6%
10.4%
10.5%
15.8%
13.0%
Deac
14.4%
24.8%
19.4%
19.9%
28.2%
23.9%
VVLT
19.0%
16.4%
17.8%
19.1%
16.1%
17.7%
VVT
93.3%
98.0%
95.5%
93.2%
98.0%
95.5%
CEGR
9.8%
8.2%
9.0%
9.8%
9.4%
9.6%
Strong
HEV
4.5%
0.5%
2.6%
4.5%
0.5%
2.6%
EV
3.6%
0.4%
2.1%
3.7%
0.4%
2.2%
PHEV
2.7%
0.5%
1.7%
2.7%
0.5%
1.7%
ATK1
4.0%
0.0%
2.1%
4.0%
0.0%
2.1%
ATK2
7.3%
5.5%
6.4%
8.9%
8.4%
8.7%
Miller
0.1%
1.3%
0.7%
0.1%
1.1%
0.6%
Stop-
Start
10.6%
10.2%
10.4%
9.9%
11.6%
10.7%
Mild
HEV
2.8%
4.2%
3.5%
3.7%
4.1%
3.9%
DSL
0.9%
1.0%
1.0%
0.9%
1.0%
1.0%
                                      12-23

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
   Mass reduction technology is applied along a continuum of possible levels, and values shown
in the tables above represent the average percentage mass reduction applied (WRtech) and
average percentage net mass reduction (WRnet). The values above do not indicate the proportion
of the fleet with the technology applied, as is the case with the other technologies shown. Not
readily apparent in the tables above is the number, or percentage, of vehicles that receive specific
levels of mass reduction.  The table below provides more detail on mass reduction technology in
our projections by showing the percentage of vehicles that receive the level of mass reduction
within the given mass reduction ranges.  Note that we account for the additional mass associated
with batteries and electrical components of EVs and PHEVs, which explains the difference
between "WRtech" and "WRnet."  "Baseline" represents the amount of mass reduction relative to
EPA's "null" or "floor" (i.e., in the case of weight reduction, EPA's "null" is the 2008 baseline
fleet used in the 2012 FRM) present in MY2014 vehicles with MY2025 projected volumes.  In
the table, we show results excluding the ZEV program vehicles because, as noted above, roughly
2.5 percent of the fleet (the fleet reflecting the ZEV program) was "built" with 20 percent mass
reduction technology applied (WRtech) and 20 percent mass reduction on net (WRnet).
 Table 12.28 Percentage of Vehicles Receiving the Mass Reduction Levels within the Indicated Ranges in the
              MY2025 Reference Case Using ICMs and AEO Reference Case Fuel Prices
Fleet
Including ZEV Program Vehicles
Excluding ZEV Program Vehicles
(as explained above)
%MR Range
<=5%
5% to <=10%
10% to <=15%
15% to <=20%
<=5%
5% to <=10%
10% to <=15%
15% to <=20%
Baseline
87.0%
9.1%
0.9%
3.0%
89.3%
9.3%
0.9%
0.5%
WRtech
57.4%
30.7%
7.3%
4.6%
58.9%
31.5%
7.5%
2.1%
WRnet
61.0%
28.0%
8.3%
2.7%
62.6%
28.8%
8.5%
0.2%
12.1.1.3.2
Control Case
   The technology penetration rates in the MY2025 control case (that is, the case where the
MY2025 standards are in effect in MY2025), again in absolute terms, are presented for cars,
trucks and the fleet, using both ICMs and RPEs, in the tables below.  We also present the
technology penetration changes, i.e., the technology added to move from compliance with the
reference case standards to the control case standards, for cars, trucks and the fleet using both
ICMs and RPEs in the tables below. All technology penetration rate tables use the AEO 2015
reference fuel price case.

   Much like both the 2012 FRM and the 2015 NAS report, the results from the control case
show that the MY 2025 standards can be met largely through the application of advanced
gasoline engines and transmissions and moderate hybridization.  The technology penetrations for
the previously identified technologies are shown in the last row of Table 12.33 for the entire
light-duty fleet. (This table presents fleet level technology penetrations using ICMs).

   For advanced gasoline engines EPA has projected that the fleet would be 33 percent 18-bar
and 24-bar turbo-charged engines and 44 percent Atkinson 2 engines. This similar penetration of
two competing engine technologies demonstrates that there are multiple  cost effective advanced
gasoline technologies available to manufacturers.  In order to acknowledge that manufacturers
                                             12-24

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
may choose to focus on turbo-downsized technology over Atkinson, EPA conducted a sensitivity
analysis restricting Atkinson 2 technology application as described in the Sensitivity Analysis
Results below.  In addition to turbo-charging and Atkinson cycle, EPA has also projected
cylinder deactivation (DEAC), variable valve timing (VVT) and cooled EGR will be prominent
engine technologies, with respective penetration rates of 54 percent, 96 percent, and 53 percent.
With respect to transmissions, EPA has projected that over 90 percent of the transmissions will
be high ratio spread (TRX21+TRX22) and 39 percent (TRX22) of these transmissions will also
implement further improvements in transmission efficiency beyond current transmissions.11

   Stop-start and Mild HEV technologies, such as 48-volt systems, are anticipated to be applied
with increasing frequency.  48-volt mild hybrids help improve the overall  efficiency of
conventional powertrains at less expense compared to strong hybridization. Stop-start is
projected to penetrate the market in 20 percent of the fleet, and Mild HEV's at an 18 percent
penetration.

   Mass reduction is also expected to be applied at moderate levels across the majority of the
fleet. For MY 2025 EPA has projected an average mass reduction technology penetration rate
for the entire fleet of 7 percent (WR Tech) which, when taking into consideration the additional
mass of electrification, yields a net mass reduction of 6 percent (WR Net).  The highest average
amount of mass reduction for an individual manufacturer is projected to be 13 percent for
Jaguar-Land Rover and the  lowest mass reduction is projected to be 5 percent for Toyota.

   For some manufacturers, strong electrification is expected to be utilized, however, for the
overall fleet EPA has projected a minimal amount of strong electrification technology
penetration. For strong HEV's, battery electric vehicles (EV), and plug-in hybrid electric
vehicles (PHEV), EPA has projected fleet technology penetration rates of 3 percent, 2 percent,
and 2 percent respectively.  The highest penetration rates for strong HEVs was projected at 11
percent for JLR, for EVs, Volkswagen has been projected to utilize 9 percent, and for PHEVs,
BMW is projected to utilize 4 percent. EPA notes that our analysis included consideration for
compliance with other related regulations including CARET s ZEV regulation that has also been
adopted by nine other states under section 177 of the Federal Clean Air Act. Therefore, some of
the EV and PHEV penetration in the following tables is ZEV program-related (2.6 percent of the
combined fleet), some is in  EPA's purchased fleet projections (1.2 percent of the combined
fleet), and some is generated by OMEGA to reach compliance (an additional 0.5 percent of the
combined fleet for a total of 4.3 percent in the AEO 2015 reference fuel price and ICM case).
See Table 12.33 where the final EV (2.6 percent) and PHEV (1.7 percent) penetrations can be
added to 4.3 percent; see Table 12.39 where the incremental EV penetration is shown as 1
percent, rounded from 0.5 percent. EPA's analysis also reflects considerable penetration of
certain advanced engine technologies such as the Atkinson-2 technology introduced since the
H EPA has used transmission designations TRX11, TRX12, TRX21 and TRX22 to represent levels of improvement
  to the transmission in the baseline fleet. As such, these transmission designations could include automatic
  transmissions, dual clutch transmissions or CVTs. The point is, TRX21 and TRX22 transmissions have wider
  ratio spreads, regardless of the type of transmission, than do TRX11 and TRX12 transmissions. Similarly, TRX12
  and TRX22 transmissions have additional efficiency improvements beyond those found in TRX11 and TRX21
  transmissions.
                                              12-25

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
2012 FRM.  Had the analysis not taken these factors into account, it is likely that estimates of
strong hybridization and electrification penetration rates would be higher.1

   The tables that follow for control case technology penetrations are:

       •  Table 12.29 Absolute Technology Penetrations for Cars in the MY2025 Control Case
          Using ICMs
       •  Table 12.30 Absolute Technology Penetrations for Cars in the MY2025 Control Case
          Using RPEs
       •  Table 12.31 Absolute Technology Penetrations for Trucks in the MY2025 Control
          Case Using ICMs
       •  Table 12.32 Absolute Technology Penetrations for Trucks in the MY2025 Control
          Case Using RPEs
       •  Table 12.33 Absolute Technology Penetrations for the Fleet in the MY2025 Control
          Case Using ICMs
       •  Table 12.34 Absolute Technology Penetrations for the Fleet in the MY2025 Control
          Case Using RPEs
   The tables that follow for control case incremental technology penetrations are:

       •  Table 12.35 Incremental Technology Penetrations for Cars in the MY2025 Central
          Analysis Using ICMs
       •  Table 12.36 Incremental Technology Penetrations for Cars in the MY2025 Central
          Analysis Using RPEs
       •  Table 12.37 Incremental Technology Penetrations for Trucks in the MY2025 Central
          Analysis Using ICMs
       •  Table 12.38 Incremental Technology Penetrations for Trucks in the MY2025 Central
          Analysis Using RPEs
       •  Table 12.39 Incremental Technology Penetrations for the Fleet in the MY2025
          Central Analysis Using ICMs
       •  Table 12.40 Incremental Technology Penetrations for the Fleet in the MY2025
          Central Analysis Using RPEs
   The final two tables show summaries of the above control case tables.

       •  Table 12.41 Summary of Absolute Technology Penetrations in the MY2025 Control
          Case
       •  Table 12.42 Summary of Incremental Technology Penetrations in the MY2025
          Control Case
1 This section is focused on describing the results of the OMEGA model for this Draft TAR. As noted in the
  Executive Summary and elsewhere, there are differences between the EPA and DOT approaches that derive
  different penetration rates for hybrid as well as other technologies. These derive from a range of factors, including
  but not limited to different penetration rates of EVs and PHEVs in the two agencies' reference fleets, and
  differences in technology effectiveness assumptions, and others.
                                              12-26

-------
                                                   EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.29  Absolute Technology Penetrations for Cars in the MY2025 Control Case Using ICMs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
9.1%
8.7%
6.1%
6.8%
4.2%
4.3%
19.9%
7.1%
11.5%
5.6%
7.3%
3.1%
0.0%
4.6%
10.0%
10.9%
6.3%
WR
Net
7.8%
8.4%
5.7%
5.8%
4.1%
4.0%
17.4%
7.1%
9.5%
5.6%
7.3%
3.1%
0.0%
4.1%
8.2%
9.9%
5.8%
TDS
18
18.1%
9.8%
50.4%
39.5%
11.5%
3.4%
0.3%
0.0%
15.4%
9.6%
27.6%
10.1%
0.0%
14.1%
11.4%
20.5%
20.5%
TDS
24
17.7%
22.1%
0.0%
0.0%
0.0%
8.5%
0.0%
0.0%
45.6%
0.0%
0.1%
0.0%
0.0%
0.0%
46.5%
11.9%
7.6%
TRX
11
1.2%
0.0%
4.4%
0.6%
11.8%
2.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
18.9%
0.0%
0.0%
4.7%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
6.5%
22.1%
85.7%
69.5%
69.4%
55.1%
0.0%
75.2%
2.1%
51.8%
90.0%
77.8%
0.0%
74.4%
0.4%
0.0%
58.9%
TRX
22
76.0%
68.3%
0.4%
22.5%
0.0%
37.1%
62.2%
0.0%
82.7%
3.0%
0.4%
0.0%
0.0%
0.3%
81.6%
91.7%
24.2%
Deac
52.6%
62.4%
40.7%
51.9%
57.9%
83.0%
60.4%
20.6%
23.6%
64.1%
63.3%
48.1%
0.0%
58.1%
24.3%
59.4%
54.3%
VVLT
18.1%
2.0%
0.0%
6.1%
25.1%
1.6%
1.7%
0.0%
20.9%
0.0%
0.0%
0.0%
0.0%
0.0%
18.0%
20.5%
6.2%
VVT
88.4%
94.3%
95.6%
97.0%
94.5%
97.1%
62.2%
96.1%
84.7%
95.5%
95.0%
96.1%
0.0%
95.5%
82.2%
91.7%
93.1%
CEGR
68.9%
66.0%
30.7%
41.5%
39.7%
71.2%
60.4%
20.6%
68.8%
48.1%
39.0%
40.3%
0.0%
42.5%
70.2%
68.8%
48.4%
Strong
HEV
0.1%
0.0%
3.7%
0.0%
12.0%
2.2%
0.0%
0.0%
0.0%
0.0%
0.6%
0.0%
0.0%
17.9%
0.4%
0.0%
4.5%
EV
7.8%
2.8%
2.3%
1.6%
2.6%
1.3%
29.9%
1.8%
11.3%
3.1%
2.7%
1.8%
100.0%
2.4%
13.6%
3.8%
4.6%
PHEV
4.6%
2.9%
2.9%
2.7%
3.0%
1.6%
11.8%
2.1%
4.3%
1.4%
2.3%
2.1%
0.0%
3.0%
3.2%
4.5%
2.7%
ATK1
0.0%
0.0%
3.7%
1.1%
6.5%
1.2%
1.4%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
18.7%
0.0%
0.0%
4.0%
ATK2
51.4%
56.5%
30.7%
44.3%
38.6%
82.8%
60.4%
96.1%
23.6%
48.1%
38.9%
4.5%
0.0%
23.7%
24.3%
56.9%
43.9%
Miller
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
60.4%
0.0%
8.3%
0.0%
0.0%
0.0%
0.0%
0.0%
17.1%
0.0%
1.4%
Stop-
Start
36.5%
37.4%
0.0%
20.2%
0.0%
25.3%
0.0%
0.0%
10.9%
0.0%
0.0%
0.0%
0.0%
0.0%
16.4%
51.6%
12.1%
Mild
HEV
49.2%
12.4%
0.0%
3.7%
0.0%
5.0%
60.7%
0.0%
71.8%
0.0%
0.0%
0.0%
0.0%
0.0%
66.7%
40.1%
10.4%
DSL
0.5%
0.0%
0.0%
0.5%
0.0%
0.0%
0.0%
0.0%
0.1%
0.0%
0.0%
0.0%
0.0%
0.0%
1.3%
0.0%
0.2%
Table 12.30 Absolute Technology Penetrations for Cars in the MY2025 Control Case Using RPEs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
5.3%
6.1%
3.5%
5.0%
2.3%
3.6%
15.3%
6.7%
6.8%
3.9%
5.9%
0.8%
0.0%
3.1%
7.9%
9.1%
4.5%
WR
Net
4.6%
5.8%
3.3%
4.7%
2.3%
3.5%
12.6%
6.7%
5.6%
3.9%
5.9%
0.8%
0.0%
2.9%
6.4%
8.2%
4.2%
TDS
18
16.2%
9.5%
50.4%
39.3%
3.6%
3.4%
0.0%
0.0%
13.0%
2.4%
27.6%
10.1%
0.0%
14.1%
9.7%
15.9%
19.3%
TDS
24
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
11
1.2%
0.0%
4.4%
0.6%
11.8%
2.2%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
18.9%
0.0%
0.0%
4.7%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
6.5%
11.7%
85.7%
60.5%
69.4%
47.2%
0.0%
75.2%
2.1%
51.8%
90.0%
77.8%
0.0%
74.4%
0.4%
0.0%
55.7%
TRX
22
74.0%
78.7%
0.4%
32.5%
0.0%
45.0%
62.4%
0.0%
80.2%
3.0%
0.4%
0.0%
0.0%
0.3%
78.0%
87.1%
27.1%
Deac
68.6%
84.8%
40.7%
52.1%
66.2%
91.5%
58.9%
20.6%
69.3%
71.3%
52.9%
48.7%
0.0%
58.1%
66.2%
71.2%
61.4%
VVLT
16.2%
7.6%
0.0%
6.1%
30.2%
3.0%
3.6%
0.0%
13.0%
0.0%
0.0%
0.0%
0.0%
0.0%
9.4%
15.9%
6.5%
WT
84.8%
94.3%
95.6%
97.2%
94.5%
97.1%
62.4%
96.1%
82.3%
95.5%
95.0%
96.1%
0.0%
95.5%
75.8%
87.1%
92.5%
CEGR
68.9%
66.0%
30.7%
41.5%
29.3%
71.2%
58.9%
20.6%
68.8%
55.3%
24.6%
40.3%
0.0%
42.5%
70.2%
68.8%
46.1%
Strong
HEV
0.1%
0.0%
3.7%
0.0%
12.0%
2.2%
0.0%
0.0%
0.0%
0.0%
0.6%
0.0%
0.0%
17.9%
0.4%
0.0%
4.5%
EV
9.7%
2.8%
2.3%
1.6%
2.6%
1.3%
29.9%
1.8%
13.9%
3.1%
2.7%
1.8%
100.0%
2.4%
15.4%
8.7%
4.8%
PHEV
4.6%
2.9%
2.9%
2.7%
3.0%
1.6%
13.6%
2.1%
4.3%
1.4%
2.3%
2.1%
0.0%
3.0%
3.2%
4.5%
2.7%
ATK1
0.0%
0.0%
3.7%
1.1%
6.5%
1.2%
3.6%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
18.7%
0.0%
0.0%
4.0%
ATK2
67.4%
78.5%
30.7%
44.3%
28.2%
83.0%
58.9%
96.1%
69.3%
55.3%
24.6%
4.5%
0.0%
23.7%
66.2%
68.8%
47.6%
Miller
1.2%
0.0%
0.0%
0.0%
0.0%
0.0%
58.9%
0.0%
11.4%
0.0%
0.0%
0.0%
0.0%
0.0%
16.9%
0.0%
1.5%
Stop-
Start
53.7%
56.7%
0.0%
20.3%
0.0%
57.6%
0.0%
0.0%
38.8%
0.0%
0.1%
0.0%
0.0%
0.0%
27.4%
51.6%
19.9%
Mild
HEV
30.2%
14.8%
0.0%
6.6%
0.0%
3.6%
58.9%
0.0%
41.4%
0.0%
0.0%
0.0%
0.0%
0.0%
54.0%
35.5%
8.4%
DSL
2.2%
0.0%
0.0%
0.5%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
6.0%
0.0%
0.5%
                                          12-27

-------
                                                    EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.31 Absolute Technology Penetrations for Trucks in the MY2025 Control Case Using ICMs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
5.5%
6.5%
5.2%
5.4%
7.4%
10.1%
14.0%
8.9%
6.8%
10.2%
8.2%
10.1%
0.0%
6.6%
9.6%
5.2%
6.9%
WR
Net
3.5%
4.8%
5.0%
4.8%
7.4%
8.8%
12.1%
8.9%
4.8%
10.2%
8.1%
9.9%
0.0%
6.6%
7.5%
3.2%
6.2%
TDS
18
25.0%
16.0%
28.9%
9.5%
55.7%
12.5%
25.0%
28.4%
25.0%
9.8%
40.8%
5.9%
0.0%
29.6%
25.7%
25.0%
24.5%
TDS
24
17.9%
34.1%
7.9%
14.2%
0.0%
27.5%
0.0%
0.0%
0.0%
0.0%
15.9%
0.0%
0.0%
21.9%
0.0%
21.4%
14.8%
TRX
11
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
1.4%
2.1%
0.0%
0.7%
0.0%
0.0%
0.3%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
0.0%
0.0%
38.4%
38.4%
89.9%
0.0%
0.0%
97.2%
0.0%
98.2%
70.1%
90.5%
0.0%
77.3%
0.9%
0.0%
43.4%
TRX
22
100.0%
99.0%
61.3%
61.4%
8.0%
98.5%
100.0%
0.0%
100.0%
0.0%
26.6%
0.0%
0.0%
19.0%
99.1%
100.0%
54.7%
Deac
56.0%
47.6%
63.0%
76.1%
42.2%
58.5%
75.0%
51.6%
72.0%
88.4%
41.5%
0.0%
0.0%
45.0%
70.1%
53.6%
52.8%
VVLT
42.9%
28.2%
7.3%
6.5%
0.0%
12.5%
25.0%
0.0%
25.0%
0.0%
1.2%
0.0%
0.0%
0.0%
25.6%
46.4%
10.5%
WT
98.9%
97.7%
99.8%
99.8%
97.9%
98.5%
100.0%
97.2%
97.0%
98.2%
98.8%
96.3%
0.0%
98.7%
95.8%
100.0%
98.5%
CEGR
75.0%
74.5%
62.7%
72.5%
25.7%
73.9%
75.0%
51.6%
75.0%
66.3%
47.5%
3.9%
0.0%
58.8%
74.3%
75.0%
58.9%
Strong
HEV
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
14.4%
0.0%
0.0%
0.0%
1.4%
2.1%
0.0%
0.7%
0.9%
0.0%
0.5%
EV
0.0%
0.3%
0.1%
0.1%
1.0%
0.7%
0.0%
1.3%
0.0%
0.6%
0.5%
1.7%
0.0%
0.7%
0.0%
0.0%
0.4%
PHEV
0.0%
0.4%
0.1%
0.1%
1.1%
0.8%
0.0%
1.5%
0.0%
1.3%
0.7%
2.0%
0.0%
0.6%
0.0%
0.0%
0.5%
ATK1
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
ATK2
56.0%
42.2%
55.7%
60.3%
25.7%
58.5%
75.0%
68.9%
72.0%
66.3%
34.9%
0.0%
0.0%
36.4%
70.1%
53.6%
45.0%
Miller
47.6%
0.0%
0.0%
0.0%
0.0%
0.0%
75.0%
0.0%
72.0%
0.0%
0.0%
0.0%
0.0%
0.0%
70.1%
33.6%
5.9%
Stop-
Start
20.0%
18.3%
63.8%
65.0%
0.0%
10.0%
8.5%
0.0%
20.0%
0.0%
12.6%
0.0%
0.0%
17.5%
19.8%
20.0%
29.5%
Mild
HEV
80.0%
70.9%
7.8%
24.1%
0.0%
52.1%
77.1%
0.0%
80.0%
0.0%
2.2%
0.0%
0.0%
0.0%
79.3%
80.0%
26.9%
DSL
1.1%
1.6%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
3.0%
0.0%
0.0%
0.0%
0.0%
0.0%
4.2%
0.0%
0.5%
Table 12.32  Absolute Technology Penetrations for Trucks in the MY2025 Control Case Using RPEs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
5.5%
5.8%
5.2%
5.4%
6.2%
7.7%
13.9%
6.3%
6.3%
10.2%
7.6%
5.6%
0.0%
6.6%
8.3%
5.0%
6.1%
WR
Net
3.5%
4.1%
5.2%
5.2%
6.2%
6.4%
12.0%
6.3%
4.3%
10.2%
7.5%
5.5%
0.0%
6.6%
6.3%
3.8%
5.6%
TDS
18
25.0%
16.0%
26.0%
7.2%
33.2%
12.5%
25.0%
24.8%
25.0%
9.8%
15.4%
5.9%
0.0%
32.1%
25.7%
25.0%
20.1%
TDS
24
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
11
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
1.4%
2.1%
0.0%
0.7%
0.0%
0.0%
0.3%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
0.0%
0.0%
38.1%
2.3%
89.9%
0.0%
0.0%
97.2%
0.0%
98.2%
77.5%
90.5%
0.0%
77.3%
0.9%
0.0%
38.0%
TRX
22
100.0%
99.0%
61.7%
97.5%
8.0%
98.5%
100.0%
0.0%
100.0%
0.0%
18.5%
0.0%
0.0%
19.0%
99.1%
100.0%
60.1%
Deac
73.9%
81.7%
73.8%
92.6%
64.7%
86.0%
75.0%
55.3%
72.0%
88.4%
82.9%
25.9%
0.0%
64.5%
70.1%
75.0%
73.8%
VVLT
25.0%
16.0%
2.1%
6.5%
0.0%
12.5%
25.0%
0.0%
25.0%
0.0%
1.2%
0.0%
0.0%
0.0%
25.6%
25.0%
7.0%
VVT
98.9%
97.7%
99.8%
99.8%
97.9%
98.5%
100.0%
97.2%
97.0%
98.2%
98.8%
96.3%
0.0%
98.7%
95.8%
100.0%
98.5%
CEGR
75.0%
74.5%
65.5%
69.9%
48.1%
73.9%
75.0%
55.3%
75.0%
66.3%
73.0%
29.5%
0.0%
56.3%
74.3%
75.0%
64.2%
Strong
HEV
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
16.8%
0.0%
0.0%
0.0%
1.4%
2.1%
0.0%
0.7%
0.9%
0.0%
0.6%
EV
0.0%
0.3%
0.1%
0.1%
1.0%
0.7%
0.0%
1.3%
0.0%
0.6%
0.5%
1.7%
0.0%
0.7%
0.0%
0.0%
0.4%
PHEV
0.0%
0.4%
0.1%
0.1%
1.1%
0.8%
0.0%
1.5%
0.0%
1.3%
0.7%
2.0%
0.0%
0.6%
0.0%
0.0%
0.5%
ATK1
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
ATK2
73.9%
76.3%
66.4%
71.8%
48.1%
86.0%
75.0%
72.5%
72.0%
66.3%
73.7%
25.9%
0.0%
54.6%
70.1%
75.0%
64.8%
Miller
47.6%
0.0%
0.0%
0.0%
0.0%
0.0%
75.0%
0.0%
72.0%
0.0%
0.0%
0.0%
0.0%
0.0%
70.1%
14.0%
5.8%
Stop-
Start
20.0%
21.2%
80.2%
82.3%
8.0%
10.0%
8.5%
3.6%
20.0%
0.0%
54.0%
0.0%
0.0%
36.9%
19.8%
51.3%
41.9%
Mild
HEV
80.0%
70.9%
2.6%
9.9%
0.0%
52.1%
74.7%
0.0%
80.0%
0.0%
0.7%
0.0%
0.0%
0.0%
79.3%
48.7%
23.4%
DSL
1.1%
1.6%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
3.0%
0.0%
0.0%
0.0%
0.0%
0.0%
4.2%
0.0%
0.5%
                                           12-28

-------
                                                     EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.33 Absolute Technology Penetrations for the Fleet in the MY2025 Control Case Using ICMs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
8.2%
7.1%
5.6%
6.1%
5.7%
5.0%
15.2%
7.7%
9.7%
7.2%
7.7%
8.5%
0.0%
5.5%
9.8%
7.9%
6.6%
WR
Net
6.8%
5.9%
5.3%
5.3%
5.7%
4.6%
13.2%
7.7%
7.8%
7.2%
7.6%
8.4%
0.0%
5.2%
7.9%
6.4%
6.0%
TDS
18
19.7%
14.2%
37.9%
24.6%
32.1%
4.5%
19.8%
8.7%
19.1%
9.6%
32.9%
6.9%
0.0%
21.1%
16.8%
22.8%
22.4%
TDS
24
17.8%
30.5%
4.6%
7.1%
0.0%
10.8%
0.0%
0.0%
28.2%
0.0%
6.3%
0.0%
0.0%
9.9%
28.9%
16.8%
11.0%
TRX
11
0.9%
0.0%
1.9%
0.3%
6.3%
2.0%
0.0%
0.0%
0.0%
0.0%
0.5%
1.6%
0.0%
10.7%
0.0%
0.0%
2.6%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
5.0%
6.6%
58.2%
54.0%
79.0%
48.5%
0.0%
81.9%
1.3%
68.4%
82.1%
87.7%
0.0%
75.7%
0.6%
0.0%
51.5%
TRX
22
81.7%
89.8%
35.8%
41.9%
3.7%
44.4%
92.1%
0.0%
89.3%
1.9%
10.8%
0.0%
0.0%
8.8%
88.2%
96.0%
38.7%
Deac
53.4%
52.0%
53.7%
63.9%
50.6%
80.0%
72.0%
30.1%
42.1%
72.8%
54.7%
10.8%
0.0%
52.2%
41.6%
56.4%
53.6%
VVLT
24.0%
20.4%
4.3%
6.3%
13.4%
2.9%
20.1%
0.0%
22.5%
0.0%
0.5%
0.0%
0.0%
0.0%
20.9%
34.0%
8.3%
VVT
90.9%
96.7%
98.0%
98.4%
96.1%
97.3%
92.1%
96.5%
89.4%
96.5%
96.5%
96.2%
0.0%
96.9%
87.4%
96.0%
95.7%
CEGR
70.4%
71.9%
49.3%
57.0%
33.1%
71.5%
72.0%
30.1%
71.2%
54.6%
42.4%
12.0%
0.0%
49.9%
71.7%
72.0%
53.4%
Strong
HEV
0.1%
0.0%
1.5%
0.0%
6.4%
2.0%
11.4%
0.0%
0.0%
0.0%
0.9%
1.6%
0.0%
10.1%
0.6%
0.0%
2.6%
EV
6.0%
1.0%
1.0%
0.9%
1.8%
1.3%
6.3%
1.6%
7.0%
2.2%
1.8%
1.7%
100.0%
1.6%
8.5%
1.8%
2.6%
PHEV
3.5%
1.2%
1.3%
1.4%
2.1%
1.5%
2.5%
1.9%
2.7%
1.4%
1.7%
2.0%
0.0%
1.9%
2.0%
2.1%
1.7%
ATK1
0.0%
0.0%
1.5%
0.5%
3.4%
1.1%
0.3%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
10.2%
0.0%
0.0%
2.1%
ATK2
52.5%
46.4%
45.2%
52.2%
32.6%
79.9%
72.0%
87.8%
42.1%
54.6%
37.4%
1.0%
0.0%
29.5%
41.6%
55.2%
44.4%
Miller
11.3%
0.0%
0.0%
0.0%
0.0%
0.0%
72.0%
0.0%
32.6%
0.0%
0.0%
0.0%
0.0%
0.0%
37.2%
17.5%
3.6%
Stop-
Start
32.6%
24.0%
37.1%
42.5%
0.0%
23.5%
6.7%
0.0%
14.4%
0.0%
5.0%
0.0%
0.0%
7.9%
17.7%
35.1%
20.4%
Mild
HEV
56.5%
53.5%
4.5%
13.9%
0.0%
10.7%
73.7%
0.0%
74.9%
0.0%
0.9%
0.0%
0.0%
0.0%
71.5%
60.9%
18.3%
DSL
0.6%
1.1%
0.0%
0.3%
0.0%
0.0%
0.0%
0.0%
1.2%
0.0%
0.0%
0.0%
0.0%
0.0%
2.4%
0.0%
0.3%
Table 12.34 Absolute Technology Penetrations for the Fleet in the MY2025 Control Case Using RPEs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
5.4%
5.9%
4.5%
5.2%
4.1%
4.1%
14.2%
6.6%
6.6%
6.1%
6.6%
4.5%
0.0%
4.7%
8.1%
7.0%
5.3%
WR
Net
4.3%
4.6%
4.4%
4.9%
4.1%
3.8%
12.1%
6.6%
5.1%
6.1%
6.5%
4.4%
0.0%
4.6%
6.3%
5.9%
4.8%
TDS
18
18.3%
14.1%
36.2%
23.3%
17.4%
4.5%
19.8%
7.6%
17.5%
5.0%
22.8%
6.9%
0.0%
22.3%
15.7%
20.6%
19.7%
TDS
24
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
11
0.9%
0.0%
1.9%
0.3%
6.3%
2.0%
0.0%
0.0%
0.0%
0.0%
0.5%
1.6%
0.0%
10.7%
0.0%
0.0%
2.6%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
5.0%
3.5%
58.0%
31.5%
79.0%
41.5%
0.0%
81.9%
1.3%
68.4%
85.0%
87.7%
0.0%
75.7%
0.6%
0.0%
47.3%
TRX
22
80.2%
92.9%
36.0%
64.9%
3.7%
51.4%
92.1%
0.0%
87.8%
1.9%
7.6%
0.0%
0.0%
8.8%
86.0%
93.8%
42.9%
Deac
69.9%
82.6%
59.9%
72.3%
65.5%
90.8%
71.6%
31.2%
70.3%
77.4%
64.8%
31.0%
0.0%
61.0%
67.7%
73.2%
67.3%
VVLT
18.3%
13.5%
1.2%
6.3%
16.1%
4.2%
20.5%
0.0%
17.5%
0.0%
0.5%
0.0%
0.0%
0.0%
15.5%
20.6%
6.7%
VVT
88.2%
96.7%
98.0%
98.5%
96.1%
97.3%
92.1%
96.5%
87.9%
96.5%
96.5%
96.2%
0.0%
96.9%
83.4%
93.8%
95.4%
CEGR
70.4%
71.9%
51.0%
55.7%
38.1%
71.5%
71.6%
31.2%
71.2%
59.2%
43.8%
31.9%
0.0%
48.8%
71.7%
72.0%
54.7%
Strong
HEV
0.1%
0.0%
1.5%
0.0%
6.4%
2.0%
13.3%
0.0%
0.0%
0.0%
0.9%
1.6%
0.0%
10.1%
0.6%
0.0%
2.6%
EV
7.4%
1.0%
1.0%
0.9%
1.8%
1.3%
6.3%
1.6%
8.6%
2.2%
1.8%
1.7%
100.0%
1.6%
9.6%
4.2%
2.7%
PHEV
3.5%
1.2%
1.3%
1.4%
2.1%
1.5%
2.9%
1.9%
2.7%
1.4%
1.7%
2.0%
0.0%
1.9%
2.0%
2.1%
1.7%
ATK1
0.0%
0.0%
1.5%
0.5%
3.4%
1.1%
0.7%
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
10.2%
0.0%
0.0%
2.1%
ATK2
69.0%
77.0%
51.5%
58.0%
37.5%
83.3%
71.6%
88.9%
70.3%
59.2%
44.0%
21.1%
0.0%
37.7%
67.7%
72.0%
55.8%
Miller
12.3%
0.0%
0.0%
0.0%
0.0%
0.0%
71.6%
0.0%
34.5%
0.0%
0.0%
0.0%
0.0%
0.0%
37.0%
7.3%
3.6%
Stop-
Start
45.6%
31.8%
46.6%
51.2%
3.7%
51.9%
6.7%
1.1%
31.6%
0.0%
21.4%
0.0%
0.0%
16.7%
24.5%
51.4%
30.4%
Mild
HEV
42.1%
54.2%
1.5%
8.2%
0.0%
9.4%
71.4%
0.0%
56.1%
0.0%
0.3%
0.0%
0.0%
0.0%
63.5%
42.4%
15.6%
DSL
1.9%
1.1%
0.0%
0.3%
0.0%
0.0%
0.0%
0.0%
1.2%
0.0%
0.0%
0.0%
0.0%
0.0%
5.3%
0.0%
0.5%
                                            12-29

-------
                                                      EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.35  Incremental Technology Penetrations for Cars in the MY2025 Central Analysis Using ICMs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
2%
1%
2%
2%
1%
0%
4%
0%
1%
1%
1%
0%
0%
1%
1%
1%
1%
WR
Net
3%
1%
1%
1%
1%
0%
2%
0%
-1%
1%
1%
0%
0%
1%
0%
0%
1%
TDS
18
-29%
-26%
11%
-9%
12%
-10%
-51%
0%
-47%
0%
9%
1%
0%
14%
-40%
-64%
-5%
TDS
24
7%
22%
0%
0%
0%
9%
-4%
0%
30%
0%
0%
0%
0%
0%
34%
8%
6%
TRX
11
0%
0%
-19%
-9%
-64%
-1%
0%
-75%
0%
-10%
-26%
-78%
0%
-50%
0%
0%
-23%
TRX
12
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
TRX
21
-56%
-49%
19%
-4%
64%
-19%
-51%
75%
-62%
8%
26%
78%
0%
49%
-60%
-73%
7%
TRX
22
53%
51%
0%
13%
0%
20%
38%
0%
55%
2%
0%
0%
0%
0%
56%
73%
16%
Deac
23%
11%
41%
45%
49%
38%
40%
21%
11%
44%
63%
48%
0%
58%
7%
56%
40%
VVLT
-31%
-9%
0%
-6%
-69%
2%
-50%
0%
-41%
0%
-5%
0%
0%
-1%
-26%
20%
-13%
WT
-2%
0%
0%
0%
0%
0%
-13%
0%
-6%
0%
0%
0%
0%
0%
1%
0%
0%
CEGR
40%
53%
31%
41%
39%
61%
36%
21%
41%
48%
39%
5%
0%
24%
35%
62%
39%
Strong
HEV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
EV
4%
0%
0%
0%
0%
0%
15%
0%
7%
0%
0%
0%
0%
0%
9%
0%
1%
PHEV
0%
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK1
0%
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK2
35%
43%
31%
44%
39%
73%
40%
0%
11%
48%
39%
5%
0%
24%
7%
54%
37%
Miller
0%
0%
0%
0%
0%
0%
60%
0%
8%
0%
0%
0%
0%
0%
16%
0%
1%
Stop-
Start
-31%
36%
0%
9%
0%
24%
-35%
0%
-66%
0%
0%
0%
0%
0%
-46%
46%
1%
Mild
HEV
37%
12%
0%
3%
0%
5%
20%
0%
59%
0%
0%
0%
0%
0%
38%
40%
8%
DSL
-3%
0%
0%
0%
0%
0%
0%
0%
-1%
0%
0%
0%
0%
0%
-11%
0%
-1%
Table 12.36 Incremental Technology Penetrations for Cars in the MY2025 Central Analysis Using RPEs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
0%
1%
0%
1%
0%
0%
2%
0%
0%
0%
0%
0%
0%
0%
1%
0%
0%
WR
Net
0%
0%
0%
1%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
-1%
0%
TDS
18
-33%
-26%
19%
-9%
4%
-10%
-49%
0%
-49%
-7%
18%
1%
0%
14%
-36%
-68%
-4%
TDS
24
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
TRX
11
0%
0%
-14%
-1%
-48%
0%
0%
-53%
0%
-4%
-14%
-77%
0%
-17%
0%
0%
-13%
TRX
12
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
TRX
21
-55%
-60%
17%
-13%
51%
-27%
-49%
53%
-61%
8%
17%
77%
0%
17%
-59%
-73%
-4%
TRX
22
50%
61%
-3%
15%
0%
27%
38%
0%
53%
-4%
-3%
0%
0%
0%
52%
69%
17%
Deac
30%
39%
41%
44%
57%
18%
35%
21%
41%
44%
53%
49%
0%
58%
32%
63%
42%
VVLT
-35%
-8%
0%
-6%
-64%
3%
-43%
0%
-43%
0%
-10%
0%
0%
-1%
-24%
16%
-13%
VVT
-5%
0%
0%
0%
0%
0%
-10%
0%
-8%
0%
0%
0%
0%
0%
-3%
-5%
-1%
CEGR
40%
53%
31%
41%
28%
61%
35%
21%
41%
55%
25%
5%
0%
24%
35%
60%
36%
Strong
HEV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
EV
6%
0%
0%
0%
0%
0%
12%
0%
9%
0%
0%
0%
0%
0%
9%
5%
1%
PHEV
0%
0%
0%
0%
0%
0%
3%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK1
0%
0%
0%
0%
0%
0%
4%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK2
40%
65%
31%
44%
28%
73%
35%
0%
41%
55%
25%
5%
0%
24%
37%
60%
39%
Miller
1%
0%
0%
0%
0%
0%
59%
0%
11%
0%
0%
0%
0%
0%
16%
0%
1%
Stop-
Start
-20%
55%
0%
9%
0%
56%
-35%
0%
-5%
0%
0%
0%
0%
0%
-34%
34%
10%
Mild
HEV
18%
15%
0%
6%
0%
4%
21%
0%
-3%
0%
0%
0%
0%
0%
25%
36%
5%
DSL
-1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-6%
0%
0%
                                             12-30

-------
                                                       EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.37 Incremental Technology Penetrations for Trucks in the MY2025 Central Analysis Using ICMs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
0%
1%
0%
0%
2%
3%
1%
3%
0%
2%
1%
5%
0%
1%
1%
0%
1%
WR
Net
-1%
-1%
0%
-1%
2%
2%
1%
3%
-1%
2%
1%
4%
0%
1%
0%
-1%
1%
TDS
18
-44%
-34%
-38%
-19%
56%
-37%
-45%
28%
-41%
10%
-16%
3%
0%
-22%
-40%
-48%
-18%
TDS
24
8%
22%
8%
14%
0%
27%
0%
0%
-12%
0%
16%
0%
0%
22%
-8%
10%
12%
TRX
11
0%
0%
-9%
0%
-36%
0%
0%
-19%
0%
0%
-16%
-90%
0%
-16%
0%
0%
-14%
TRX
12
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
TRX
21
-71%
-78%
-41%
-41%
28%
-79%
-70%
19%
-70%
20%
-6%
90%
0%
0%
-69%
-76%
-28%
TRX
22
71%
79%
50%
41%
8%
79%
70%
0%
70%
-20%
23%
0%
0%
15%
69%
76%
42%
Deac
38%
14%
63%
11%
-13%
10%
45%
52%
58%
88%
42%
0%
0%
45%
55%
38%
28%
VVLT
-33%
26%
7%
6%
-98%
12%
-45%
0%
-54%
-53%
-1%
0%
0%
0%
-49%
31%
-6%
VVT
2%
1%
0%
0%
0%
0%
0%
0%
5%
10%
0%
0%
0%
0%
6%
0%
0%
CEGR
44%
55%
63%
71%
26%
62%
45%
52%
41%
66%
48%
0%
0%
53%
40%
48%
51%
Strong
HEV
0%
0%
0%
0%
0%
0%
3%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
EV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
PHEV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK1
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK2
38%
30%
56%
58%
26%
46%
45%
0%
58%
66%
35%
0%
0%
36%
55%
38%
40%
Miller
45%
0%
0%
0%
0%
0%
45%
0%
58%
0%
0%
0%
0%
0%
55%
34%
5%
Stop-
Start
-35%
-11%
64%
65%
0%
-2%
-30%
0%
-30%
0%
13%
0%
0%
17%
-30%
-23%
19%
Mild
HEV
35%
71%
8%
24%
0%
52%
27%
0%
30%
0%
2%
0%
0%
0%
30%
61%
23%
DSL
-2%
-1%
0%
0%
0%
0%
0%
0%
-5%
0%
0%
0%
0%
0%
-6%
0%
0%
Table 12.38 Incremental Technology Penetrations for Trucks in the MY2025 Central Analysis Using RPEs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
0%
0%
0%
0%
0%
1%
2%
1%
0%
2%
1%
0%
0%
1%
0%
0%
0%
WR
Net
-1%
-1%
0%
0%
0%
-1%
1%
1%
0%
2%
1%
0%
0%
1%
0%
-1%
0%
TDS
18
-41%
-31%
-33%
-22%
33%
-25%
-45%
25%
-42%
10%
-11%
3%
0%
26%
-40%
-47%
-13%
TDS
24
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
TRX
11
0%
0%
-3%
0%
-20%
0%
0%
-19%
0%
0%
-4%
-86%
0%
-5%
0%
0%
-9%
TRX
12
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
TRX
21
-77%
-77%
-42%
-78%
12%
-79%
-70%
19%
-70%
20%
1%
86%
0%
0%
-69%
-80%
-35%
TRX
22
77%
78%
45%
78%
8%
79%
70%
0%
70%
-20%
3%
0%
0%
5%
69%
80%
44%
Deac
46%
35%
74%
27%
9%
25%
45%
55%
47%
86%
82%
26%
0%
64%
47%
47%
46%
VVLT
-36%
12%
2%
6%
-98%
12%
-45%
0%
-42%
-51%
-1%
0%
0%
0%
-40%
1%
-9%
WT
4%
1%
0%
0%
0%
0%
0%
0%
5%
7%
0%
0%
0%
0%
6%
0%
1%
CEGR
41%
50%
66%
68%
48%
49%
45%
55%
42%
66%
72%
26%
0%
50%
40%
47%
55%
Strong
HEV
0%
0%
0%
0%
0%
0%
5%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
EV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
PHEV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK1
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK2
46%
54%
66%
70%
48%
61%
45%
4%
47%
66%
73%
26%
0%
55%
47%
47%
56%
Miller
35%
0%
0%
0%
0%
0%
45%
0%
65%
0%
0%
0%
0%
0%
58%
14%
5%
Stop-
Start
-62%
-10%
80%
76%
8%
10%
-30%
4%
-30%
0%
54%
0%
0%
37%
-30%
5%
30%
Mild
HEV
64%
70%
3%
10%
0%
52%
25%
0%
30%
0%
1%
0%
0%
0%
30%
29%
19%
DSL
-4%
-1%
0%
0%
0%
0%
0%
0%
-5%
0%
0%
0%
0%
0%
-6%
0%
0%
                                              12-31

-------
                                                        EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.39 Incremental Technology Penetrations for the Fleet in the MY2025 Central Analysis Using ICMs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
1%
1%
1%
1%
1%
1%
2%
1%
1%
2%
1%
4%
0%
1%
1%
1%
1%
WR
Net
2%
0%
1%
0%
1%
0%
1%
1%
-1%
2%
1%
3%
0%
1%
0%
0%
1%
TDS
18
-33%
-31%
-17%
-14%
32%
-13%
-46%
9%
-45%
4%
-1%
3%
0%
-2%
-40%
-56%
-11%
TDS
24
7%
22%
5%
7%
0%
11%
-1%
0%
14%
0%
6%
0%
0%
10%
18%
9%
9%
TRX
11
0%
0%
-13%
-4%
-51%
-1%
0%
-58%
0%
-7%
-22%
-87%
0%
-34%
0%
0%
-19%
TRX
12
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
TRX
21
-59%
-70%
-16%
-23%
47%
-26%
-66%
58%
-65%
12%
13%
87%
0%
27%
-64%
-75%
-10%
TRX
22
57%
71%
29%
27%
4%
27%
63%
0%
61%
-6%
10%
0%
0%
7%
61%
75%
28%
Deac
27%
13%
54%
28%
20%
34%
44%
30%
29%
60%
55%
11%
0%
52%
25%
47%
34%
VVLT
-31%
16%
4%
0%
-83%
3%
-46%
0%
-46%
-19%
-3%
0%
0%
-1%
-35%
26%
-10%
VVT
-1%
0%
0%
0%
0%
0%
-3%
0%
-2%
4%
0%
0%
0%
0%
3%
0%
0%
CEGR
41%
54%
49%
56%
33%
61%
43%
30%
41%
55%
42%
1%
0%
37%
37%
55%
44%
Strong
HEV
0%
0%
0%
0%
0%
0%
2%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
EV
3%
0%
0%
0%
0%
0%
3%
0%
4%
0%
0%
0%
0%
0%
6%
0%
1%
PHEV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK1
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK2
35%
34%
45%
51%
33%
69%
44%
0%
29%
55%
37%
1%
0%
29%
25%
46%
38%
Miller
11%
0%
0%
0%
0%
0%
48%
0%
27%
0%
0%
0%
0%
0%
30%
17%
3%
Stop-
Start
-32%
3%
37%
37%
0%
21%
-31%
0%
-52%
0%
5%
0%
0%
8%
-40%
10%
10%
Mild
HEV
36%
53%
5%
14%
0%
11%
26%
0%
48%
0%
1%
0%
0%
0%
35%
51%
15%
DSL
-3%
0%
0%
0%
0%
0%
0%
0%
-2%
0%
0%
0%
0%
0%
-9%
0%
-1%
Table 12.40 Incremental Technology Penetrations for the Fleet in the MY2025 Central Analysis Using RPEs

BMW
FCA
FORD
GM
HONDA
HYUNDAI/KIA
JLR
MAZDA
MERCEDES-BENZ
MITSUBISHI
NISSAN
SUBARU
TESLA
TOYOTA
VOLKSWAGEN
VOLVO
Fleet
WR
Tech
0%
0%
0%
1%
0%
0%
2%
0%
0%
1%
0%
0%
0%
1%
0%
0%
0%
WR
Net
0%
-1%
0%
0%
0%
0%
1%
0%
0%
1%
0%
0%
0%
1%
0%
-1%
0%
TDS
18
-35%
-29%
-11%
-15%
17%
-12%
-46%
8%
-46%
-1%
7%
3%
0%
20%
-38%
-57%
-8%
TDS
24
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
TRX
11
0%
0%
-8%
-1%
-35%
0%
0%
-43%
0%
-2%
-10%
-84%
0%
-12%
0%
0%
-11%
TRX
12
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
TRX
21
-60%
-72%
-17%
-45%
32%
-33%
-66%
43%
-65%
12%
11%
84%
0%
10%
-63%
-77%
-19%
TRX
22
57%
73%
25%
46%
4%
33%
63%
0%
59%
-10%
-1%
0%
0%
2%
59%
75%
30%
Deac
34%
36%
60%
36%
35%
19%
43%
31%
43%
59%
65%
31%
0%
61%
38%
55%
43%
VVLT
-35%
6%
1%
0%
-80%
4%
-45%
0%
-43%
-18%
-6%
0%
0%
-1%
-30%
8%
-11%
WT
-3%
0%
0%
0%
0%
0%
-2%
0%
-3%
3%
0%
0%
0%
0%
0%
-2%
0%
CEGR
40%
51%
51%
54%
38%
60%
43%
31%
41%
59%
44%
21%
0%
36%
37%
53%
45%
Strong
HEV
0%
0%
0%
0%
0%
0%
4%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
EV
5%
0%
0%
0%
0%
0%
3%
0%
6%
0%
0%
0%
0%
0%
6%
2%
1%
PHEV
0%
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK1
0%
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
ATK2
41%
57%
51%
57%
38%
72%
43%
1%
43%
59%
44%
21%
0%
38%
41%
53%
47%
Miller
9%
0%
0%
0%
0%
0%
48%
0%
32%
0%
0%
0%
0%
0%
32%
7%
3%
Stop-
Start
-30%
10%
47%
42%
4%
51%
-31%
1%
-15%
0%
21%
0%
0%
17%
-32%
19%
20%
Mild
HEV
29%
53%
2%
8%
0%
9%
24%
0%
9%
0%
0%
0%
0%
0%
27%
32%
12%
DSL
-2%
0%
0%
0%
0%
0%
0%
0%
-2%
0%
0%
0%
0%
0%
-6%
0%
0%
                                               12-32

-------
                                                EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.41 Summary of Absolute Technology Penetrations in the MY2025 Control Case
Indirect Cost
Approach
ICM
ICM
ICM
RPE
RPE
RPE
C/T/Fleet
C
T
Fleet
C
T
Fleet
WR
Tech
6.3%
6.9%
6.6%
4.5%
6.1%
5.3%
WR
Net
5.8%
6.2%
6.0%
4.2%
5.6%
4.8%
IDS
18
20.5%
24.5%
22.4%
19.3%
20.1%
19.7%
IDS
24
7.6%
14.8%
11.0%
0.0%
0.0%
0.0%
TRX
11
4.7%
0.3%
2.6%
4.7%
0.3%
2.6%
TRX
12
0.0%
0.0%
0.0%
0.0%
0.0%
0.0%
TRX
21
58.9%
43.4%
51.5%
55.7%
38.0%
47.3%
TRX
22
24.2%
54.7%
38.7%
27.1%
60.1%
42.9%
Deac
54.3%
52.8%
53.6%
61.4%
73.8%
67.3%
VVLT
6.2%
10.5%
8.3%
6.5%
7.0%
6.7%
VVT
93.1%
98.5%
95.7%
92.5%
98.5%
95.4%
CEGR
48.4%
58.9%
53.4%
46.1%
64.2%
54.7%
Strong
HEV
4.5%
0.5%
2.6%
4.5%
0.6%
2.6%
EV
4.6%
0.4%
2.6%
4.8%
0.4%
2.7%
PHEV
2.7%
0.5%
1.7%
2.7%
0.5%
1.7%
ATK1
4.0%
0.0%
2.1%
4.0%
0.0%
2.1%
ATK2
43.9%
45.0%
44.4%
47.6%
64.8%
55.8%
Miller
1.4%
5.9%
3.6%
1.5%
5.8%
3.6%
Stop-
Start
12.1%
29.5%
20.4%
19.9%
41.9%
30.4%
Mild
HEV
10.4%
26.9%
18.3%
8.4%
23.4%
15.6%
DSL
0.2%
0.5%
0.3%
0.5%
0.5%
0.5%
Table 12.42 Summary of Incremental Technology Penetrations in the MY2025 Control Case
Indirect Cost Approach
ICM
ICM
ICM
RPE
RPE
RPE
C/T/Fleet
C
T
Fleet
C
T
Fleet
WR
Tech
1%
1%
1%
0%
0%
0%
WR
Net
1%
1%
1%
0%
0%
0%
TDS
18
-5%
-18%
-11%
-4%
-13%
-8%
TDS
24
6%
12%
9%
0%
0%
0%
TRX
11
-23%
-14%
-19%
-13%
-9%
-11%
TRX
12
0%
0%
0%
0%
0%
0%
TRX
21
7%
-28%
-10%
-4%
-35%
-19%
TRX
22
16%
42%
28%
17%
44%
30%
Deac
40%
28%
34%
42%
46%
43%
WLT
-13%
-6%
-10%
-13%
-9%
-11%
VVT
0%
0%
0%
-1%
1%
0%
CEGR
39%
51%
44%
36%
55%
45%
Strong
HEV
0%
0%
0%
0%
0%
0%
EV
1%
0%
1%
1%
0%
1%
PHEV
0%
0%
0%
0%
0%
0%
ATK1
0%
0%
0%
0%
0%
0%
ATK2
37%
40%
38%
39%
56%
47%
Miller
1%
5%
3%
1%
5%
3%
Stop-
Start
1%
19%
10%
10%
30%
20%
Mild
HEV
8%
23%
15%
5%
19%
12%
DSL
-1%
0%
-1%
0%
0%
0%
                                       12-33

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
   Not readily apparent in the technology penetration tables above is the number, or percentage,
of vehicles that receive specific levels of mass reduction. Table 12.43 below provides more
detail on mass reduction technology using the same approach as described in the text
accompanying Table 12.28.

 Table 12.43 Percentage of Vehicles Receiving the Mass Reduction Levels within the Indicated Ranges in the
               MY2025 Control Case Using ICMs and AEO Reference Case Fuel Prices
Fleet
Including ZEV Program Vehicles
Excluding ZEV Program Vehicles
%MR Range
<=5%
5% to <=10%
10% to <=15%
15% to <=20%
<=5%
5% to <=10%
10% to <=15%
15% to <=20%
Baseline
87.0%
9.1%
0.9%
3.0%
89.3%
9.3%
0.9%
0.5%
WRtech
57.4%
30.7%
7.3%
4.6%
58.9%
31.5%
7.5%
2.1%
WRnet
61.0%
28.0%
8.3%
2.7%
62.6%
28.8%
8.5%
0.2%
12.1.1.4
Comparisons to the 2012 Final Rule
   Of interest is how the costs estimated in this Draft TAR analysis compare to those presented
in the 2012 FRM. In that analysis, since we were setting standards for MY2017-2025, we did not
present costs relative to a reference case consisting of the MY2021 standards.  Instead, we
presented costs relative to a reference case consisting of the MY2016 standards. In Table 12.44
we have broken out the Draft TAR costs/vehicle along with the closest matching costs/vehicle
from the 2012 FRM. The entries of perhaps most interest are those shown for the incremental
costs to bring the fleet down to the 2025 standards, shown as $1070 for the 2012 FRM and $894
for the Draft TAR. Because the baseline fleets are completely different, comparisons of the costs
to bring the baseline fleets down to the 2016 standards are not valid comparisons.  Instead, the
relative values of these entries simply show that the 2014 fleet is nearly complying with the 2016
standards, as one would expect.  The same is true for the bottom row showing total costs. The
costs to bring the 2008 fleet, projected forward to MY2025, into compliance with the 2025
standards should be considerably higher than the costs to bring the 2014 fleet, projected forward
to MY2025, into compliance with those standards. This is reflected in the bottom-row values in
the table. The differences in the costs to bring the respective baseline fleets down to the each
incrementally lower standard level are driven by many factors including, but not limited to:
car/truck fleet mix and footprint characteristics are more favorable to lower costs because of the
relatively larger number of car-like trucks that emit more like cars but are actually subject to the
less stringent truck curve; new and very cost effective technologies like Atkinson 2 and mild
hybrid 48V technologies that were not even considered in the 2012 FRM; updated and more
comprehensive studies informing our mass reduction cost estimates; inclusion of ZEV required
EV and PHEV sales which was not considered for the FRM.  To better understand the impact
some of these factors have on the overall analysis, EPA has also performed several sensitivity
analyses which are described below in Chapter 12.2.4.
                                             12-34

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
          Table 12.44 Cost per Vehicle Comparison - 2012 FRM (2010$) vs Draft TAR (2013$)
Note: Due to large differences in the
baseline fleets used (2008 vs. 2014),
the 2012 FRM values and the Draft TAR
results are not directly comparable.
Cost to bring the baseline fleet down to
the 2016 standards
Incremental cost to bring that fleet
down to the 2021 standards
Incremental cost to bring that baseline
fleet down to the 2025 standards
Total costs to bring the baseline fleet
down to the 2025 standards
FRM
(2008 baseline fleet in
MY2025)
$719
$766
$1070
$2555
Draft TAR
(2014 baseline fleet in
MY2025)
$279
$393
$894
$1565
Note: The $719 value can be found in EPA's final RIA (EPA-420-R-12-016) at Table 3.6-1; the $766 value can be
found in EPA's final RIA at Table 3.6-2 and is actually a MY2021 cost presented here as a proxy for a MY2025
cost; the $1070 value is calculated as $2555 (see final RIA Table 7.4-5) minus $766 minus $719; the $393 value is
calculated as $671 (see Table 12.97, "Reference Case in MY2025" entry for the Combined Fleet) minus $279; the
$894 value can be found in Table 12.17 and the $1565 value can be found in Table 12.97, "Control Case in
MY2025" entry for the Combined Fleet.

   We can also consider the technology penetration rate differences between the 2012 FRM and
this Draft TAR.  Here we focus only on the final, absolute technology penetrations projected in
the 2012 FRM and those projected in this Draft TAR in the ICM-based central analysis. The
absolute technology penetrations for the technologies generally considered to be of most interest
are shown in the table below.
           Table 12.45 Final Technology Penetration Comparison - 2012 FRM vs Draft TAR
Technology
Gasoline direct injection engine
8+ speeds & improved CVTs
Turbocharged and downsized gasoline
engine
Higher compression ratio/naturally
aspirated gasoline engine (Atkinson-2)
Stop-start
Mild HEV
Strong HEV
EV+PHEV
2012 FRM
94%
91%
93%
n/a
15%
26%
5%
2%
Draft TAR
79%
90%
33%
44%
20%
18%
3%
4%
Note: 2012 FRM values taken from EPA's final RIA Table 3.5-25; Atkinson-2 was not considered in the 2012 FRM;
mild HEV used a 110/115V battery in the 2012 FRM but uses a 48V battery in this Draft TAR.
   This table highlights two important results: (1) EPA's 2012 FRM analysis featured a high
penetration of turbocharged/downsized engine technology, a technology that is projected less in
EPA's Draft TAR analysis due to the inclusion of the new and more cost-effective Atkinson-2
technology which provides dual non-electrified pathways toward compliance with the MY2022-
2025 standards (both turbocharging/downsizing and Atkinson-2); and, (2) just two years into the
2012-2016 GHG program, a new technology—Atkinson-2—which was not previously
considered by the agencies, has emerged as one of the most promising non-electrified
technologies capable of playing a major role in compliance with the standards through 2025.
Further, while not as highly projected as Atkinson-2 in our analysis, the mild HEV 48V
                                               12-35

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
technology represents yet another cost effective technology that can provide another pathway
toward compliance. EPA has confidence that other technologies will emerge in the coming years
and we will consider further developments as the midterm evaluation progresses.

12.1.2 Sensitivity Analysis Results
12.1.2.1
Reference Case:  CO2 Targets
   The different AEO 2015 fuel price cases (shown in Chapter 3, Figure 3.3) carry with them
unique fleet projections since higher fuel prices are projected to result in fewer truck and more
car sales, while lower fuel prices are projected to result in fewer car sales and more truck sales.
As a result of these fleet mix differences, the manufacturer-specific footprint based standards
would result in different fleet-wide CCh target values for each AEO 2015 fuel price case and
projected fleet. While we have conducted additional sensitivity runs beyond varying the fuel
price projections, only these two fuel price sensitivities (high and low) result in unique CCh
target values.  All other sensitivity runs use the AEO 2015 reference case fuel prices, fleets and
resultant targets.
          Table 12.46  Reference Case CCh Targets in MY2025 for Each Sensitivity Case (g/mi)

Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Car
AEO
Ref
177.5
182.3
179.6
178.8
172.8
177.1
189.7
175.2
180.0
164.8
173.3
170.0
205.7
174.5
174.6
182.0
176.9
AEO
High
177.2
182.2
179.4
178.6
172.6
176.9
189.8
175.2
179.9
164.8
173.2
169.7
205.7
174.4
174.5
182.0
176.8
AEO
Low
177.5
182.3
179.7
178.9
172.9
177.1
189.6
175.2
180.0
164.8
173.3
170.1
205.7
174.6
174.6
182.0
177.0
Truck
AEO
Ref
237.0
247.2
280.0
277.3
232.9
227.9
235.0
223.4
237.0
208.4
243.0
210.5
0.0
246.3
230.4
227.7
251.3
AEO
High
236.8
246.2
278.3
276.4
232.4
227.9
234.7
223.1
236.8
208.3
242.1
210.4
0.0
245.1
230.3
227.7
249.9
AEO
Low
237.0
247.5
280.6
277.6
233.0
227.8
235.0
223.5
237.0
208.4
243.2
210.5
0.0
246.7
230.5
227.7
251.7
Combined
AEO
Ref
191.7
227.9
237.9
227.9
200.8
183.1
225.5
190.0
201.7
180.4
200.9
201.4
205.7
207.0
195.7
205.8
212.4
AEO
High
187.9
221.9
227.2
217.7
195.2
181.4
222.6
186.5
196.9
176.8
194.5
198.4
205.7
199.9
191.1
201.9
204.9
AEO
Low
193.6
230.3
242.5
232.4
203.4
184.1
226.6
191.7
204.0
182.1
203.8
202.5
205.7
210.2
197.8
207.6
215.7
                                              12-36

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
12.1.2.2       Control Case: CO2 Targets

           Table 12.47 Control Case CCh Targets in MY2025 for Each Sensitivity Case (g/mi)

Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Car
AEO
Ref
147.7
151.8
149.5
148.8
143.7
147.3
158.1
145.8
149.8
137.0
144.1
141.3
171.6
145.2
145.2
151.5
147.2
AEO
High
147.5
151.7
149.3
148.6
143.6
147.2
158.2
145.7
149.7
136.9
144.0
141.1
171.6
145.1
145.2
151.5
147.1
AEO
Low
147.7
151.8
149.6
148.9
143.8
147.4
158.0
145.8
149.8
137.0
144.2
141.4
171.6
145.2
145.2
151.6
147.3
Truck
AEO
Ref
193.8
202.3
229.7
227.4
190.4
186.2
192.1
182.4
193.8
169.9
198.8
171.7
0.0
201.5
188.3
186.1
205.7
AEO
High
193.7
201.5
228.2
226.6
190.0
186.2
191.9
182.2
193.6
169.9
198.1
171.6
0.0
200.6
188.2
186.0
204.6
AEO
Low
193.8
202.6
230.1
227.6
190.5
186.1
192.1
182.5
193.8
169.9
199.0
171.7
0.0
201.9
188.4
186.1
206.1
Combined
AEO
Ref
158.7
187.3
196.1
188.0
165.5
152.0
185.0
157.0
166.6
148.8
165.8
164.9
171.6
170.7
161.5
169.5
175.1
AEO
High
155.7
182.6
187.5
179.9
161.1
150.6
182.8
154.4
162.9
146.0
160.8
162.7
171.6
165.1
158.0
166.5
169.2
AEO
Low
160.2
189.2
199.7
191.6
167.5
152.7
185.9
158.3
168.4
150.1
168.1
165.7
171.6
173.3
163.2
170.9
177.8
   Note that none of the total fleet targets presented in Table 12.47 achieve the 163 g/mi CCh
target (54.5 mpg, if all reductions achieved through fuel economy improvements) projected in
the 2012 FRM. This is due to changes in the fleet makeup, mainly-car/truck mix and also
footprint characteristics in the AEO 2015 fleet projections relative to the 2012 FRM projections.
12.1.2.3
Cost per Vehicle and Technology Penetrations
   In the previous section, EPA presented our projections for the technology penetrations and
cost per vehicle for the MY2025 central analysis control case. We recognize there are many
uncertainties involved when making projections to MY2025, including the makeup of the future
fleet, which will be influenced in part by future gasoline prices, which technologies
manufacturers will actually adopt, how manufacturers will respond to compliance with the
standards given the range of credit programs available, including credit trading across
manufacturers. As a way to inform how changes in such factors would affect our analysis of the
MY2025 standards, we have conducted a wide range of sensitivity analyses, including:

       1) AEO 2015 high fuel price case, which changes both fuel prices and projected fleet
          characteristics (using both ICMs and RPEs).
       2) AEO 2015 low fuel price case, which changes both fuel prices and  projected fleet
          characteristics (using both ICMs and RPEs).
       3) "Perfect" credit trading across all manufacturers.  This sensitivity should represent the
          most cost effective case since any manufacturer in need of credits is assumed to
          acquire them if they exist (using ICMs).
       4) No Car/Truck transfers across a single manufacturer's fleet, which forces cars to meet
          the car curve standards and trucks to meet the truck curve standards (using ICMs).
                                              12-37

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
          This sensitivity illustrates a more restrictive scenario, since the GHG program in fact
          allows full transfers across a manufacturer's car and truck fleets, and thus highlights
          the importance of this flexibility provision.
       5)  No additional mass reduction beyond that included in the projected baseline fleet.
          That is, no mass reduction allowed to comply with MY2021  or MY2025 standards.
          Though EPA believes our mass reduction estimates are fully feasible, this sensitivity
          shows the impacts of our updated mass reduction costs on the results (using ICMs).A
          non-Atkinson engine technology path which sets a penetration cap on Atkinson-2
          technology at 10 percent in both the reference and control cases. This sensitivity
          shows the impacts of manufacturers choosing a path less dependent on that
          technology (using ICMs).
Table 12.48 MY2025 Absolute Technology Penetrations & Incremental Costs for Cars in Each OMEGA Run
                                         (2013$)
Technology
VVT
VVLT
Deac
TRX11
TRX12
TRX21
TRX22
TDS18
TDS24
ATK2
Cooled EGR
Miller
Stop-Start
Mild Hybrid
Full Hybrid
REEV
EV
WRtech
WRnet
$/vehicle
AEO
Ref
ICM
93%
6%
54%
5%
0%
59%
24%
20%
8%
44%
48%
1%
12%
10%
4%
3%
5%
6%
6%
$707
AEO
High
ICM
93%
6%
57%
5%
0%
60%
23%
18%
7%
47%
51%
1%
19%
8%
5%
2%
5%
6%
5%
$701
AEO
Low
ICM
93%
6%
54%
5%
0%
61%
22%
20%
8%
43%
48%
1%
16%
7%
4%
3%
5%
6%
6%
$707
AEO
Ref
RPE
93%
6%
61%
5%
0%
56%
27%
19%
0%
48%
46%
2%
20%
8%
4%
3%
5%
4%
4%
$789
AEO
High
RPE
93%
6%
65%
5%
0%
60%
23%
16%
0%
52%
49%
1%
25%
7%
5%
2%
5%
4%
4%
$778
AEO
Low
RPE
92%
7%
60%
5%
0%
61%
21%
19%
0%
47%
44%
1%
22%
7%
4%
3%
5%
5%
4%
$782
Perfect
Trading
ICM
93%
6%
54%
5%
0%
78%
6%
29%
3%
44%
49%
0%
7%
0%
4%
3%
3%
6%
5%
$549
NoC/T
Transfers
ICM
93%
10%
63%
5%
0%
42%
42%
12%
8%
55%
57%
0%
29%
16%
4%
3%
4%
7%
6%
$775
No
Additional
MR
Beyond
Baseline
Levels
ICM
93%
8%
56%
5%
0%
55%
28%
18%
9%
49%
55%
2%
16%
13%
4%
3%
5%
3%
3%
$709
Non-
ATK2
Path
ICM
93%
13%
42%
5%
0%
39%
44%
14%
24%
10%
35%
0%
14%
20%
4%
3%
5%
7%
6%
$828
                                              12-38

-------
                                    EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.49 MY2025 Absolute Technology Penetrations & Incremental Costs for Trucks in Each OMEGA
                                      Run (2013$)
Technology
VVT
VVLT
Deac
TRX11
TRX12
TRX21
TRX22
TDS18
TDS24
ATK2
Cooled EGR
Miller
Stop-Start
Mild Hybrid
Full Hybrid
REEV
EV
WRtech
WRnet
$/vehicle
AEO
Ref
ICM
99%
11%
53%
0%
0%
43%
55%
25%
15%
45%
59%
6%
30%
27%
1%
1%
0%
7%
6%
$1099
AEO
High
ICM
99%
15%
53%
0%
0%
37%
61%
23%
18%
45%
62%
8%
26%
30%
0%
0%
1%
7%
6%
$1144
AEO
Low
ICM
98%
13%
51%
0%
0%
44%
54%
24%
17%
42%
58%
6%
28%
25%
1%
1%
1%
7%
6%
$1077
AEO
Ref
RPE
99%
7%
74%
0%
0%
38%
60%
20%
0%
65%
64%
6%
42%
23%
1%
1%
0%
6%
6%
$1267
AEO
High
RPE
99%
8%
74%
0%
0%
36%
62%
19%
2%
66%
68%
6%
37%
28%
1%
0%
1%
6%
6%
$1304
AEO
Low
RPE
98%
7%
73%
0%
0%
42%
56%
20%
0%
63%
62%
6%
40%
23%
1%
1%
1%
6%
6%
$1251
Perfect
Trading
ICM
98%
11%
66%
0%
0%
35%
63%
14%
17%
56%
69%
0%
48%
26%
0%
1%
0%
7%
6%
$1211
NoC/T
Transfers
ICM
99%
7%
51%
0%
0%
49%
49%
23%
18%
36%
53%
6%
24%
22%
2%
1%
0%
7%
6%
$1086
No
Additional
MR Beyond
Baseline
Levels
ICM
98%
14%
59%
0%
0%
31%
67%
16%
21%
52%
71%
8%
30%
33%
1%
1%
1%
3%
2%
$1137
Non-
ATK2
Path
ICM
98%
27%
35%
0%
0%
16%
82%
15%
42%
10%
52%
1%
17%
57%
1%
1%
1%
7%
6%
$1269
                                             12-39

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.50 MY2025 Absolute Technology Penetrations & Incremental Costs for the Fleet in Each OMEGA
                                        Run (2013$)
Technology
VVT
VVLT
Deac
TRX11
TRX12
TRX21
TRX22
TDS18
TDS24
ATK2
Cooled EGR
Miller
Stop-Start
Mild Hybrid
Full Hybrid
REEV
EV
WRtech
WRnet
$/vehicle
AEO
Ref
ICM
96%
8%
54%
3%
0%
52%
39%
22%
11%
44%
53%
4%
20%
18%
3%
2%
3%
7%
6%
$894
AEO
High
ICM
95%
9%
55%
3%
0%
51%
38%
20%
11%
46%
55%
4%
22%
16%
3%
2%
3%
6%
6%
$872
AEO
Low
ICM
96%
10%
52%
2%
0%
52%
38%
22%
13%
42%
53%
4%
23%
16%
2%
2%
3%
6%
6%
$899
AEO
Ref
RPE
95%
7%
67%
3%
0%
47%
43%
20%
0%
56%
55%
4%
30%
16%
3%
2%
3%
5%
5%
$1017
AEO
High
RPE
95%
7%
69%
3%
0%
51%
38%
17%
1%
58%
56%
3%
29%
15%
3%
2%
3%
5%
5%
$980
AEO
Low
RPE
95%
7%
67%
2%
0%
51%
39%
20%
0%
55%
53%
4%
31%
15%
2%
2%
3%
5%
5%
$1025
Perfect
Trading
ICM
95%
8%
60%
3%
0%
58%
33%
22%
10%
50%
59%
0%
27%
13%
3%
2%
2%
6%
6%
$865
NoC/T
Transfers
ICM
96%
8%
57%
3%
0%
45%
45%
17%
13%
46%
55%
3%
26%
19%
3%
2%
2%
7%
6%
$923
No
Additional
MR Beyond
Baseline
Levels
ICM
96%
11%
57%
3%
0%
44%
46%
17%
15%
50%
62%
4%
23%
22%
3%
2%
3%
3%
2%
$913
Non-
ATK2
Path
ICM
95%
19%
38%
3%
0%
28%
62%
15%
32%
10%
43%
1%
15%
38%
3%
2%
3%
7%
6%
$1038
12.1.2.4
Observations on Sensitivity Analyses
   EPA notes the following observations on each of the sensitivity analyses shown above.
   1. Fuel prices have little impact on the cost per vehicle outcomes. This result is driven by the
fact that the projected fleet changes depending on the projected fuel price. The AEO 2015 high
fuel price case has more cars than the reference price case, while the low fuel price case has
more trucks than the reference price case. This observation holds true within the ICM fuel price
cases and within the RPE fuel price cases.

   2. Fuel prices have little impact on the technology penetration outcomes.  Within the ICM fuel
price cases, the technology penetrations vary only slightly. The same is true with the RPE fuel
price cases.

   3. Higher fuel prices do not result in substantially different fleet electrification.  Full electric
and plug-in hybrid electric vehicle penetrations are essentially constant across all sensitivities.
This is largely driven by the EVs and PHEVs projected in the reference fleet as a result of the
ZEV program.  Only the mild hybrid technology shows notable differences, ranging from 13
percent to 38 percent of the fleet depending on the sensitivity case.
                                              12-40

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
   4. Using RPEs to account for indirect costs increases $/vehicle, as would be expected, but
only on the order of $100 to $125 per vehicle, depending on fuel price case.

   5. The $/vehicle result is not heavily dependent on mass reduction  and, therefore, the mass
reduction cost curves. Disallowing any mass reduction beyond that estimated in the baseline
fleet increased $/vehicle by just $19 ($2 per car, $38 per truck, $19 combined).  There are
enough technologies available with similar cost effectiveness such that the fleet compliance costs
are not dependent on any one of those technologies.

   6. Limiting estimated penetration of the Atkinson-2 engine technology would increase
estimated cost per vehicle from $894 to $1,038, a $144 increase.

   7. While the case where car/truck transfers has little impact on overall $/vehicle, the limitation
of transfers impacts car costs more significantly increasing their costs from $707 to $775 (+$68)
while decreasing truck costs from $1099 to $1086 (-$13).  This indicates that, in the central
analysis, it is more cost effective to reduce truck emissions (as discussed in Section 12.1.1.4 and
in observation 8 below) and transfer over compliance credits to the less cost effective car fleet.
This can also be seen in Table 12.3  and Table 12.4 which show achieved car CCh higher than
respective targets and achieved truck CCh lower than respective targets. Elimination of transfers
also drives the car fleet further into the advanced technologies (TRX22, ATK2, stop-start, mild
HEV) while simultaneously limiting advanced technology penetrations on trucks.

   8. The perfect trading sensitivity illustrates the potential value of trading across firms and
illustrates the greater value of truck credits given the higher VMT of trucks when determining
the credit. The overall $/vehicle impact is not great ($894 down to $865), but the car $/vehicle
decreases from $707 down to $549  (-$158) while the truck $/vehicle increases from $1099 to
$1211 (+$112). OMEGA is putting more technology on trucks to generate credits that can be
used to offset under compliance (and less technology) on cars.  This also illustrates the
movement of the fleet to car-like trucks that emit at levels  more like cars and  have car-like (i.e.,
generally less costly) technologies for use in reducing CCh emissions  but are  on the less stringent
truck curve.  Those car-like trucks can cost effectively generate credits that can then be traded to
another firm.

12.1.3 Payback Period & Lifetime Savings

   Here EPA looks at the cost of owning a new vehicle complying with the MY2025 standards
and the payback period - the point at which savings exceed costs.  For example, relative to the
reference case (i.e., the MY2021 standards), a new MY2025 vehicle is estimated to cost roughly
$900 to $1,000 more due to the addition of new GHG reducing/fuel economy improving
technology.  This new technology will result in lower fuel  consumption and, therefore, savings in
fuel expenditures. But how many months or years would pass before  the fuel savings  exceed the
cumulative costs?

   The tables below present EPA's  estimates of increased costs associated with owning a new
MY2025 vehicle. For purposes of this analysis, we are using a "sales weighted average vehicle"
which means the combined car/truck fleet, weighted by sales on the cost side  and usage on the
fuel savings side, to arrive at a  single weighted vehicle analysis. The  table uses results from the
OMEGA Inventory, Costs and  Benefits Tool  analysis discussed in the section 12.2. Included in
the analysis are maintenance costs (see Chapter 5.3.2.3), sales taxes and insurance costs (see
                                              12-41

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
Chapter 10).  This analysis does not include other impacts such as reduced refueling events, or
other societal impacts, such as the potential rebound miles driven or the value of driving those
rebound miles, or noise, congestion and accidents, since the focus is meant to be on those factors
consumers likely think about most while in the showroom considering a new car purchase, and
on those factors that result in more or fewer dollars in their pockets. As noted, to estimate the
cumulative vehicle costs, we have included not only the sales tax on the new car purchase but
also the increased insurance  premiums that would result from the more valuable vehicle (see
Chapter 10).  The payback periods were calculated using both 3 percent and 7 percent discount
rates with lifetime discounted costs shown in the last 2 rows of the table, again at both 3 percent
and 7 percent discount rates.

   As shown in these tables, payback occurs in the 5th year of ownership in the ICM case and the
6th year in the RPE case, regardless of the discount rate used. Note that, in the first table, the cost
per vehicle is shown as $881 when the cost per vehicle presented earlier was $894. The $881
value is $894 discounted at 3 percent to the mid-year point of the first year of ownership.
 Table 12.51 Payback Period for the Sales Weighted Average MY2025 Vehicle in the Central Analysis using
               ICMs Relative to the Reference Case Standards (3% discounting, 2013$)
Vehicle Age




0
1
2
3
4
5
6
7
Delta Cost
per Vehicle



$881
$0
$0
$0
$0
$0
$0
$0
Delta Taxes
per Vehicle



$48
$0
$0
$0
$0
$0
$0
$0
Delta
Insurance
per Vehicle


$16
$16
$15
$14
$13
$12
$11
$11
Delta
Purchase
Costs per
Vehicle

$945
$16
$15
$14
$13
$12
$11
$11
Delta
Maintenance
Costs per
Vehicle

$5
$4
$4
$4
$4
$4
$3
$3
Delta Fuel
Costs per
Vehicle


-$239
-$231
-$222
-$214
-$202
-$191
-$178
-$167
Cumulative
Delta
Operating
Costs per
Vehicle
$711
$501
$298
$103
-$82
-$257
-$420
-$573
       Note: Costs are discounted to the first mid-year of vehicle ownership.
                                              12-42

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.52 Payback Period for the Sales Weighted Average MY2025 Vehicle in the Central Analysis using
               RPEs Relative to the Reference Case Standards (3% discounting, 2013$)
Age




0
1
2
3
4
5
6
7
Delta Cost
per Vehicle



$1,002
$0
$0
$0
$0
$0
$0
$0
Delta Taxes
per Vehicle



$55
$0
$0
$0
$0
$0
$0
$0
Delta
Insurance
per Vehicle


$19
$18
$17
$16
$15
$14
$13
$12
Delta
Purchase
Costs per
Vehicle

$1,075
$18
$17
$16
$15
$14
$13
$12
Delta
Maintenance
Costs per
Vehicle

$5
$5
$4
$4
$4
$4
$4
$3
Delta Fuel
Costs per
Vehicle


-$238
-$230
-$221
-$213
-$201
-$190
-$178
-$167
Cumulative
Delta
Operating
Costs per
Vehicle
$842
$634
$434
$241
$58
-$115
-$276
-$428
      Note: Costs are discounted to the first mid-year of vehicle ownership.
Table 12.53 Payback Period for the Sales Weighted Average MY2025 Vehicle in the Central Analysis using
               ICMs Relative to the Reference Case Standards (7% discounting, 2013$)
Age




0
1
2
3
4
5
6
7
Delta Cost
per Vehicle



$864
$0
$0
$0
$0
$0
$0
$0
Delta Taxes
per Vehicle



$47
$0
$0
$0
$0
$0
$0
$0
Delta
Insurance
per Vehicle


$16
$15
$14
$12
$11
$10
$9
$8
Delta
Purchase
Costs per
Vehicle

$928
$15
$14
$12
$11
$10
$9
$8
Delta
Maintenance
Costs per
Vehicle

$5
$4
$4
$4
$3
$3
$3
$2
Delta Fuel
Costs per
Vehicle


-$234
-$218
-$202
-$187
-$170
-$155
-$139
-$125
Cumulative
Delta
Operating
Costs per
Vehicle
$698
$499
$315
$144
-$12
-$154
-$281
-$396
      Note: Costs are discounted to the first mid-year of vehicle ownership.
                                                 12-43

-------
                                       EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.54 Payback Period for the Sales Weighted Average MY2025 Vehicle in the Central Analysis using
                RPEs Relative to the Reference Case Standards (7% discounting, 2013$)
Age




0
1
2
3
4
5
6
7
Delta Cost
per Vehicle



$983
$0
$0
$0
$0
$0
$0
$0
Delta Taxes
per Vehicle



$54
$0
$0
$0
$0
$0
$0
$0
Delta
Insurance
per Vehicle


$18
$17
$15
$14
$13
$11
$10
$9
Delta
Purchase
Costs per
Vehicle

$1,055
$17
$15
$14
$13
$11
$10
$9
Delta
Maintenance
Costs per
Vehicle

$5
$4
$4
$4
$3
$3
$3
$2
Delta Fuel
Costs per
Vehicle


-$234
-$218
-$201
-$187
-$170
-$154
-$139
-$125
Cumulative
Delta
Operating
Costs per
Vehicle
$826
$629
$448
$279
$125
-$15
-$142
-$255
       Note:  Costs are discounted to the first mid-year of vehicle ownership.


   EPA has also calculated the payback periods using the AEO 2015 High and Low fuel price
scenarios, at both the 3 percent and 7 percent discount rates.  Those results are shown in the
tables below and show, again, that payback occurs in the 5th year of ownership for the ICM cases
and in the 6th year when using RPEs, regardless of discount rate.
 Table 12.55 Payback Period for the Sales Weighted Average MY2025 Vehicle using AEO High Fuel Prices
              and ICMs Relative to the Reference Case Standards (3% discounting, 2013$)
Age




0
1
2
3
4
5
6
7
Delta Cost
per Vehicle



$859
$0
$0
$0
$0
$0
$0
$0
Delta Taxes
per Vehicle



$47
$0
$0
$0
$0
$0
$0
$0
Delta
Insurance
per Vehicle


$16
$16
$15
$14
$13
$12
$11
$10
Delta
Purchase
Costs per
Vehicle

$922
$16
$15
$14
$13
$12
$11
$10
Delta
Maintenance
Costs per
Vehicle

$4
$4
$4
$3
$3
$3
$3
$3
Delta Fuel
Costs per
Vehicle


-$225
-$218
-$209
-$202
-$191
-$181
-$170
-$159
Cumulative
Delta
Operating
Costs per
Vehicle
$701
$502
$311
$126
-$49
-$215
-$370
-$516
       Note:  Costs are discounted to the first mid-year of vehicle ownership.
                                                12-44

-------
                                       EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.56 Payback Period for the Sales Weighted Average MY2025 Vehicle using AEO High Fuel Prices
             and ICMs Relative to the Reference Case Standards (7% discounting, 2013$)
Age




0
1
2
3
4
5
6
7
Delta Cost
per Vehicle



$843
$0
$0
$0
$0
$0
$0
$0
Delta Taxes
per Vehicle



$46
$0
$0
$0
$0
$0
$0
$0
Delta
Insurance
per Vehicle


$16
$15
$13
$12
$11
$10
$9
$8
Delta
Purchase
Costs per
Vehicle

$904
$15
$13
$12
$11
$10
$9
$8
Delta
Maintenance
Costs per
Vehicle

$4
$4
$3
$3
$3
$2
$2
$2
Delta Fuel
Costs per
Vehicle


-$221
-$206
-$190
-$177
-$161
-$147
-$132
-$119
Cumulative
Delta
Operating
Costs per
Vehicle
$687
$500
$326
$164
$17
-$118
-$239
-$348
      Note:  Costs are discounted to the first mid-year of vehicle ownership.
Table 12.57 Payback Period for the Sales Weighted Average MY2025 Vehicle using AEO Low Fuel Prices
             and ICMs Relative to the Reference Case Standards (3% discounting, 2013$)
Age




0
1
2
3
4
5
6
7
Delta Cost
per Vehicle



$886
$0
$0
$0
$0
$0
$0
$0
Delta Taxes
per Vehicle



$48
$0
$0
$0
$0
$0
$0
$0
Delta
Insurance
per Vehicle


$17
$16
$15
$14
$13
$12
$11
$11
Delta
Purchase
Costs per
Vehicle

$951
$16
$15
$14
$13
$12
$11
$11
Delta
Maintenance
Costs per
Vehicle

$5
$4
$4
$4
$4
$4
$3
$3
Delta Fuel
Costs per
Vehicle


-$244
-$236
-$227
-$219
-$206
-$195
-$182
-$170
Cumulative
Delta
Operating
Costs per
Vehicle
$711
$495
$287
$87
-$102
-$281
-$449
-$605
      Note:  Costs are discounted to the first mid-year of vehicle ownership.
                                                12-45

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.58 Payback Period for the Sales Weighted Average MY2025 Vehicle using AEO Low Fuel Prices
             and ICMs Relative to the Reference Case Standards (7% discounting, 2013$)
Age




0
1
2
3
4
5
6
7
Delta Cost
per Vehicle



$869
$0
$0
$0
$0
$0
$0
$0
Delta Taxes
per Vehicle



$47
$0
$0
$0
$0
$0
$0
$0
Delta
Insurance
per Vehicle


$16
$15
$14
$12
$11
$10
$9
$8
Delta
Purchase
Costs per
Vehicle

$933
$15
$14
$12
$11
$10
$9
$8
Delta
Maintenance
Costs per
Vehicle

$4
$4
$4
$4
$3
$3
$3
$2
Delta Fuel
Costs per
Vehicle


-$240
-$223
-$206
-$191
-$174
-$158
-$142
-$128
Cumulative
Delta
Operating
Costs per
Vehicle
$698
$494
$305
$129
-$30
-$175
-$306
-$423
       Note: Costs are discounted to the first mid-year of vehicle ownership.
   The table below shows the cumulative increased lifetime savings associated with the
standards using each the 3 fuel price cases, both ICMs and RPEs, and at both the 3 percent and 7
percent discount rates.  Note that the values shown in the table include added costs associated
with maintenance, insurance and taxes, and the fuel savings resulting from less fuel usage.
These analyses compare the lifetime savings associated with a vehicle meeting the MY2025
standards under the various control cases to a vehicle meeting the MY2021 standards in MY2025
(the reference case). Lifetime savings across the central analysis scenarios range from $879 (for
the AEO 2015 Reference/RPE/7 percent discounting case) to $1,621 (for the AEO 2015
Reference/ICM/3 percent discounting case). Note that comparisons to the 2012 FRM lifetime
savings metrics are difficult, because in the FRM establishing standards for MY2017-2025, we
were comparing a vehicle meeting the 2025 standards to a vehicle meeting the 2016 standards as
the reference case, and thus, the accumulated lifetime savings were significantly higher (on the
order of $5,700 - $7,400 in  2010 dollars). The lifetime savings reflected in this Draft TAR for a
vehicle meeting the 2025 standards compared to a vehicle meeting the 2021 standards are
naturally covering a much smaller fraction of accumulated fuel savings as compared to the FRM
analysis.
Table 12.59 Lifetime Net Savings Associated with the Indicated Control Case Relative to the Reference Case
                        for the Sales-Weighted Average MY2025 Vehicle
Case
AEO Reference Fuel Price Case Using ICMs
AEO Reference Fuel Price Case Using RPEs
AEO High Fuel Price Case Using ICMs
AEO Low Fuel Price Case Using ICMs
Lifetime Savings
3% discounting
$1,621
$1,460
$1,506
$1,679
Lifetime Savings
7% discounting
$1,030
$879
$948
$1,072
                                              12-46

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
12.2  EPA's Projected Impacts on Emissions Inventories & Fuel
Consumption

12.2.1 Analytical Tools Used

   As in the 2012 final rule establishing MY2017-2025 standards, EPA used its OMEGA
Inventory Costs and Benefits Tool (ICBT) to project the emissions and fuel consumption impacts
of this analysis. The projections of the emission inventory and fleetwide fuel consumption are
conducted in the OMEGA ICBTJ which produces a national scale analysis of the impacts
(emission inventory and fuel consumption impacts, monetized co-benefits) of the analyzed
program.  The OMEGA ICBT incorporates the inputs discussed in Chapter 4 (baseline fleet),
Chapter 5 (technology costs and effectiveness) and Chapter 10 (vehicle miles traveled (VMT),
rebound, and other economic inputs).

   The remainder of this chapter provides a summary of the analytical inputs, methodology, and
the results of the analysis.

12.2.2 Inputs to the Emissions and Fuel Consumption Analysis

12.2.2.1       Methods

   EPA estimated GHG impacts from several sources including: (a) the impact of the standards
on tailpipe CO2 emissions, (b) projected improvements in the efficiency of vehicle air
conditioning systems, (c) reductions in direct emissions of the potent greenhouse gas refrigerant
HFC-134a from air conditioning systems, (d) "upstream" emission reductions from gasoline
extraction, production and distribution processes as a result of reduced gasoline demand
associated with standards, and (e) "upstream" emission increases from power plants as electric
powertrain vehicles are projected to increase slightly as a result of the MY2022-2025 standards.
EPA additionally accounted for the greenhouse gas impacts of additional vehicle miles traveled
(VMT) due to  the "rebound" effect discussed in Chapter 10.

   EPA's estimates of non-GHG emission impacts from the MY2022-2025 standards are broken
down by the three drivers of these changes: a) "downstream" emission changes, reflecting the
estimated effects of VMT rebound (discussed in Chapter 10) and decreased consumption of
motor vehicle fuel; b) "upstream" emission reductions due to decreased extraction, production
and distribution of motor vehicle gasoline; c)  "upstream" emission increases from power plants
as electric powertrain vehicles are projected to be slightly more prevalent in future years.K  For
all criteria and air toxic pollutants, the overall impact of the MY2022-2025 standards is small
compared to total U.S. inventories across all sectors.
1 Essentially the relevant ICBT elements are a post-processing tool to OMEGA used to incorporate inventory and
  cost-specific data not needed in OMEGA for use in this analysis.
K Note that the reference case used by EPA includes vehicle sales in response to the ZEV program. As such,
  increased power plant emissions associated with those ZEV-program vehicle sales are not attributable to the
  2022-2025 GHG standards. However, OMEGA projects a very small increase in EV and PHEV sales above those
  needed for ZEV compliance; the increased power plant emissions due to those additional EV/PHEV vehicles are
  attributable to the 2022-2025 GHG standards. Note that EPA has not yet updated the electricity emissions factors
  from those used in the 2012 FRM, though it is possible that emissions factors would change in the future due in
  part to EPA's Clean Power Plan regulations. This issue is discussed further in Chapter 11.5.
                                              12-47

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
   Although electric vehicles have zero tailpipe emissions, EPA assumes that manufacturers will
plan for these vehicles in their regulatory compliance strategy for criteria pollutant and air toxics
emissions, and will not over-comply with applicable Tier 3 emissions standards for non-GHG air
pollutants.  Since the Tier 3 emissions standards are fleet-average standards, EPA assumes that if
a manufacturer introduces EVs into its fleet, then it would correspondingly compensate through
changes to vehicles elsewhere in its fleet, rather than produce an overall lower fleet-average
emissions level.  Consequently, consistent with the 2012 FRM, EPA assumes neither tailpipe
pollutant (other than CCh), evaporative emissions, nor brake and tire wear particulate matter
reductions from the introduction of electric vehicles into the fleet.

   Two basic elements feed into the OMEGA ICBT calculation of vehicle tailpipe emissions.
These elements are vehicle miles traveled (VMT) and emission rates, where the total emissions
are the vehicle miles traveled multiplied by the emission rate in grams/mile.  This equation is
adjusted in calculations for various emissions, but provides the basic form used throughout this
analysis. As an example, in an analysis of a single calendar year,  the emissions  equation is
repeatedly applied to determine the contribution of each model year in the calendar year's
particular fleet. Appropriate VMT and emission factors by age are applied to each model year
within the calendar year, and the products are then summed.  Similarly, to determine the
emissions of a single model year, appropriate VMT and emission  factors by age are  applied to
each calendar year between when the model year fleet is produced and projected to be scrapped.

   Tailpipe sulfur dioxide (SCh) emissions, which are largely controlled by the sulfur content of
the fuel, are an exception to this basic equation.  Decreasing the quantity of fuel consumed
decreases tailpipe SCh emissions proportionally to the decrease in fuel combusted. Therefore,
rather than multiplying the SCh emission factor by miles traveled, we multiply by gallons
consumed.  As such, the SCh emission factor is expressed in terms of grams/gallon rather than
grams/mile.
12.2.2.2
Global Warming Potentials
   In general, when we refer to the four inventoried greenhouse gases on an equivalent basis,
Global Warming Potentials (GWPs) are used. In simple terms, GWPs provide a common basis
with which to combine several gases with different heat trapping abilities into a single inventory.
When expressed in CCh equivalent (CChe) terms, each gas is weighted by its heat trapping
ability relative to that of carbon dioxide. The GWPs used are shown in Table 12.60.L
                Table 12.60 Global Warming Potentials (GWP) for Inventoried GHGs
GHG
C02
CH4
N2O
HFC (R134a)
GWP
(C02e)
1
25
298
1430
L As with the MY 2017-2025 Light Duty rule and the MY 2014-2018 Medium and Heavy Duty rule, the GWPs used
  in this rule are consistent with 100-year time frame values in the 2007 Intergovernmental Panel on Climate
  Change (IPCC) Fourth Assessment Report (AR4).
                                              12-48

-------
                                    EPA's Analysis of the MY2022-2025 GHG Standards
12.2.2.3       Years Considered

   This analysis presents the projected impacts of the standards in calendar years 2025, 2030,
2040 and 2050. We also present the emission impacts over the estimated full lifetime of MYs
2022-2025 vehicles. The program was quantified as the difference in mass emissions between a
control case under the final MY2022-2025 standards and a reference case under the MY2021
standards in place indefinitely.  As such, negative values represent emissions decreases due to
the policy and positive values represent emissions increases due to the policy.

12.2.2.4      Fleet Activity

12.2.2.4.1     Vehicle Sales, Survival Schedules, and VMT

   Vehicle sales projections from MY2014 through MY2030 are discussed in Chapter 4.
Vehicle survival schedules and VMT by vehicle age were updated to be consistent with the most
recent publicly released EPA MOVES model (MOVES2014a).  These updates are described in
more detail in Chapter 10.

12.2.2.5       Upstream Emission Factors

12.2.2.5.1    Gasoline Production and Transport Emission Rates

   The gasoline production and transport sector is composed of four distinct components:

       •   Domestic crude oil production and transport

       •   Petroleum production and refining emissions

       •   Production of energy for refinery use

       •   Gasoline transport, storage and distribution

   For this Draft TAR analysis, the emission factors associated with on-road combustion
emissions allocated to gasoline transport and distribution were updated based on the emission
factors calculated as part of the HD GHG Phase 2 rule.5 Refinery related emissions were
updated to reconcile the emission totals with those in the most recent national emission
inventory.6 Otherwise, the upstream emission rate analysis remains the same as that performed
in the 2012 FRM Regulatory Impacts Analysis (RIA), Chapters 4.2 and 4.6.7 Table 12.61,
below, shows the gasoline upstream emission rates used in the cost-benefit calculations for this
analysis.
                                             12-49

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
                         Table 12.61 Gasoline Production Emission Rates
Pollutant
CO
NOx
PM2.5
PM10
SOx
VOC
1,3- Butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
Naphthalene
CH4
N2O
CO2
Emission Rate
(g/MMbtuofElO
gasoline)
5.472145
13.87269
2.07292
6.048208
8.089376
47.4966
0.001442
0.009798
0.000816
0.322958
0.081647
0.015177
95.454
0.369224
19145.2
12.2.2.5.2     Electricity Generation Emission Rates

       For the 2012 FRM, EPA conducted an Integrated Planning Model (IPM) analysis of the
electricity sector in order to gauge the impacts upon the power grid of the additional electric
charging projected to be needed to meet the MY2017-2025 standards.8  Since the 2012 final rule,
EPA has adopted a GHG program for electricity generation, known as the Clean Power Plan.M
These rules are expected to significantly decrease GHG emissions associated with future
electricity generation. The 2012 FRM's IPM modeling projected that the average power plant
electricity GHG emissions factor in 2030 for vehicle electricity use would be 0.445 grams/watt-
hour.9  The overall vehicle electricity GHG emissions factor was projected to be 0.534
grams/watt-hour when using a multiplicative value of 1.20 to account for feedstock-related GHG
emissions upstream of the power plant.  EPA is currently exploring whether there are appropriate
updates to these projected emissions factors for the incremental electricity that would be
necessary for electric vehicle operation  in the 2030 timeframe, which we plan to assess in more
detail further in the midterm evaluation process. For this Draft TAR, EPA is continuing to apply
the FRM IPM results as a representation of the electrical grid in the time period surrounding
2030. The emission factors are shown in Table 12.62 below.

       The 2030 IPM results were post-processed to develop gram per kWh emission factors for
use in the OMEGA model and inventory cost-benefit analysis. For those emissions that IPM
does not generate, we relied upon the National Emissions Inventory (NEI) for air toxic emissions
and eGrid for N2O and CH4. There are  also additional emissions attributable to feedstock
generation, or the gathering and transport of fuel to the power plant. Emission factors from the
version of GREET 1.8c (as modified for the EPA upstream analysis discussed above) were used
to generate feedstock emission factors.  Retail electricity price projections from the 2030 FRM
M EPA issued a final GHG emissions program, known as the Clean Power Plan, addressing fossil fuel-fired electric
  generating units. 80 FR 64661, October 23, 2015.
                                              12-50

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
IPM run were used in our analysis of electricity fuel costs to drivers. More information
regarding the integration of GREET emission factors and IPM modeling can be found in the
FRMRIA, Chapter 4.6.

               Table 12.62 Emission Factors Used in Analysis of Electricity Generation
Pollutant
VOC
CO
NOx
PM2.5
S02
C02
N2O
CH4
1,3-butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
IPM
(g/kWh)
8.28E-03
2.89E-01
1.13E-01
5.81E-03
1.90E-01
4.45E+02
6.76E-03
8.60E-03
O.OE+00
5.5E-05
2.8E-05
1.3E-04
3.0E-05
Feedstock
(g/kWh)
4.69E-02
5.01E-02
1.27E-01
6.51E-02
4.69E-02
3.55E+01
6.81E-04
3.31E+00
O.OOE+00
9.47E-06
3.15E-05
1.41E-03
7.51E-06
Total
(g/kWh)
5.52E-02
3.39E-01
2.41E-01
7.09E-02
2.37E-01
4.80E+02
7.44E-03
3.32E+00
O.OOE+00
6.40E-05
5.95E-05
1.54E-03
3.79E-05
12.2.2.6
Reference Case CO2 g/tni & kWh/mi
   As described in Section 12.1, EPA assumes that the reference case fleet continues to meet the
MY2021 standards indefinitely. Importantly, we model the fleet as meeting the reference (or
control) case targets rather than the achieved CCh values as reported by the OMEGA core model.
We do this because we consider OMEGA core model results to be a possible, feasible path
toward compliance and not necessarily the actually path that any given manufacturer will choose.
For that reason, we choose to model the target values.  Compliance flexibilities such as A/C
credits and fleet averaging are included in the modeling.  The A/C direct credit is added here to
the 2-cycle target value to arrive at the 2-cycle tailpipe CO2 value because, while that credit
results in real GHG reductions, it does not result in real tailpipe CO2 reductions (or real on-road
fuel economy improvements).  The benefits of off-cycle and A/C indirect credits  are implicitly
included in the values below because they result in real CO2 reductions.  The CO2 targets
presented here were also presented in  Section 12.1.1. The fleet CO2 g/mi and kWh/mi  emission
rates used for inventory modeling are  as shown in the tables below. In the CO2 g/mi tables, the
on-road tailpipe CO2 values are the values used in generating CO2 inventory impacts in the
reference case. The "gap" noted in the tables below is the gap between compliance and real
world fuel economy/tailpipe CO2, discussed further in Chapter 10.1. Entries change slightly year-
over-year due to  fleet changes.
                                             12-51

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
          Table 12.63 Reference Case Car On-Road CCh g/mi Used in All OMEGA ICBT Runs
MY
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2-cycle CO2
Target, g/mi
177.0
177.0
177.0
176.9
177.0
176.9
176.9
176.8
176.7
176.7
A/C Direct
Credit, g/mi
13.8
13.8
13.8
13.8
13.8
13.8
13.8
13.8
13.8
13.8
Adjusted 2-cycle
Tailpipe CO2, g/mi
190.8
190.8
190.8
190.7
190.8
190.7
190.7
190.6
190.5
190.5
Adjusted
MPG
46.6
46.6
46.6
46.6
46.6
46.6
46.6
46.6
46.6
46.6
Gap
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
On-road
MPG
35.9
35.9
36.0
36.0
36.0
36.0
36.0
36.0
36.0
36.0
On-road CO2
Tailpipe, g/mi
236.1
236.2
236.1
236.0
236.1
236.0
235.9
235.9
235.8
235.8
Note: The on-road values reflect adjustments for both the historical 2-cycle-to-5-cycle gap as well as the projected
ethanol content in retail gasoline, and corresponding energy content. The on-road CC>2 is calculated by dividing
8488, the estimated CO2 grams/gallon from combustion of a gallon of retail gasoline, by the on-road MPG.

         Table 12.64 Reference Case Truck On-Road COi g/mi Used in All OMEGA ICBT Runs
MY
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2-cycle CO2
Target, g/mi
251.1
251.1
250.8
250.8
250.9
250.9
251.3
251.3
250.9
250.9
A/C Direct
Credit, g/mi
17.2
17.2
17.2
17.2
17.2
17.2
17.2
17.2
17.2
17.2
Adjusted 2-cycle
Tailpipe CO2, g/mi
268.3
268.3
268.0
268.0
268.1
268.1
268.5
268.5
268.1
268.1
Adjusted
MPG
33.1
33.1
33.2
33.2
33.1
33.2
33.1
33.1
33.1
33.2
Gap
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
On-road
MPG
25.6
25.6
25.6
25.6
25.6
25.6
25.6
25.5
25.6
25.6
On-road CO2
Tailpipe, g/mi
332.1
332.1
331.6
331.6
331.8
331.7
332.2
332.2
331.8
331.7
Note: The on-road values reflect adjustments for both the historical 2-cycle-to-5-cycle gap as well as the projected
ethanol content in retail gasoline, and corresponding energy content. The on-road CO2 is calculated by dividing
8488, the estimated CO2 grams/gallon from combustion of a gallon of retail gasoline, by the on-road MPG.
   The reference case electricity consumption rates, including both electricity consumption by
ZEV program vehicles and consumption by the very small fraction of EV and PHEV vehicles
projected by OMEGA toward compliance with the reference case standards are shown in the
table below. EPA accounts for all electricity consumed by the vehicle. For calculations of GHG
emissions from electricity generation, the total energy consumed from the battery is divided by
0.9 to account for charging losses. This factor is included in the values presented in the table
below. Within the OMEGA ICBT, a transmission loss divisor of 0.93 is applied  to account for
losses during transmission, the result being electricity demand at the electric plant.  Both values
were discussed in the 2012 FRM; the approach in  this analysis is unchanged.10 The estimate of
charging losses is based upon engineering judgment and manufacturer CBI. The estimate of
transmission losses is consistent, although not identical to the 8 percent estimate used in GREET,
as well as the 6 percent estimate in eGrid 2010.11'12 The upstream emission factor discussed
above in Section 12.2.2.5.2 is applied to total electricity production, rather than simply power
                                               12-52

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
consumed at the wheel.N  It is assumed that electrically powered vehicles drive the same drive
schedule as the rest of the fleet.0 Note that the values shown in the table already include a 0.8
on-road "gap" since the gap was considered in determining battery sizing and consumption.1"

   Because the kWh/mi inputs to the OMEGA ICBT differ based on fuel price case and whether
ICMs or RPEs are used in each set of inputs are shown below.  The values shown in the kWh/mi
table are the values used to generate upstream emission inventory impacts in the applicable
reference case.

  Table 12.65 Reference Case Car & Truck On-Road kWh/mi Consumption used in the Indicated OMEGA
                                        ICBT Runs


MY
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
ICMs
AEO Ref
Car
0.01260
0.01385
0.01510
0.01635
0.01760
0.01760
0.01760
0.01760
0.01760
0.01760
Truck
0.00137
0.00167
0.00198
0.00228
0.00259
0.00259
0.00259
0.00259
0.00259
0.00259
AEO High
Car
0.01226
0.01353
0.01481
0.01608
0.01735
0.01735
0.01735
0.01735
0.01735
0.01735
Truck
0.00132
0.00162
0.00192
0.00222
0.00252
0.00252
0.00252
0.00252
0.00252
0.00252
AEO Low
Car
0.01324
0.01464
0.01604
0.01745
0.01885
0.01885
0.01885
0.01885
0.01885
0.01885
Truck
0.00141
0.00171
0.00201
0.00232
0.00262
0.00262
0.00262
0.00262
0.00262
0.00262
RPEs
AEO Ref
Car
0.01296
0.01427
0.01559
0.01690
0.01821
0.01821
0.01821
0.01821
0.01821
0.01821
Truck
0.00137
0.00167
0.00198
0.00228
0.00259
0.00259
0.00259
0.00259
0.00259
0.00259
   For this Draft TAR analysis, EPA has considered the ZEV program in California and Section
177 states in the reference case for this analysis. That analysis fleet is described in detail in
Chapter 4.  Our central analysis also treats EVs and the electricity portion of PHEV operation as
zero emitting for compliance purposes (although their upstream emissions are considered in our
GHG emission inventory estimates). Given the ZEV program sales, it appears that some
manufacturers are likely to exceed the sales levels beyond which net upstream emissions would
have to be considered in their compliance determination.13 However, other manufacturers appear
unlikely to exceed that limit.  In the current version of OMEGA, EPA does not have the
capability to apply upstream emissions to only some manufacturers' fleets and not others.  This is
a change we plan to implement in future updates to the OMEGA model.
12.2.2.7
Control Case CO2 g/mi & kWh/mi
   As noted above, we model the fleet as meeting the compliance targets rather than the achieved
CO2 values as reported by the OMEGA core model.  We do this because we consider OMEGA
core model results to be a possible path toward compliance and not necessarily the path that will
result.  For that reason, we choose to model the target values since those represent the levels that
are actually required.  The off-cycle credits are implicitly included in the values below, as are all
A/C credits, because their use is assumed in meeting the "2-cycle CO2 Target" values shown.
N By contrast, consumer electricity costs would not include the power lost during transmission. While consumers
  indirectly pay for this lost power through higher rates, this power does not appear on their electric meter.
0 The validity of this assumption will depend on the use of electric vehicles by their purchasers.
p See Chapter 5 for details on EPA's battery sizing methodology.
                                              12-53

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
The A/C direct credit is added here to the 2-cycle target value to arrive at the adjusted 2-cycle
tailpipe CCh value because, while that credit results in real GHG reductions, it does not result in
real tailpipe CCh reductions (or real on-road fuel economy improvements). The CCh targets
presented here were also presented in Section 12.1.1.  The fleet CCh g/mi and kWh/mi emission
rates used for inventory modeling are as shown in the tables below.  In the CCh g/mi tables, the
on-road tailpipe CCh value is the value used in generating CCh inventory impacts in the control
case.  The "Gap" noted in the tables below is the gap between compliance and real world fuel
economy/tailpipe CCh, discussed in Chapter 10.1.  The gap, as shown, is applied to adjusted
MPG values.

           Table 12.66  Control Case Car On-Road CCh g/mi Used in All OMEGA ICBT Runs
MY
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2-cycle CO2
Target, g/mi
171.1
164.8
158.7
152.8
147.3
147.2
147.2
147.1
147.1
147.1
A/C Direct
Credit, g/mi
13.8
13.8
13.8
13.8
13.8
13.8
13.8
13.8
13.8
13.8
Adjusted 2-cycle
Tailpipe CO2, g/mi
184.9
178.6
172.5
166.6
161.1
161.0
161.0
160.9
160.9
160.9
Adjusted
MPG
48.1
49.8
51.5
53.3
55.2
55.2
55.2
55.2
55.2
55.2
Gap
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
On-road MPG
37.1
38.4
39.8
41.2
42.6
42.6
42.6
42.6
42.6
42.6
On-road CO2
Tailpipe, g/mi
228.9
221.1
213.5
206.3
199.4
199.4
199.3
199.3
199.2
199.2
Note:  The on-road values reflect adjustments for both the historical 2-cycle-to-5-cycle gap as well as the projected
ethanol content in retail gasoline, and corresponding energy content. The on-road CO2 is calculated by dividing
8488, the estimated CO2 grams/gallon from combustion of a gallon of retail gasoline, by the on-road MPG; off-cycle
credits are not shown in the table since they are assumed to have been used in meeting the 2-cycle CCh Targets.
          Table 12.67 Control Case Truck On-Road COi g/mi Used in All OMEGA ICBT Runs
MY
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2-cycle CO2
Target, g/mi
242.0
232.3
222.7
213.8
205.5
205.5
205.8
205.8
205.5
205.5
A/C Direct
Credit, g/mi
17.2
17.2
17.2
17.2
17.2
17.2
17.2
17.2
17.2
17.2
Adjusted 2-cycle
Tailpipe CO2, g/mi
259.2
249.5
239.9
231.0
222.7
222.7
223.0
223.0
222.7
222.7
Adjusted
MPG
34.3
35.6
37.0
38.5
39.9
39.9
39.9
39.9
39.9
39.9
Gap
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
0.77
On-road
MPG
26.5
27.5
28.6
29.7
30.8
30.8
30.8
30.8
30.8
30.8
On-road CO2
Tailpipe, g/mi
321.0
309.0
297.1
286.1
275.8
275.7
276.1
276.1
275.8
275.7
Note:  The on-road values reflect adjustments for both the historical 2-cycle-to-5-cycle gap as well as the projected
ethanol content in retail gasoline, and corresponding energy content.  The on-road CO2 is calculated by dividing
8488, the estimated CO2 grams/gallon from combustion of a gallon of retail gasoline, by the on-road MPG; off-
cycle credits are not shown in the table since they are assumed to have been used in meeting the 2-cycle CCh Targets
and because they provide real-world COa reductions so do not need to be backed out as do the A/C leakage, or A/C
direct credit, values.
                                                 12-54

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
   The table below shows the control case electricity emission factors, including both electricity
consumption by ZEV program vehicles and projected EV and PHEV vehicles generated by
OMEGA toward compliance with the control case standards. These consumption levels include
charging losses (a 90 percent divisor) and the OMEGA ICBT applies a 93 percent transmission
loss divisor (not included in the values below). Note that the values shown in the table already
include a 0.8 on-road "gap" since the gap was considered in determining battery sizing and
consumption.

   The control case kWh/mi inputs to the OMEGA ICBT are shown in the table below.  Because
fuel prices, and choice of ICMs or RPEs, impact the projected penetration of EV and PHEV
vehicles, unique kWh/mi inputs are presented for each combination fuel price and indirect cost
scenario. The values shown in the kWh/mi table are the values used to generate upstream
emission inventory impacts in the applicable control case.
  Table 12.68 Reference Case Car & Truck On-Road kWh/mi Consumption used in the Indicated OMEGA
                                       ICBT Runs


MY
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
ICMs
AEO Ref
Car
0.01350
0.01524
0.01698
0.01871
0.02045
0.02045
0.02045
0.02045
0.02045
0.02045
Truck
0.00137
0.00167
0.00198
0.00228
0.00259
0.00259
0.00259
0.00259
0.00259
0.00259
AEO High
Car
0.01304
0.01476
0.01649
0.01822
0.01995
0.01995
0.01995
0.01995
0.01995
0.01995
Truck
0.00132
0.00178
0.00223
0.00268
0.00314
0.00314
0.00314
0.00314
0.00314
0.00314
AEO Low
Car
0.01412
0.01599
0.01786
0.01973
0.02160
0.02160
0.02160
0.02160
0.02160
0.02160
Truck
0.00141
0.00186
0.00230
0.00275
0.00320
0.00320
0.00320
0.00320
0.00320
0.00320
RPEs
AEO Ref
Car
0.01423
0.01607
0.01790
0.01974
0.02157
0.02157
0.02157
0.02157
0.02157
0.02157
Truck
0.00137
0.00167
0.00198
0.00228
0.00259
0.00259
0.00259
0.00259
0.00259
0.00259
   It is important to emphasize that these CO2 and kWh emission rate projections are based on
EPA's current projections of a wide range of inputs, including the mix of cars and trucks, as well
as the mix of vehicle footprint values in varying years. It is of course possible that the actual
CO2 emissions values, as well as the actual use of incentives and credits, will be either higher or
lower than these projections.
12.2.2.8
Criteria Pollutant and Select Toxic Pollutant Emission Rates
   For the analysis of criteria emissions in this rule, EPA estimates the increases in emissions of
each criteria air pollutant from additional vehicle use by multiplying the increase in total miles
driven by cars and light trucks of each model year and age by their estimated emission rates per
vehicle-mile of each pollutant.  These emission rates differ between cars and light trucks,
between gasoline and diesel vehicles, and by age. With the exception of SO2, EPA calculated
the increase in emissions of these criteria pollutants from added car and light truck use by
multiplying the estimated increases in vehicle use during each year over their expected lifetimes
by per-mile emission rates appropriate to each vehicle type, fuel used, model year, and age as of
that future year.
                                             12-55

-------
                                    EPA's Analysis of the MY2022-2025 GHG Standards
   The relevant emission rates were estimated by EPA using the most recent version of the
Motor Vehicle Emission Simulator (MOVES2014a).  The MOVES model assumes that the per-
mile rates at which these pollutants are emitted are determined by EPA regulations and the
effectiveness of after-treatment of engine exhaust emissions, and are thus unaffected by changes
in car and light truck fuel economy. As a consequence, the downstream impacts of required
increases in fuel economy on emissions of these pollutants from car and light truck use are
determined entirely by the increases in driving that result from the fuel economy rebound effect.

   Emission factors in the MOVES database are expressed in the form of grams per vehicle-hour
of operation. To convert these emission factors to grams per mile, MOVES was run for the year
2050, and was  programmed to report aggregate emissions from vehicle start, running, brake and
tire wear and crankcase exhaust operations.  EPA analysts ran MOVES for every calendar year
from 2014 to the year 2050 in order to generate emission factors for each age of each model
year.  Separate estimates were developed for each vehicle type, as well as for a winter and a
summer month in order to reflect the effects of temporal variation in temperature and other
relevant variables on emissions. All calendar years were run using national averages calculated
from the aggregation of the county level default estimates (national aggregation).

   The MOVES emissions estimates were then summed to the model year level and divided by
total distance traveled by vehicles of that model year in order to produce per-mile emission
factors for each pollutant. The resulting emission rates represent average values across the
nation, and incorporate  variation in temperature and other operating conditions affecting
emissions over an entire calendar year. These national average rates also reflect county-specific
differences in fuel composition, as well as in the presence and type of vehicle inspection and
maintenance programs.  Average emission rates were assumed not to increase after 30 years of
age.

   Emission rates for the criteria pollutant SO2 were calculated by using average fuel sulfur
content estimates supplied by EPA, together with the simplifying assumption that the entire
sulfur content of fuel is emitted in the form of SO2.  These calculations assumed that national
average gasoline and diesel sulfur levels would remain at current levels, because there are no
current regulations that will change those levels, and we have no expectation that the market will
cause such  changes on its own.

12.2.3 Outputs of the Emissions and Fuel Consumption Analysis

   In this section. EPA presents the emissions inventory impacts, fuel, and  electricity
consumption results.  Section 12.2.3.1 shows impacts in a given calendar year resulting from the
control case analysis. These results are not cumulative, and are presented to show the continued
impacts of the  analysis beyond the control case years. Section 12.2.3.2 shows impacts for a
given model year cohort of vehicles, as well as cumulative sums of impacts due to vehicle model
years included in the control case (over the whole vehicle lifetime, as discussed in Chapter 10).
Tables presenting emissions inventory impacts are generally shown as reductions, such that
emission decreases would be shown as a positive number. Tables presenting fuel and energy
consumption are shown as absolute impact,  such that fuel or energy consumption decreases
would be show as a negative number.  See specific table notes for more direction. Discussion of
the inputs to this analysis can be found in section 12.2.2, above.
                                             12-56

-------
                                         EPA's Analysis of the MY2022-2025 GHG Standards
12.2.3.1
Calendar Year Results
   Table 12.69 Annual Emissions Reductions of the MY2022-2025 Standards on GHGs in Select Calendar
                                        Years (MMTCChe)Q
Calendar Year
Net GHG
Net CO2
Net other GHG
Downstream GHG
CO2 (excluding A/C)
A/C- indirect CO2
A/C- direct H PCs
CH4 (rebound effect)
N2O (rebound effect)
Fuel Production and
Distribution GHG
Fuel Production and
Distribution CO2
Fuel Production and
Distribution CH4
Fuel Production and
Distribution N2O
Electricity Upstream GHG
Electricity Upstream CO2
Electricity Upstream CH4
Electricity Upstream N2O
2025
40.7
39.9
0.9
32.4
32.3
0.1
0
0
0
9.1
8
1
0.1
-0.9 to -0.8
-0.8 to -0.7
-0.1
0.0
2030
102
100
2.3
81.6
81.3
0.3
0
0
0
22.8
20.2
2.5
0.1
-2.3 to -1.9
-1.9 to -1.6
-0.3
0.0
2040
186
182
4.1
148
147
0.6
0
0
-0.1
41.5
36.7
4.6
0.2
-4.1 to -3.5
-3.5 to -3.0
-0.6 to -0.5
0.0
2050
234
229
5.2
186
185
0.7
0
0
-0.1
52.3
46.2
5.8
0.3
-5.1 to -4.4
-4.3 to -3.7
-0.7 to -0.6
0.0
Note: These values are expressed as emission reductions, such that positive values imply an emissions decrease, and
negative values imply an emissions increase.
Q With the exception of upstream electricity generation due to differing technology mix, the differences in total
  inventory between ICM and RPE cases are negligible and have been omitted. Results are consistent with the ICM
  case where ranges are not shown.
                                                   12-57

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
     Table 12.70 Annual Emission Reductions of the MY2022-2025 Standards on GHGs (MMT CChe)
Calendar Year
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
Sum
CO2
2.6
8.0
16.1
26.7
39.9
52.8
65.4
77.6
89.2
100
111
121
130
139
148
156
163
170
176
182
187
192
197
202
207
211
216
220
225
229
4060
HFC
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
CH4
0.0
0.2
0.4
0.6
0.9
1.2
1.5
1.7
2.0
2.2
2.5
2.7
2.9
3.1
3.3
3.5
3.6
3.8
3.9
4.0
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
5.0
5.1
90.4
N2O
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
2.2
Total
2.7
8.2
16.4
27.3
40.8
54.0
66.9
79.4
91.3
102
113
124
133
143
151
159
167
174
180
186
191
197
202
207
211
216
221
225
230
234
4153
Note:  These values are expressed as emission reductions,
negative values imply an emissions increase.
such that positive values imply an emissions decrease, and
                                                 12-58

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.71  Annual Emission Reductions of the MY2022-2025 Standards on non-GHG Criteria Pollutants in
                                           Select Years


Total
Downstream
(Rebound)
Fuel production
& distribution
Electricity

Pollutant
VOC
CO
NOx
PM2.5
SOx
VOC
CO
NOx
PM2.5
SOx
VOC
CO
NOx
PM2.5
SOx
VOC
CO
NOx
PM2.5
SOx
CY2030
Impacts
(short tons)
53672
-30665
13763
2066.5
8512.5
-1419
-35762
-1483
-80.5
-16.5
55298
6370
16151
2413
9418
-207
-1273
-905
-266
-889
% of US
Inventory"
0.091
-0.038
0.089
0.034
0.131
-0.002
-0.044
-0.010
-0.001
0.000
0.094
0.008
0.104
0.040
0.145
0.000
-0.002
-0.006
-0.004
-0.014
CY2040
Impacts
(short tons)
96711
-69582
24334
3704
15426.4
-3203
-78807
-3304
-186
-29.6
100293
11554
29294
4377
17082
-379
-2329
-1656
-487
-1626
% of US
Inventory
0.164
-0.086
0.157
0.061
0.238
-0.005
-0.098
-0.021
-0.003
0.000
0.170
0.014
0.189
0.072
0.264
-0.001
-0.003
-0.011
-0.008
-0.025
Note:  These values are expressed as emission reductions,
negative values imply an emissions increase.
such that positive values imply an emissions decrease, and
                                                  12-59

-------
                                       EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.72 Annual Emission Reductions of the MY2022-2025 Standards on Select Toxic Pollutants in Select
                                            Years


Total
Downstream
(Rebound)
Fuel production
& distribution
Electricity

Pollutant
1,3- Butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
1,3- Butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
1,3- Butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
1,3- Butadiene
Acetaldehyde
Acrolein
Benzene
Formaldehyde
CY2030
Impacts
(short tons)
-8.7
-5.3
-1.0
311.6
59.6
-10.4
-16.5
-1.8
-58.6
-35.4
1.7
11.4
1.0
376
95.1
0
-0.2
-0.2
-5.8
-0.1
% of US
Inventory
-0.014
-0.001
-0.002
0.110
0.004
-0.017
-0.002
-0.004
-0.021
-0.003
0.003
0.001
0.002
0.133
0.007
0.000
0.000
0.000
-0.002
0.000
CY2040
Impacts
(short tons)
-20.0
-17.2
-2.8
539.4
91.5
-23
-37.5
-4.1
-132
-80.2
3.0
20.7
1.7
682
172
0
-0.4
-0.4
-10.6
-0.3
% of US
Inventory
-0.033
-0.002
-0.006
0.190
0.007
-0.038
-0.005
-0.008
-0.047
-0.006
0.005
0.002
0.003
0.241
0.013
0.000
0.000
-0.001
-0.004
0.000
Note:  These values are expressed as emission reductions, such that positive values imply an emissions decrease, and
negative values imply an emissions increase.
   The fuel consumption analysis relied on the same set of fleet and activity inputs as the
emission analysis.  EPA modeled the entire fleet as using petroleum gasoline (consistent with
OMEGA model results showing a lack of projected diesel penetration in the central analysis),
and used a conversion factor of 8887 grams of CCh per gallon of petroleum gasoline in order to
determine the quantity of fuel savings. The term petroleum gasoline is used here to mean fuel
with 115,000 BTU/gallon.  This is different than retail fuel, which is typically blended with
ethanol and has a lower energy content as discussed earlier in Section 12.2.2.7.
R The total US inventory for selected pollutants (in short tons) was derived from the EPA National Emissions
  Inventory (NEI) 2011 (https://www.epa.gov/air-emissions-inventories/national-emissions-inventory)
                                                12-60

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
     Table 12.73 Annual Impacts of the MY2022-2025 Standards on Fuel and Electricity Consumption
Calendar Year
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
Sum
Petroleum Gasoline
(billion gallons)
-0.25
-0.77
-1.54
-2.56
-3.82
-5.05
-6.26
-7.43
-8.54
-9.59
-10.60
-11.57
-12.48
-13.34
-14.15
-14.91
-15.60
-16.25
-16.84
-17.39
-17.92
-18.41
-18.88
-19.34
-19.78
-20.21
-20.64
-21.07
-21.49
-21.92
-389
Petroleum Gasoline
(billion barrels)
-0.01
-0.02
-0.04
-0.06
-0.09
-0.12
-0.15
-0.18
-0.20
-0.23
-0.25
-0.28
-0.30
-0.32
-0.34
-0.35
-0.37
-0.39
-0.40
-0.41
-0.43
-0.44
-0.45
-0.46
-0.47
-0.48
-0.49
-0.50
-0.51
-0.52
-9.26
Electricity
(billion kWh)
0.11
0.29
0.54
0.86
1.26
1.66
2.05
2.44
2.81
3.17
3.52
3.85
4.17
4.46
4.74
5.00
5.23
5.44
5.63
5.80
5.96
6.11
6.25
6.39
6.53
6.66
6.79
6.92
7.04
7.17
129
Note:  These values are expressed as absolute inventory changes, such that negative values imply a decrease in
consumption, and positive values imply an increase in consumption.
12.2.3.2
Model Year Lifetime Results
   Table 12.74 MY Lifetime Emission Reductions of the MY2022-2025 Standards on GHGs (MMT CChe)
Model Year
2021
2022
2023
2024
2025
Sum
Downstream
(including A/C)
27.4
56.9
85.7
114
144
428
Fuel Production
& Distribution
7.7
15.9
24.0
32.0
40.2
120
Electricity
-0.9
-1.4
-2.0
-2.6
-3.2
-10.0
Total
34.2
71.4
108
144
181
538
Note:  These values are expressed as emission reductions, such that positive values imply an emissions decrease, and
negative values imply an emissions increase.
                                                  12-61

-------
                                       EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.75 MY Lifetime Emission Reductions of the MY2022-2025 Standards on Select non-GHG Criteria
                                           Pollutants
                                          (Short tons)
Model Year
2021
2022
2023
2024
2025
Sum
voc
17,635
36,730
55,546
74,346
93,600
277,857
CO
-19,775
-37,876
-52,658
-64,598
-73,959
-248,864
NOx
3,977
8,644
13,398
18,295
23,445
67,760
PM2.5
650
1,407
2,141
2,866
3,600
10,663
S02
2,752
5,904
8,969
12,000
15,069
44,693
Note:  These values are expressed as emission reductions, such that positive values imply an emissions decrease, and
negative values imply an emissions increase.
  Table 12.76 MY Lifetime Emission Reductions of the MY2022-2025 Standards on Select Toxic Pollutants
                                          (Short tons)
Model Year
2021
2022
2023
2024
2025
Sum
Benzene
89
190
293
399
512
1,482
1,3 Butadiene
-5.3
-10.4
-14.7
-18.4
-21.6
-70.4
Formaldehyde
22.0
46.8
72.4
99.1
127
368
Acetaldehyde
-16.0
-31.5
-44.2
-54.9
-63.9
-210
Acrolein
-0.8
-1.6
-2.1
-2.6
-3.0
-10.1
Note:  These values are expressed as emission reductions, such that positive values imply an emissions decrease, and
negative values imply an emissions increase.
   Table 12.77  MY Lifetime Impacts of the MY2022-2025 Standards on Fuel and Electricity Consumption
Model Year
2021
2022
2023
2024
2025
Sum
Retail Gasoline
(billion gallons)
-3.2
-6.7
-10.1
-13.4
-16.9
-50.3
Retail Gasoline
(billion barrels)
-0.1
-0.2
-0.2
-0.3
-0.4
-1.2
Electricity
(billion kWh)
1.4
2.3
3.2
4.2
5.3
16.4
Note:  These values are expressed as absolute inventory changes, such that negative values imply a decrease in
consumption, and positive values imply an increase in consumption.
12.2.4 Sensitivity Analysis Results

   In this section, EPA presents the central case emissions impact analysis results using AEO
2015 reference fuel price cases (shown in Section 12.2.3) with two additional analyses based on
the low and high fuel price cases found in the  Annual Energy Outlook 2015 report (see Chapter
10 for more discussion regarding these fuel price cases).  These additional analyses provide a
good bracket around the uncertainty in fuel price projections and shows the magnitude of the
effect of differing fuel price projections on emission impacts. Similarly to Section 12.2.3,
Section 12.2.4.1 shows non-cumulative calendar year results for all three fuel price cases, and
Section 12.2.4.2 shows model year lifetime and cumulative sum results for all three fuel price
cases.
                                                12-62

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
12.2.4.1
Calendar Year Case Comparison Results
Table 12.78 Annual Emission Reductions of the MY2022-2025 Standards and AEO Fuel Price Cases on Total
                                       GHGs (MMT CChe)
Calendar Year
2022
2025
2030
2040
2050
AEO Low Fuel
Price Case
8.3
41.8
106
193
244
Central Case
AEO Reference
Fuel Price Case
8.2
40.8
102
186
234
AEO High Fuel
Price Case
7.9
39.1
96.6
172
216
Note:  These values are expressed as emission reductions, such that positive values imply an emissions decrease, and
negative values imply an emissions increase.
            Table 12.79 Annual Impacts of the MY2022-2025 Standards on Fuel Consumption
Calendar
Year
2022
2025
2030
2040
2050
AEO Low Fuel Price Case
Petroleum Gasoline
(Billon Gallons)
-0.78
-3.91
-9.91
-18.09
-22.87
Electricity
(Billion kWh)
0.28
1.35
3.44
6.28
7.81
Central Case - AEO Reference Fuel
Price Case
Petroleum Gasoline
(Billion Gallons)
-0.77
-3.82
-9.59
-17.39
-21.92
Electricity
(Billion kWh)
0.29
1.26
3.17
5.80
7.17
AEO High Price Case
Petroleum Gasoline
(Billon Gallons)
-0.75
-3.67
-9.09
-16.22
-20.29
Electricity
(Billion kWh)
0.33
1.53
3.83
6.90
8.55
Note:  These values are expressed as absolute inventory changes, such that negative values imply a decrease in
consumption, and positive values imply an increase in consumption.
           12.2.4.1    Model Year Lifetime Case Comparison Results

Table 12.80 MY Lifetime Emission Reductions of the MY2022-2025 Standards and AEO Fuel Price Cases on
                                     Total GHGs (MMT CChe)
Model Year
2021
2022
2023
2024
2025
Sum
AEO Low Fuel
Price Case
34.7
72.7
110
148
186
551
Central Case
AEO Reference
Fuel Price Case
34.2
71.3
108
144
181
538
AEO High Fuel
Price Case
32.9
69.0
104
137
172
514
Note:  The values shown in the table above are expressed as emission reductions, such that negative values imply an
emissions increase while positive values imply an emissions decrease.
                                                  12-63

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
   Table 12.81 MY Lifetime Impacts of the MY2022-2025 Standards and AEO Fuel Price Cases on Fuel
                                       Consumption
Calendar
Year
2021
2022
2023
2024
2025
Sum
AEO Low Fuel Price Case
Petroleum
Gasoline (billion
gallons)
-3.3
-6.8
-10.3
-13.8
-17.4
-51.6
Electricity
(billion
kWh)
1.3
2.3
3.4
4.6
5.9
17.6
Central Case AEO Reference Fuel
Price Case
Petroleum
Gasoline (billion
gallons)
-3.2
-6.7
-10.1
-13.4
-16.9
-50.3
Electricity
(billion
kWh)
1.4
2.3
3.2
4.2
5.3
16.4
AEO High Fuel Price Case
Petroleum
Gasoline (billion
gallons)
-3.1
-6.5
-9.7
-12.9
-16.1
-48.4
Electricity
(billion
kWh)
1.5
2.7
3.9
5.2
6.6
19.9
Note: These values are expressed as absolute inventory changes, such that negative values imply a decrease in
consumption, and positive values imply an increase in consumption.

12.3   EPA's Benefit-Cost Analysis Results

   In Section 12.3.1, EPA presents results of its model year analysis, which looks at the lifetimes
of MY2021-2025 vehicles.  In Section 12.3.2, EPA presents results of its calendar year analysis,
which looks at annual impacts through the year 2050. The inventory inputs used to generate the
monetized benefits presented here are discussed in Section 12.2.  The monetary inputs used to
generate the monetized benefits and costs presented here are discussed in Chapter 10 where we
present $/ton, $/gallon and $/mile premiums that are applied to the inventory inputs to generate
the benefit cost analysis results.

12.3.1  Model Year Analysis

   In our MY analysis, we look at the impacts over the lifetimes of MY2021-2025 vehicles.8  All
values  are discounted at 3 percent and 7 percent discount rates with the exception of the social
costs of greenhouse gases which are discounted at the discount rate used in their generation. All
values  are discounted back to CY 2015.

12.3.1.1      AEO 2015 Reference Fuel Price Case Using ICMs

   In the central analysis, we use AEO 2015 reference fuel prices and fleet projections, and, as
noted, we include our estimate of EV and PHEV sales required by the ZEV program in the
reference and control case fleets.  Importantly, Table 12.82 shows that technology and
maintenance costs are estimated at roughly $35 billion and benefits excluding fuel savings are
estimated at roughly $41 billion (using the 3 percent average SC-GHG value). In other words,
even without fuel savings, benefits outweigh costs.  Similarly, Table 12.83 shows that
technology and maintenance costs are estimated at roughly $25 billion and benefits excluding
fuel savings are estimated at roughly $30 billion (using the 3 percent average SC-GHG value).
In other words,  even without fuel savings, benefits outweigh costs.
 1 See Chapter 12.1.1.1.2 for details on why MY2021 is included in our Control Case.
                                              12-64

-------
                                          EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.82 MY Lifetime Costs & Benefits Using AEO Reference Fuel Prices and ICMs (3 Percent Discount
                                      Rate, Billions of 2013$)a'b'c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG5%Avg
SC-GHG3%Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2021
-$2.5
$0.0
$5.9
$0.3
-$0.6
$0.7
$0.5
$0.3 -$0.8

$0.3
$1.2
$1.9
$3.7

$5.0
$6.0
$6.6
$8.4
2022
-$4.7
-$0.2
$12.0
$0.6
-$1.2
$1.4
$1.0
$0.7 -$1.6

$0.6
$2.5
$3.9
$7.6

$10.7
$12.6
$14.0
$17.7
2023
-$6.8
-$0.3
$17.9
$0.9
-$1.7
$2.0
$1.4
$1.1 -$2.5

$0.9
$3.8
$5.9
$11.3

$16.2
$19.0
$21.1
$26.6
2024
-$8.8
-$0.5
$23.7
$1.3
-$2.2
$2.6
$1.9
$1.4 -$3.2

$1.2
$5.0
$7.7
$15.0

$21.5
$25.2
$28.0
$35.2
2025
-$10.8
-$0.6
$29.4
$1.6
-$2.7
$3.2
$2.3
$1.8 -$4.0

$1.4
$6.2
$9.6
$18.6

$26.8
$31.5
$35.0
$43.9
Sum
-$33.6
-$1.6
$88.8
$4.7
-$8.3
$9.8
$7.1
$5.4 -$12.1

$4.4
$18.6
$29.0
$56.1

$80.1
$94.3
$104.8
$131.8
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range  of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag.  Benefit-per-ton values were estimated for the years
2016,  2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply  to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond).  See Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5  percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2021-2025), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $13-$15; for Average SC-CO2 at 3%: $46-$50; for
Average SC-CO2 at 2.5%: $69-$75; and for 95th percentile SC-CO2 at 3%: $140-$150. For the years 2021-2025, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $620-$700; for Average SC-CH4 at 3%: $1,400-
$1,500; for Average SC-CH4 at 2.5%: $1,800-$2,000; and for 95th percentile SC-CH4 at 3%: $3,600-$4,100. For the
years 2021-2025, the SC-N2O estimates range as follows: for Average SC-N2O at 5%: $5,300-$6,000; for Average
SC-N2O at 3%: $17,000-$19,000; for Average SC-N2O at 2.5%: $25,000-$26,000; and for 95th percentile SC-N2O
at 3%: $44,000-$48,000. Chapter 10.7 also presents these SC-GHG estimates.
                                                    12-65

-------
                                          EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.83 MY Lifetime Costs & Benefits Using AEO Reference Fuel Prices and ICMs, (7 Percent Discount
                                      Rate, Billions of 2013$)a'b'c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG5%Avg
SC-GHG3%Avg
SC-GHG2.5%Avg
SC-GHG3%95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2021
-$2.0
$0.0
$3.5
$0.2
-$0.4
$0.4
$0.3
$0.2 -$0.4

$0.3
$1.2
$1.9
$3.7

$2.7
$3.6
$4.3
$6.1
2022
-$3.5
-$0.1
$7.0
$0.4
-$0.7
$0.8
$0.6
$0.4 - $0.9

$0.6
$2.5
$3.9
$7.6

$5.6
$7.5
$8.9
$12.6
2023
-$4.9
-$0.2
$10.1
$0.5
-$1.0
$1.1
$0.8
$0.6 -$1.3

$0.9
$3.8
$5.9
$11.3

$8.2
$11.1
$13.2
$18.7
2024
-$6.1
-$0.2
$12.8
$0.7
-$1.2
$1.4
$1.0
$0.7 -$1.6

$1.2
$5.0
$7.7
$15.0

$10.6
$14.4
$17.2
$24.4
2025
-$7.2
-$0.3
$15.3
$0.8
-$1.4
$1.6
$1.2
$0.9 -$1.9

$1.4
$6.2
$9.6
$18.6

$12.8
$17.6
$21.0
$30.0
Sum
-$23.8
-$0.9
$48.7
$2.6
-$4.7
$5.4
$3.9
$2.7 -$6.1

$4.4
$18.6
$29.0
$56.1

$40.0
$54.2
$64.7
$91.7
Notes
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag.  Benefit-per-ton values were estimated for the years
2016, 2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond).  See Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5 percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2021-2025), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $13-$15; for Average SC-CO2 at 3%: $46-$50; for
Average SC-CO2 at 2.5%: $69-$75; and for 95th percentile SC-CO2 at 3%: $140-$150. For the years 2021-2025, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $620-$700; for Average SC-CH4 at 3%: $1,400-
$1,500; for Average SC-CH4 at 2.5%: $1,800-$2,000; and for 95th percentile SC-CH4 at 3%: $3,600-$4,100. For the
years 2021-2025, the SC-N2O estimates range as follows: for Average SC-N2O at 5%: $5,300-$6,000; for Average
SC-N2O at 3%: $17,000-$19,000; for Average SC-N2O at 2.5%: $25,000-$26,000; and for 95th percentile SC-N2O
at 3%: $44,000-$48,000. Chapter 10.7 also presents these SC-GHG estimates.
                                                    12-66

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
12.3.1.2
AEO 2015 Reference Fuel Price Case Using RPEs
   In the central analysis, we use AEO 2015 reference fuel prices and fleet projections, and we
include our estimate of EV and PHEV sales required by the ZEV program in the reference and
control case fleets.  Importantly, Table 12.84 shows that technology and maintenance costs are
estimated at roughly $39 billion and benefits excluding fuel savings are estimated at roughly $40
billion (using the 3  percent average SC-GHG value).  In other words, even without fuel savings,
benefits outweigh costs.  Similarly, Table 12.85 shows that technology and maintenance costs
are estimated at roughly $28 billion and benefits excluding fuel savings are estimated at roughly
$30 billion (using the 3 percent average SC-GHG value).  In other words,  even without fuel
savings, benefits outweigh costs.

 Table 12.84 MY Lifetime Costs & Benefits Using AEO Reference Fuel Prices and RPEs (3 Percent Discount
                                    Rate, Billions of 2013$)a'b'c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2021
-$2.6
$0.0
$5.9
$0.3
-$0.6
$0.7
$0.5
$0.3 - $0.7

$0.3
$1.2
$1.9
$3.6

$4.9
$5.8
$6.5
$8.3
2022
-$5.2
-$0.2
$12.0
$0.6
-$1.2
$1.4
$1.0
$0.7 -$1.6

$0.6
$2.5
$3.9
$7.5

$10.2
$12.1
$13.5
$17.1
2023
-$7.6
-$0.3
$17.9
$0.9
-$1.7
$2.0
$1.4
$1.1 -$2.4

$0.9
$3.7
$5.8
$11.3

$15.3
$18.1
$20.2
$25.7
2024
-$10.0
-$0.5
$23.7
$1.3
-$2.2
$2.6
$1.9
$1.4 -$3.2

$1.2
$4.9
$7.7
$14.9

$20.2
$24.0
$26.8
$34.0
2025
-$12.2
-$0.6
$29.4
$1.6
-$2.7
$3.1
$2.3
$1.8 -$3.9

$1.4
$6.1
$9.6
$18.5

$25.2
$29.9
$33.4
$42.3
Sum
-$37.6
-$1.6
$88.8
$4.7
-$8.3
$9.8
$7.1
$5.3 -$11.8

$4.4
$18.5
$28.9
$55.8

$75.9
$90.1
$100.4
$127.4
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al., 2012). The
range of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag. Benefit-per-ton values were estimated for the years
2016, 2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond).  See Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5 percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2021-2025), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $13-$15; for Average SC-CO2 at 3%: $46-$50; for
Average SC-CO2 at 2.5%: $69-$75; and for 95th percentile SC-CO2 at 3%: $140-$150. For the years 2021-2025,  the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $620-$700; for Average SC-CH4 at 3%: $1,400-
$1,500; for Average SC-CH4 at 2.5%: $1,800-$2,000; and for 95th percentile SC-CH4 at 3%: $3,600-$4,100. For the
                                                  12-67

-------
                                          EPA's Analysis of the MY2022-2025 GHG Standards
years 2021-2025, the SC-N2O estimates range as follows: for Average SC-N2O at 5%: $5,300-$6,000; for Average
SC-N2O at 3%: $17,000-$19,000; for Average SC-N2O at 2.5%: $25,000-$26,000; and for 95th percentile SC-N2O
at 3%: $44,000-$48,000. Chapter 10.7 also presents these SC-GHG estimates.
 Table 12.85 MY Lifetime Costs & Benefits Using AEO Reference Fuel Prices and RPEs, (7 Percent Discount
                                      Rate, Billions of 2013$)a- b>c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2021
-$2.0
$0.0
$3.5
$0.2
-$0.4
$0.4
$0.3
$0.2 - $0.4

$0.3
$1.2
$1.9
$3.6

$2.6
$3.5
$4.2
$5.9
2022
-$3.9
-$0.1
$7.0
$0.4
-$0.7
$0.8
$0.6
$0.4 -$0.8

$0.6
$2.5
$3.9
$7.5

$5.3
$7.2
$8.5
$12.2
2023
-$5.5
-$0.2
$10.1
$0.5
-$1.0
$1.1
$0.8
$0.5 -$1.2

$0.9
$3.7
$5.8
$11.3

$7.6
$10.5
$12.6
$18.0
2024
-$6.9
-$0.3
$12.8
$0.7
-$1.2
$1.4
$1.0
$0.7 -$1.6

$1.2
$4.9
$7.7
$14.9

$9.8
$13.6
$16.4
$23.5
2025
-$8.2
-$0.3
$15.3
$0.8
-$1.4
$1.6
$1.2
$0.8 -$1.9

$1.4
$6.1
$9.6
$18.5

$11.8
$16.5
$20.0
$28.9
Sum
-$26.6
-$0.9
$48.7
$2.6
-$4.6
$5.4
$3.9
$2.6 -$5.9

$4.4
$18.5
$28.9
$55.8

$37.1
$51.3
$61.7
$88.6
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag.  Benefit-per-ton values were estimated for the years
2016, 2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond).  See Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5 percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2021-2025), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $13-$15; for Average SC-CO2 at 3%: $46-$50; for
Average SC-CO2 at 2.5%: $69-$75; and for 95th percentile SC-CO2 at 3%: $140-$150. For the years 2021-2025, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $620-$700; for Average SC-CH4 at 3%: $1,400-
$1,500; for Average SC-CH4 at 2.5%: $1,800-$2,000; and for 95th percentile SC-CH4 at 3%: $3,600-$4,100. For the
years 2021-2025, the SC-N2O estimates range as follows: for Average SC-N2O at 5%: $5,300-$6,000; for Average
SC-N2O at 3%: $17,000-$19,000; for Average SC-N2O at 2.5%: $25,000-$26,000; and for 95th percentile SC-N2O
at 3%: $44,000-$48,000. Chapter 10.7 also presents these SC-GHG estimates.
                                                    12-68

-------
                                    EPA's Analysis of the MY2022-2025 GHG Standards
12.3.1.3      AEO 2015 High Fuel Price Case Using ICMs

   In the AEO high fuel price analysis, we use AEO 2015 high fuel prices and fleet projections,
and we include our estimate of EV and PHEV sales required by the ZEV program in the
reference and control case fleets. Importantly, Table 12.86 shows that technology and
maintenance costs are estimated at roughly $32 billion and benefits excluding fuel savings are
estimated at roughly $36 billion (using the 3 percent average SC-GHG value).  In other words,
even without fuel savings, benefits outweigh costs.  Similarly, Table 12.87 shows that
technology and maintenance costs are estimated at roughly $22 billion and benefits excluding
fuel savings are estimated at roughly $27 billion (using the 3 percent average SC-GHG value).
In other words, even without fuel savings, benefits outweigh costs.
                                             12-69

-------
                                          EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.86 MY Lifetime Costs & Benefits Using AEO High Fuel Prices and ICMs (3 Percent Discount Rate,
                                        Billions of 2013$) a'b'c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG5%Avg
SC-GHG3%Avg
SC-GHG2.5%Avg
SC-GHG3%95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2021
-$2.4
$0.0
$5.6
$0.3
-$0.6
$0.7
$0.5
$0.3 - $0.7

$0.3
$1.2
$1.8
$3.5

$4.9
$5.8
$6.5
$8.2
2022
-$4.6
-$0.1
$11.7
$0.6
-$1.2
$1.3
$0.9
$0.7 -$1.6

$0.6
$2.4
$3.8
$7.3

$10.4
$12.2
$13.6
$17.1
2023
-$6.7
-$0.2
$17.3
$0.9
-$1.7
$1.9
$1.4
$1.0 -$2.3

$0.9
$3.6
$5.6
$10.9

$15.4
$18.2
$20.2
$25.5
2024
-$8.7
-$0.4
$22.7
$1.2
-$2.2
$2.5
$1.8
$1.3 -$3.0

$1.1
$4.7
$7.4
$14.3

$20.3
$23.9
$26.6
$33.5
2025
-$10.6
-$0.5
$28.1
$1.5
-$2.6
$3.0
$2.2
$1.7 -$3.7

$1.4
$5.8
$9.2
$17.7

$25.2
$29.6
$32.9
$41.5
Sum
-$30.6
-$1.2
$79.8
$4.2
-$7.7
$8.8
$6.4
$4.7 -$10.6

$3.9
$16.6
$26.0
$50.1

$71.3
$84.0
$93.3
$117.5
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range  of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag.  Benefit-per-ton values were estimated for the years
2016,  2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply  to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond).  See Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5  percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2021-2025), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $13-$15; for Average SC-CO2 at 3%: $46-$50; for
Average SC-CO2 at 2.5%: $69-$75; and for 95th percentile SC-CO2 at 3%: $140-$150. For the years 2021-2025, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $620-$700; for Average SC-CH4 at 3%: $1,400-
$1,500; for Average SC-CH4 at 2.5%: $1,800-$2,000; and for 95th percentile SC-CH4 at 3%: $3,600-$4,100. For the
years 2021-2025, the SC-N2O estimates range as follows: for Average SC-N2O at 5%: $5,300-$6,000; for Average
SC-N2O at 3%: $17,000-$19,000; for Average SC-N2O at 2.5%: $25,000-$26,000; and for 95th percentile SC-N2O
at 3%: $44,000-$48,000. Chapter 10.7 also presents these SC-GHG estimates.
                                                    12-70

-------
                                         EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.87 MY Lifetime Costs & Benefits Using AEO High Fuel Prices and ICMs (7 Percent Discount Rate,
                                       Billions of 2013$) a'b'c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG5%Avg
SC-GHG3%Avg
SC-GHG2.5%Avg
SC-GHG3%95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2021
-$1.9
$0.0
$3.4
$0.2
-$0.4
$0.4
$0.3
$0.2 - $0.4

$0.3
$1.2
$1.8
$3.5

$2.7
$3.5
$4.2
$5.9
2022
-$3.5
-$0.1
$6.8
$0.4
-$0.7
$0.8
$0.6
$0.4 -$0.8

$0.6
$2.4
$3.8
$7.3

$5.4
$7.3
$8.6
$12.2
2023
-$4.9
-$0.1
$9.7
$0.5
-$1.0
$1.1
$0.8
$0.5 -$1.2

$0.9
$3.6
$5.6
$10.9

$7.9
$10.6
$12.6
$17.9
2024
-$6.0
-$0.2
$12.3
$0.6
-$1.2
$1.4
$1.0
$0.7 -$1.5

$1.1
$4.7
$7.4
$14.3

$10.0
$13.7
$16.3
$23.2
2025
-$7.1
-$0.3
$14.6
$0.8
-$1.4
$1.6
$1.2
$0.8 -$1.8

$1.4
$5.8
$9.2
$17.7

$12.0
$16.5
$19.8
$28.3
Sum
-$21.5
-$0.7
$43.5
$2.3
-$4.3
$4.8
$3.5
$2.3 -$5.3

$3.9
$16.6
$26.0
$50.1

$35.4
$48.1
$57.5
$81.6
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag. Benefit-per-ton values were estimated for the years
2016, 2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond).  See Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5 percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2021-2025), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $13-$15; for Average SC-CO2 at 3%: $46-$50; for
Average SC-CO2 at 2.5%: $69-$75; and for 95th percentile SC-CO2 at 3%: $140-$150. For the years 2021-2025, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $620-$700; for Average SC-CH4 at 3%: $1,400-
$1,500; for Average SC-CH4 at 2.5%: $1,800-$2,000; and for 95th percentile SC-CH4  at 3%: $3,600-$4,100. For the
years 2021-2025, the SC-N2O estimates range as follows: for Average SC-N2O at 5%:  $5,300-$6,000; for Average
SC-N2O at 3%: $17,000-$19,000; for Average SC-N2O at 2.5%: $25,000-$26,000; and for 95th percentile SC-N2O
at 3%: $44,000-$48,000. Chapter 10.7 also presents these SC-GHG estimates.
12.3.1.4       AEO 2015 Low Fuel Price Case Using ICMs

   In the AEO low fuel price analysis, we use AEO 2015 low fuel prices and fleet projections,
and we include our estimate of EV and PHEV sales required by the ZEV program in the
reference and control case fleets. Importantly, Table 12.88 shows that technology and
maintenance costs are estimated at roughly $33 billion and benefits excluding fuel savings are
estimated at roughly $39 billion (using the 3 percent average  SC-GHG value). In other words,
                                                  12-71

-------
                                         EPA's Analysis of the MY2022-2025 GHG Standards
even without fuel savings, benefits outweigh costs.  Similarly, Table 12.89 shows that
technology and maintenance costs are estimated at roughly $23 billion and benefits excluding
fuel savings are estimated at roughly $29 billion (using the 3 percent average SC-GHG value).
In other words, even without fuel savings, benefits outweigh costs.

 Table 12.88 MY Lifetime Costs & Benefits Using AEO Low Fuel Prices and ICMs (3 Percent Discount Rate,
                                        Billions of 2013$) a'b'c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2021
-$2.4
-$0.1
$5.9
$0.3
-$0.6
$0.7
$0.5
$0.4 -$0.8

$0.3
$1.2
$1.9
$3.7

$5.3
$6.2
$6.9
$8.7
2022
-$4.6
-$0.2
$12.2
$0.6
-$1.2
$1.4
$1.0
$0.7 -$1.7

$0.6
$2.6
$4.0
$7.7

$11.1
$13.1
$14.5
$18.2
2023
-$6.8
-$0.3
$18.3
$1.0
-$1.7
$2.1
$1.5
$1.1 -$2.5

$0.9
$3.8
$6.0
$11.6

$16.7
$19.7
$21.8
$27.4
2024
-$8.9
-$0.5
$24.3
$1.3
-$2.2
$2.7
$1.9
$1.5 -$3.3

$1.2
$5.1
$8.0
$15.4

$22.3
$26.2
$29.0
$36.4
2025
-$10.9
-$0.6
$30.3
$1.6
-$2.7
$3.2
$2.4
$1.8 -$4.1

$1.5
$6.3
$9.9
$19.1

$27.8
$32.7
$36.3
$45.5
Sum
-$31.2
-$1.5
$85.2
$4.5
-$7.8
$9.4
$6.8
$5.2 -$11.6

$4.2
$17.8
$27.8
$53.8

$77.9
$91.6
$101.6
$127.5
Note:a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter
10.6 for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al., 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag.  Benefit-per-ton values were estimated for the years
2016, 2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond). See  Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The  same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5 percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2021-2025), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $13-$15; for Average SC-CO2 at 3%: $46-$50; for
Average SC-CO2 at 2.5%: $69-$75; and for 95th percentile SC-CO2 at 3%: $140-$150. For the years 2021-2025, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $620-$700; for Average SC-CH4 at 3%: $1,400-
$1,500; for Average SC-CH4 at 2.5%: $1,800-$2,000; and for 95th percentile SC-CH4 at 3%: $3,600-$4,100. For the
years 2021-2025, the SC-N2O estimates range as follows: for Average SC-N2O  at 5%: $5,300-$6,000; for Average
SC-N2O at 3%: $17,000-$19,000; for Average SC-N2O at 2.5%: $25,000-$26,000; and for 95th percentile SC-N2O
at 3%: $44,000-$48,000. Chapter 10.7 also presents these SC-GHG estimates.
                                                   12-72

-------
                                          EPA's Analysis of the MY2022-2025 GHG Standards
 Table 12.89 MY Lifetime Costs & Benefits Using AEO Low Fuel Prices and ICMs (7 Percent Discount Rate,
                                         Billions of 2013$) a'b'c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG5%Avg
SC-GHG3%Avg
SC-GHG2.5%Avg
SC-GHG3%95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2021
-$1.9
$0.0
$3.6
$0.2
-$0.4
$0.4
$0.3
$0.2 - $0.4

$0.3
$1.2
$1.9
$3.7

$2.9
$3.8
$4.5
$6.3
2022
-$3.5
-$0.1
$7.1
$0.4
-$0.7
$0.8
$0.6
$0.4 -$0.9

$0.6
$2.6
$4.0
$7.7

$5.9
$7.8
$9.2
$13.0
2023
-$4.9
-$0.2
$10.3
$0.5
-$1.0
$1.2
$0.8
$0.6 -$1.3

$0.9
$3.8
$6.0
$11.6

$8.6
$11.5
$13.6
$19.2
2024
-$6.2
-$0.2
$13.1
$0.7
-$1.2
$1.4
$1.1
$0.7 -$1.6

$1.2
$5.1
$8.0
$15.4

$11.1
$15.0
$17.8
$25.2
2025
-$7.3
-$0.3
$15.8
$0.8
-$1.4
$1.7
$1.3
$0.9 - $2.0

$1.5
$6.3
$9.9
$19.1

$13.4
$18.2
$21.8
$31.0
Sum
-$21.8
-$0.9
$46.3
$2.4
-$4.4
$5.1
$3.7
$2.6 -$5.8

$4.2
$17.8
$27.8
$53.8

$38.9
$52.5
$62.5
$88.4
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range  of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag.  Benefit-per-ton values were estimated for the years
2016,  2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply  to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond).  See Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5  percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2021-2025), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $13-$15; for Average SC-CO2 at 3%: $46-$50; for
Average SC-CO2 at 2.5%: $69-$75; and for 95th percentile SC-CO2 at 3%: $140-$150. For the years 2021-2025, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $620-$700; for Average SC-CH4 at 3%: $1,400-
$1,500; for Average SC-CH4 at 2.5%: $1,800-$2,000; and for 95th percentile SC-CH4 at 3%: $3,600-$4,100. For the
years 2021-2025, the SC-N2O estimates range as follows: for Average SC-N2O at 5%: $5,300-$6,000; for Average
SC-N2O at 3%: $17,000-$19,000; for Average SC-N2O at 2.5%: $25,000-$26,000; and for 95th percentile SC-N2O
at 3%: $44,000-$48,000. Chapter 10.7 also presents these SC-GHG estimates.
                                                    12-73

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
12.3.1.5
Summary of MY Lifetime Benefit-Cost Analysis Results
   The table below summarizes EPA's MY lifetime BCA results. Importantly, the fuel savings
do not vary in the AEO 2015 reference fuel price case regardless of choice of ICM or RPE since
these metrics are tied directly to standard level targets (rather than achieved) values.  The slight
variations that do exist in the benefits category in the AEO 2015 reference fuel price case is the
result of slightly different projected EV/PHEV penetration above and beyond the ZEV program.
The different penetrations result in different electricity demands and, therefore, different
upstream emission impacts.  The differences in all categories when comparing across fuel  price
cases are the result of the different fleet makeups across fuel prices, different ZEV program sales
projections across fuel prices cases, and the different fuel prices themselves.
      Table 12.90 MY Lifetime Costs &  Benefits in the Central & Sensitivity Cases (Billions of 2013$)

Vehicle Program
Maintenance
Fuel
Benefits
Net Benefits
3 Percent Discount Rate
AEO Low
(ICMs)
-$31.2
-$1.5
$85.2
$39.1
$91.6
AEO Ref
(ICMs&RPEs)
-$37.6 to -$33.6
-$1.6 to -$1.6
$88.8 to $88.8
$40.4 to $40.7
$90.1 to $94.3
AEO High
(ICMs)
-$30.6
-$1.2
$79.8
$36.0
$84.0
7 Percent Discount Rate
AEO Low
(ICMs)
-$21.8
-$0.9
$46.3
$28.9
$52.5
AEO Ref
(ICMs&RPEs)
-$26.6 to -$23.8
-$0.9 to -$0.9
$48.7 to $48.7
$30.0 to $30.2
$51.3 to $54.2
AEO High
(ICMs)
-$21.5
-$0.7
$43.5
$26.8
$48.1
Note: AEO Reference fuel price case shows ranges generated using both ICMs and RPEs in calculating indirect
technology costs; Benefits and Net Benefits values presented here use the mid-point value of the non-GHG range for
the applicable discount rate and the central SC-GHG values (average SC-CCh, average SC-CH4, average SC-N2O,
each at 3 percent) discounted at 3 percent in all cases.

   Importantly, Table 12.90 shows that, in all cases, the net benefits are greater than the fuel
savings. In other words,  even excluding fuel savings, the benefits of the standards outweigh the
costs.  It is also important to note in the table above that the net benefits are actually lowest in the
high fuel price case.  This is counterintuitive.  This result  is driven by the lower share of trucks
projected in the high fuel price case whereas the low fuel  price case has a higher share of trucks.
Trucks drive more miles  so, in general, more trucks in the fleet results in more GHG and fuel
reductions (and associated fuel savings) and, thus, more net benefits. Fewer trucks, as in the
high fuel price case, results in fewer net benefits.  Importantly, EPA would not suggest that to
maximize net benefits we should all buy trucks. Instead, the analysis projects those relatively
higher net benefits in a world consisting of such a high share of trucks.  If the car/truck mix is
not so  dependent upon fuel price as estimated in AEO 2015 (i.e., if the low fuel price case had a
fleet mix like that of the reference or high fuel price case), then the net benefits of the low fuel
price case would be lower, as one might initially expect.
12.3.2  Calendar Year Analysis

   In our calendar year (CY) analysis, EPA looks  at the impacts year-over-year through the year
2050.  All annual values are presented without discounting and the stream of values for the years
2021 through 2050 are then discounted back to the year 2015 at both 3 and 7 percent discount
rates, with the exception that all social costs of greenhouse gases are discounted at the discount
rate used in their generation.
                                              12-74

-------
                                         EPA's Analysis of the MY2022-2025 GHG Standards
12.3.2.1
AEO 2015 Reference Fuel Price Case Using ICMs
   In the central analysis, we use AEO 2015 reference fuel prices and fleet projections, and we
include our estimate of EV and PHEV sales required by the ZEV program in the reference case
fleet.

  Table 12.91 Annual Costs & Benefits Using AEO Reference Fuel Prices and ICMs (Billions of 2013$)a-b-c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2025
-$14.7
-$0.2
$9.8
$0.5
-$1.0
$1.1
$0.8
$0.6 -$1.5

$0.4
$1.5
$2.4
$4.5

-$2.3
-$1.1
-$0.3
$1.9
2030
-$14.8
-$0.5
$27.0
$1.4
-$2.6
$2.9
$2.2
$1.6 -$4.0

$0.9
$3.6
$5.6
$10.8

$19.3
$22.0
$24.0
$29.2
2040
-$16.8
-$1.0
$61.6
$3.4
-$4.7
$6.5
$4.4
$2.9 -$7.2

$1.3
$5.8
$9.2
$17.7

$59.6
$64.2
$67.5
$76.0
2050
-$18.8
-$1.3
$77.6
$4.3
-$5.9
$8.1
$6.2
$3.6 -$9.0

$1.2
$6.3
$10.2
$19.2

$77.7
$82.8
$86.7
$95.7
NPV, 3%
-$240.5
-$10.7
$611.3
$33.5
-$50.3
$64.9
$46.9
$34.2 -$76.4

$27.9
$128.9
$204.7
$392.4

$538.3
$639.2
$715.1
$902.8
NPV, 7%
-$114.8
-$4.4
$248.0
$13.5
-$21.0
$26.5
$19.2
$12.7 -$28.5

$27.9
$128.9
$204.7
$392.4

$215.5
$316.4
$392.3
$580.0
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag.  Benefit-per-ton values were estimated for the years
2016, 2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond). See  Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5 percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2025-2050), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $15-$26; for Average SC-CO2 at 3%: $50-$69; for
Average SC-CO2 at 2.5%: $75-$95; and for 95th percentile SC-CO2 at 3%: $150-$210. For the years 2025-2050, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $710-$1400; for Average SC-CH4 at 3%: $1,500-
$2,700; for Average SC-CH4 at 2.5%: $2,000-$3,400; and for 95th percentile SC-CH4 at 3%: $4,100-$7,300. For the
years 2025-2050, the SC-N2O estimates range as follows: for Average SC-N2O at 5%: $6,000-$ 12,000; for Average
SC-N2O at 3%: $19,000-$30,000; for Average SC-N2O at 2.5%: $26,000-$41,000; and for 95th percentile SC-N2O
at 3%: $48,000-$79,000. Chapter 10.7 also presents these SC-GHG estimates.
                                                   12-75

-------
                                         EPA's Analysis of the MY2022-2025 GHG Standards
12.3.2.2       AEO 2015 Reference Fuel Price Case Using RPEs

  Table 12.92 Annual Costs & Benefits Using AEO Reference Fuel Prices and RPEs (Billions of 2013$)a-b-c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG5%Avg
SC-GHG3%Avg
SC-GHG2.5%Avg
SC-GHG3%95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2025
-$16.7
-$0.2
$9.8
$0.5
-$1.0
$1.1
$0.8
$0.6 -$1.4

$0.4
$1.5
$2.3
$4.5

-$4.3
-$3.2
-$2.3
-$0.2
2030
-$16.9
-$0.6
$27.1
$1.4
-$2.6
$2.9
$2.2
$1.6 -$3.9

$0.9
$3.6
$5.6
$10.8

$17.2
$19.9
$21.9
$27.1
2040
-$19.0
-$1.0
$61.6
$3.4
-$4.7
$6.5
$4.4
$2.8 -$7.0

$1.3
$5.8
$9.2
$17.6

$57.2
$61.7
$65.1
$73.6
2050
-$21.4
-$1.3
$77.6
$4.3
-$5.9
$8.1
$6.2
$3.6 -$8.9

$1.2
$6.3
$10.2
$19.2

$75.0
$80.0
$84.0
$92.9
NPV, 3%
-$272.8
-$11.0
$611.4
$33.5
-$50.2
$64.8
$46.9
$33.5 -$74.9

$27.8
$128.4
$204.1
$391.2

$504.7
$605.3
$681.0
$868.0
NPV, 7%
-$130.0
-$4.5
$248.1
$13.5
-$20.9
$26.4
$19.2
$12.5 -$27.9

$27.8
$128.4
$204.1
$391.2

$199.7
$300.3
$376.0
$563.0
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al., 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag. Benefit-per-ton values were estimated for the years
2016, 2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond).  See Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5 percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to  the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2025-2050), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $15-$26; for Average SC-CO2 at 3%: $50-$69; for
Average SC-CO2 at 2.5%: $75-$95; and for 95th percentile SC-CO2 at 3%: $150-$210. For the years 2025-2050, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $710-$1400; for Average SC-CH4 at 3%: $1,500-
$2,700; for Average SC-CH4 at 2.5%: $2,000-$3,400; and for 95th percentile SC-CH4 at 3%: $4,100-$7,300. For the
years 2025-2050, the  SC-N2O estimates range as follows: for Average SC-N2O at 5%: $6,000-$ 12,000; for Average
SC-N2O at 3%: $19,000-$30,000; for Average SC-N2O at 2.5%: $26,000-$41,000; and for 95th percentile SC-N2O
at 3%: $48,000-$79,000. Chapter 10.7 also presents these SC-GHG estimates.
12.3.2.3
AEO 2015 High Fuel Price Case Using ICMs
   In the AEO high fuel price analysis, we use AEO 2015 high fuel prices and fleet projections,
and we include our estimate of EV and PHEV sales required by the ZEV program in the
reference case fleet.
                                                   12-76

-------
                                         EPA's Analysis of the MY2022-2025 GHG Standards
    Table 12.93 Annual Costs & Benefits Using AEO High Fuel Prices and ICMs (Billions of 2013$)a'b'c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2025
-$14.4
-$0.2
$9.4
$0.5
-$1.0
$1.1
$0.8
$0.6 -$1.4

$0.4
$1.5
$2.3
$4.4

-$2.2
-$0.9
$0.1
$3.1
2030
-$14.1
-$0.4
$25.6
$1.4
-$2.5
$2.8
$2.1
$1.5 -$3.7

$0.8
$3.4
$5.3
$10.2

$19.2
$22.8
$25.2
$33.6
2040
-$15.9
-$0.8
$57.4
$3.2
-$4.5
$6.0
$4.1
$2.6 -$6.5

$1.2
$5.4
$8.5
$16.4

$58.2
$65.6
$70.1
$88.9
2050
-$17.9
-$1.1
$71.8
$4.0
-$5.6
$7.5
$5.8
$3.2 -$8.1

$1.1
$5.8
$9.4
$17.7

$76.7
$86.9
$93.0
$120.9
NPV, 3%
-$230.4
-$8.8
$572.0
$31.4
-$48.7
$60.6
$43.9
$31.0 -$69.3

$26.0
$119.9
$190.5
$365.1

$536.4
$630.3
$700.9
$875.5
NPV, 7%
-$110.2
-$3.6
$232.7
$12.7
-$20.4
$24.8
$18.0
$11.6 -$25.9

$26.0
$119.9
$190.5
$365.1

$213.6
$307.5
$378.1
$552.7
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al, 2012). The
range of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag.  Benefit-per-ton values were estimated for the years
2016, 2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond). See  Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5 percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time. Corresponding to the years in this table (2025-2050), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%: $15-$26; for Average SC-CO2 at 3%: $50-$69; for
Average SC-CO2 at 2.5%: $75-$95; and for 95th percentile SC-CO2 at 3%: $150-$210. For the years 2025-2050, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%: $710-$1400; for Average SC-CH4 at 3%: $1,500-
$2,700; for Average SC-CH4 at 2.5%: $2,000-$3,400; and for 95th percentile SC-CH4 at 3%: $4,100-$7,300. For the
years 2025-2050, the SC-N2O estimates range as follows: for Average SC-N2O  at 5%: $6,000-$ 12,000; for Average
SC-N2O at 3%: $19,000-$30,000; for Average SC-N2O at 2.5%: $26,000-$41,000; and for 95th percentile SC-N2O
at 3%: $48,000-$79,000. Chapter 10.7 also presents these SC-GHG estimates.
12.3.2.4
AEO 2015 Low Fuel Price Case Using ICMs
   In the AEO low fuel price analysis, we use AEO 2015 low fuel prices and fleet projections,
and we include our estimate of EV and PHEV sales required by the ZEV program in the
reference case fleet.
                                                   12-77

-------
                                         EPA's Analysis of the MY2022-2025 GHG Standards
     Table 12.94 Annual Costs & Benefits Using AEO Low Fuel Prices and RPEs (Billions of 2013$)a'b'c

Vehicle Program
Maintenance
Pre-tax Fuel
Energy Security
Crashes, Noise, Congestion
Travel Value
Refueling
Non-GHG
GHG
SC-GHG5%Avg
SC-GHG3%Avg
SC-GHG2.5%Avg
SC-GHG3%95th
Net Benefits
SC-GHG 5% Avg
SC-GHG 3% Avg
SC-GHG 2.5% Avg
SC-GHG 3% 95th
2025
-$14.9
-$0.2
$10.1
$0.5
-$1.0
$1.1
$0.8
$0.6 -$1.5

$0.4
$1.6
$2.4
$4.7

-$1.8
-$0.4
$0.6
$3.9
2030
-$15.2
-$0.5
$28.0
$1.5
-$2.6
$3.0
$2.3
$1.6 -$4.1

$0.9
$3.7
$5.8
$11.2

$21.3
$25.2
$27.9
$37.1
2040
-$17.2
-$1.0
$64.0
$3.6
-$4.8
$6.7
$4.6
$3.0 -$7.4

$1.3
$6.0
$9.5
$18.4

$65.9
$74.1
$79.2
$100.3
2050
-$19.3
-$1.3
$80.9
$4.5
-$6.0
$8.5
$6.5
$3.8 -$9.4

$1.3
$6.5
$10.7
$20.0

$87.8
$99.3
$106.3
$137.8
NPV, 3%
-$245.2
-$10.7
$634.8
$34.8
-$51.3
$67.3
$48.7
$35.3 -$79.0

$29.0
$133.7
$212.5
$407.2

$539.4
$644.1
$722.9
$917.6
NPV, 7%
-$116.8
-$4.4
$257.3
$14.0
-$21.4
$27.4
$19.9
$13.2 -$29.4

$29.0
$133.7
$212.5
$407.2

$216.5
$321.3
$400.1
$594.8
Notes:
a The non-GHG benefits presented in this table are based on PM2.5-related benefit per ton values (see Chapter 10.6
for more information); the range of benefits are derived from two premature mortality estimates - the American
Cancer Society cohort study (Krewski et al, 2009) and the Harvard Six-Cities study (Lepeule et al, 2012).  The
range of benefits also assumes either a 3 percent or 7 percent discount rate in the valuation of PM-related premature
mortality to account for a twenty-year segmented cessation lag. Benefit-per-ton values were estimated for the years
2016, 2020, 2025 and 2030. We hold values constant for intervening years (e.g., the 2016 values are assumed to
apply to years 2017-2019; 2020 values foryears 2021-2024; 2025 values foryears 2026-2029; and 2030 values for
years 2031 and beyond).  See Table 10-10 for the benefit per ton values used in this analysis.
b GHG benefit estimates include reductions in CO2, CH4, and N2O but do not include HFC reductions. Note that net
present value of reduced GHG emissions is calculated differently than other benefits. The same discount rate used to
discount the value of damages from future emissions (SC-CO2, SC-CH4, and SC-N2O, each discounted at rates of 5,
3, 2.5 percent) is used to calculate net present value of SC-CO2, SC-CH4, and SC-N2O, respectively, for internal
consistency. Refer to the Chapter 10.7 for more detail.
0 Chapter 10.7 notes that SC-GHGs increases over time.  Corresponding to the years in this table (2025-2050), the
SC-CO2 estimates range as follows: for Average SC-CO2 at 5%:  $15-$26; for Average SC-CO2 at 3%:  $50-$69; for
Average SC-CO2 at 2.5%: $75-$95; and for 95th percentile SC-CO2 at 3%: $150-$210. For the years 2025-2050, the
SC-CH4 estimates range as follows: for Average SC-CH4 at 5%:  $710-$1400; for Average SC-CH4 at 3%: $1,500-
$2,700; for Average SC-CH4 at 2.5%: $2,000-$3,400; and for 95th percentile SC-CH4 at 3%: $4,100-$7,300. For the
years 2025-2050, the SC-N2O estimates range as follows: for Average SC-N2O at 5%: $6,000-$ 12,000; for Average
SC-N2O at 3%: $19,000-$30,000; for Average SC-N2O at 2.5%:  $26,000-$41,000; and for 95th percentile SC-N2O
at 3%: $48,000-$79,000. Chapter 10.7 also presents these SC-GHG estimates.
12.3.2.5
Summary of CY Benefit-Cost Analysis Results
   In our CY analysis, EPA looks at the impacts year-over-year through the year 2050.  All
annual values are discounted back to the year 2015 at both 3 and 7 percent discount rates with
the exception that all social costs of greenhouse gases are discounted at the discount rate used in
their generation.  The table below simply summarizes the net present values presented in the
calendar year analysis tables above.
                                                   12-78

-------
                                       EPA's Analysis of the MY2022-2025 GHG Standards
  Table 12.95 CY Net Present Value Costs & Benefits in the Central & Sensitivity Cases (Billions of 2013$)

Vehicle Program
Maintenance
Fuel
Benefits
Net Benefits
3 Percent Discount Rate
AEO Low
(ICMs)
-$245.2
-$10.7
$634.8
$265.2
$644.1
AEO Ref
(ICMs&RPEs)
-$272.8 to -$240.5
-$11.0 to -$10.7
$611.3 to $611.4
$277. 7 to $279. 2
$605.3 to $639.2
AEO High
(ICMs)
-$230.4
-$8.8
$572.0
$297.4
$630.3
7 Percent Discount Rate
AEO Low
(ICMs)
-$116.8
-$4.4
$257.3
$185.2
$321.3
AEO Ref
(ICMs&RPEs)
-$130.0 to -$114.8
-$4.5 to -$4.4
$248.0 to $248.1
$186.8 to $187.6
$300.3 to $316.4
AEO High
(ICMs)
-$110.2
-$3.6
$232.7
$188.6
$307.5
Note:  AEO Reference fuel price case shows ranges generated using both ICMs and RPEs in calculating indirect
technology costs; Benefits and Net Benefits values presented here use the mid-point value of the non-GHG range for
the applicable discount rate and the central SC-GHG values (average SC-CCh, average SC-CH4, average SC-N2O,
each at 3 percent) discounted at 3 percent in all cases.

   As noted above in our MY analysis summary, it is important to note in the table above that the
net benefits are actually lowest in the high fuel price case.  This is counterintuitive.  This result is
driven by the lower share of trucks projected in the high fuel price case whereas the low fuel
price case has a higher share of trucks.

12.4  Additional OMEGA Cost Analyses

12.4.1 Cost per Vehicle Tables - Absolute and Incremental Costs

   EPA presents absolute costs for MY2025 vehicles meeting the 2021 standards (i.e., the
reference case) in Table 12.96, and for MY2025 vehicles meeting the 2025 standards (i.e., the
central analysis control case), for cars, trucks and the fleet in Table 12.97. These costs are then
compared and shown as the delta, or the incremental costs of the 2025 standards relative to the
2021  standards in MY2025. In these two tables, the absolute costs shown represent costs to
bring the projected MY2021 and MY2025 fleets into compliance with the indicated standard.  In
other words, the costs include costs that will be incurred to comply with 2015 and later MY
standards.1 Of primary interest for this analysis are the incremental costs shown in Table 12.96
and Table  12.97.  These tables present the incremental costs to comply with the control case
standards relative to meeting the reference case standards (i.e., the MY2021 standards).
T Interestingly, the absolute costs include roughly $50 to bring the projected MY2025 fleet into compliance with the
  2014 standards; in other words, the standards in place for the fleet upon which our baseline fleet is derived. This
  $50 is the result of market shifts projected to take place between MY2014 and MY2025 - those projections,
  based on AEO2015, are for a higher percentage of trucks in MY2025. The point being that, while our baseline
  fleet is derived from the MY2014 fleet, the absolute costs in our analysis include future costs just to ensure that
  the projected fleet complies with the 2014 standards.
                                                12-79

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
    Table 12.96  MY2021 Absolute and Incremental Costs per Vehicle in the Central Analysis Using AEO
        Reference Case Fuel Prices and Fleet Projections and Using both ICMs and RPEs (2013$)

Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-
Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Reference Case in MY2021
Car
$814-$888
$1015-
$1182
$501-$575
$654-$773
$363-$452
$813-$887
$3048-
$2973
$303-$409
$1567-
$1858
$541-$640
$526-$619
$307-$359
$155-$155
$418-$491
$1807-
$1980
$1288-
$1543
$697-$800
Truck
$1147-
$1283
$1221-
$1458
$703-$887
$800-$915
$543-$609
$927-$1234
$2321-
$2465
$552-$534
$1702-
$1735
$745-$726
$774-$948
$192-$282
$0-$0
$741-$912
$1606-
$1611
$1899-
$2023
$869-$1019
Combined
$904-$994
$1160-
$1376
$621-$760
$730-$847
$450-$528
$827-$930
$2459-
$2561
$378-$447
$1623-
$1807
$619-$673
$630-$758
$218-$299
$155-$155
$571-$691
$1728-
$1834
$1614-
$1799
$782-$908
Control Case in MY2021
Car
$1237-
$1290
$1079-
$1307
$606-$764
$869-$958
$431-$527
$957-$ 1042
$3989-
$4237
$429-$489
$2020-
$2210
$639-$700
$615-$711
$340-$425
$155-$155
$462-$541
$2240-
$2411
$1912-
$1988
$850-$962
Truck
$1270-
$1304
$1615-
$1840
$863-$ 1015
$937-$ 1078
$692-$778
$1418-
$1760
$2998-
$3043
$552-$649
$2035-
$2236
$922-$999
$970-$1139
$277-$367
$0-$0
$930-$1151
$2073-
$2047
$2094-
$2479
$1094-
$1253
Combined
$1246-
$1294
$1456-
$1682
$758-$913
$905-$1021
$558-$649
$1015-
$1132
$3186-
$3269
$466-$537
$2026-
$2221
$747-$815
$765-$892
$291-$380
$155-$155
$684-$831
$2174-
$2267
$2009-
$2250
$971-$1106
Delta in MY2021
Car
$402-$423
$64-$126
$105-$188
$185-$215
$68-$75
$144-$ 155
$941-$1264
$80-$126
$351-$453
$60-$98
$90-$92
$33-$66
$0-$0
$45-$50
$432-$433
$445-$624
$154-$162
Truck
$21-$123
$382-$394
$128-$159
$137-$163
$150-$169
$491-$526
$578-$677
$0-$115
$332-$501
$177-$273
$191-$196
$84-$85
$0-$0
$189-$238
$437-$468
$194-$455
$225-$234
Combined
$299-$342
$296-$306
$137-$153
$173-$174
$108-$ 120
$187-$202
$708-$727
$88-$90
$403-$413
$128-$142
$134-$135
$73-$81
$0-$0
$113-$ 140
$434-$447
$395-$450
$189-$197
Note:  In the Reference and Control cases, lower values use ICMs while higher values use RPEs; in the Delta
columns, the minimum delta forms the lower value while the maximum delta forms the higher value.
                                                 12-80

-------
                                      EPA's Analysis of the MY2022-2025 GHG Standards
   Table 12.97 MY2025 Absolute and Incremental Costs per Vehicle in the Central Analysis Using AEO
        Reference Case Fuel Prices and Fleet Projections and Using both ICMs and RPEs (2013$)

Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-
Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
Fleet
Reference Case in MY2025
Car
$644-$741
$911-$1085
$434-$538
$575-$674
$298-$327
$716-$804
$1727-
$2123
$302-$341
$1099-
$1442
$505-$601
$468-$549
$225-$315
$140-$ 140
$336-$449
$1418-
$1633
$1104-
$1327
$586-$695
Truck
$872-$965
$1051-
$1308
$630-$800
$728-$900
$474-$578
$845-$1056
$2044-
$2229
$429-$501
$1479-
$1534
$614-$666
$710-$862
$160-$226
$0-$0
$676-$759
$1359-
$1405
$1595-
$1821
$765-$918
Combined
$698-$794
$1009-
$1242
$548-$690
$652-$787
$380-$444
$732-$834
$1978-
$2207
$341-$390
$1244-
$1477
$544-$624
$564-$673
$174-$246
$140-$ 140
$490-$589
$1396-
$1547
$1360-
$1584
$671-$801
Control Case in MY2025
Car
$1724-
$1921
$1789-
$2149
$969-$1144
$1169-
$1384
$842-$896
$1447-
$1705
$5090-
$5489
$772-$880
$2482-
$2843
$1178-
$1325
$1148-
$1184
$686-$766
$140-$ 140
$884-$1004
$2751-
$3178
$2351-
$2902
$1293-
$1483
Truck
$1942-
$2153
$2451-
$2809
$1777-
$2073
$2248-
$2534
$967-$1349
$2128-
$2335
$3436-
$3821
$1081-
$1250
$2732-
$3061
$1333-
$1532
$1526-
$2080
$691-$873
$0-$0
$1547-
$1900
$2560-
$2721
$3170-
$3078
$1864-
$2184
Combined
$1776-
$1977
$2254-
$2613
$1438-
$1684
$1707-
$1957
$901-$1107
$1529-
$1780
$3782-
$4170
$866-$993
$2577-
$2926
$1234-
$1399
$1298-
$1539
$690-$849
$140-$ 140
$1184-
$1409
$2679-
$3005
$2777-
$2994
$1565-
$1818
Delta in MY2025
Car
$1080-
$1181
$879-$ 1063
$535-$606
$593-$710
$544-$569
$731-$901
$3363-
$3366
$469-$539
$1383-
$1401
$673-$724
$635-$680
$451-$461
$0-$0
$548-$555
$1333-
$1544
$1247-
$1575
$707-$789
Truck
$1070-
$1188
$1400-
$1501
$1147-
$1273
$1520-
$1633
$493-$771
$1279-
$1284
$1391-
$1592
$652-$748
$1253-
$1528
$719-$866
$816-$1218
$531-$647
$0-$0
$871-$1140
$1202-
$1316
$1257-
$1575
$1099-
$1267
Combined
$1078-
$1183
$1245-
$1371
$890-$993
$1055-
$1170
$520-$663
$797-$946
$1804-
$1963
$525-$603
$1334-
$1449
$689-$775
$734-$866
$515-$603
$0-$0
$694-$820
$1284-
$1458
$1410-
$1417
$894-$1017
Note: In the Reference and Control cases, lower values use ICMs while higher values use RPEs; in the
columns, the minimum delta forms the lower value while the maximum delta forms the higher value.
Delta
   The vehicle costs used as inputs to the OMEGA Inventory, Cost and Benefit Tool (ICBT) are
shown in the tables below.
                                               12-81

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
   Table 12.98 Reference Case Absolute Cost/Vehicle Used as Inputs to the OMEGA Inventory, Cost and
                                    Benefit Tool (2013$)

MY
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
AEO Reference Fuel Price
Case, ICMs
Car
$696
$669
$636
$601
$586
$586
$586
$585
$585
$585
Truck
$868
$843
$812
$778
$765
$765
$766
$766
$765
$765
AEO High Fuel Price
Case, ICMs
Car
$721
$696
$664
$630
$617
$617
$617
$617
$616
$616
Truck
$912
$885
$852
$816
$801
$801
$802
$802
$801
$800
AEO Low Fuel Price
Case, ICMs
Car
$675
$649
$617
$582
$569
$569
$568
$568
$568
$568
Truck
$866
$841
$809
$775
$762
$762
$763
$763
$762
$762
AEO Reference Fuel Price
Case, RPEs
Car
$799
$774
$742
$708
$695
$695
$694
$694
$694
$694
Truck
$1,019
$994
$964
$930
$918
$917
$919
$919
$918
$917
 Table 12.99 Control Case Absolute Cost/Vehicle Used as Inputs to the OMEGA Inventory, Cost and Benefit
                                       Tool (2013$)

MY
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
AEO Reference Fuel Price
Case, ICMs
Car
$850
$961
$1,067
$1,170
$1,293
$1,293
$1,292
$1,292
$1,292
$1,292
Truck
$1,094
$1,287
$1,474
$1,658
$1,864
$1,863
$1,866
$1,866
$1,864
$1,863
AEO High Fuel Price
Case, ICMs
Car
$861
$976
$1,085
$1,191
$1,319
$1,318
$1,318
$1,317
$1,317
$1,317
Truck
$1,140
$1,342
$1,538
$1,731
$1,945
$1,945
$1,948
$1,947
$1,944
$1,943
AEO Low Fuel Price
Case, ICMs
Car
$835
$946
$1,051
$1,153
$1,276
$1,276
$1,275
$1,275
$1,274
$1,274
Truck
$1,064
$1,259
$1,447
$1,633
$1,839
$1,839
$1,842
$1,842
$1,840
$1,839
AEO Reference Fuel Price
Case, RPEs
Car
$961
$1,092
$1,218
$1,340
$1,483
$1,483
$1,483
$1,482
$1,482
$1,482
Truck
$1,253
$1,486
$1,714
$1,939
$2,184
$2,184
$2,187
$2,187
$2,184
$2,184
12.4.2 Cost per Percentage Improvement in COi

   Each manufacturer's starting and ending CCh levels are shown in the tables below by car,
truck and combined fleet. Also included are EPA's estimated costs per vehicle. Using these data,
we can calculate the costs per percentage reduction in CCh emissions from the baseline case (i.e.,
the MY2025 fleet meeting the MY2014 standards) to the central analysis control case (i.e., the
MY2025 fleet meeting the MY2025 standards) and using ICMs only here.

   The results shown in these tables represent the CCh impacts and cost impacts (using ICMs) of
taking the MY2014 baseline fleet, projecting it to a MY2025 fleet meeting the MY2014
standards, and bringing that fleet into compliance with the MY2025 standards.  Note that the
costs presented here fall slightly short of the costs presented in earlier tables.  For example, Table
12.102 shows a delta cost of $1,512 while Table 12.97 shows a cost of $1,565 (using ICMs).
This difference between $1,565 and $1,512 represents the costs to bring the baseline fleet in
MY2025 into compliance with the MY2014 standards.  That cost is reflected in Table 12.97 but
                                              12-82

-------
                                        EPA's Analysis of the MY2022-2025 GHG Standards
is not reflected below since the tables below use the 2014 standards as their reference case (i.e.,
costs below are relative to meeting the 2014 standards).

  Table 12.100 CCh and Cost Changes in MY2025 using the 2014 Standards as the Reference Case and the
            2025 Standards as the Control Case for Cars (CCh in g/mi, dollar values in 2013$)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
All
Base CO2
236
271
245
257
206
242
271
205
263
220
219
234
0
205
248
265
232
Final CO2
146
160
166
169
142
149
102
149
142
148
146
174
0
143
134
154
150
Delta CO2
-90
-111
-80
-89
-63
-92
-168
-56
-121
-72
-73
-61
0
-62
-114
-110
-82
% Delta CO2
-38%
-41%
-32%
-35%
-31%
-38%
-62%
-27%
-46%
-33%
-33%
-26%
0
-30%
-46%
-42%
-35%
Cost delta
$1,664
$1,729
$908
$1,109
$782
$1,387
$5,030
$712
$2,422
$1,118
$1,088
$626
$0
$824
$2,691
$2,291
$1,233
$/%CO2
-$44
-$42
-$28
-$32
-$25
-$36
-$81
-$26
-$53
-$34
-$33
-$24
$0
-$27
-$58
-$55
-$35
          Note:  Values include use of A/C and off-cycle credits described in Table 12.6 and their costs.


  Table 12.101 CCh and Cost Changes in MY2025 using the 2014 Standards as the Reference Case and the
           2025 Standards as the Control Case for Trucks (CCh in g/mi, dollar values in 2013$)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
All
Base CO2
306
350
364
359
295
313
344
274
348
244
322
239

336
321
336
333
Final CO2
197
199
219
209
192
171
205
175
203
151
197
164

203
204
183
201
Delta CO2
-109
-150
-145
-150
-103
-142
-140
-99
-145
-93
-125
-75

-133
-117
-152
-133
% Delta CO2
-35%
-43%
-40%
-42%
-35%
-45%
-41%
-36%
-42%
-38%
-39%
-32%

-39%
-36%
-45%
-40%
Cost delta
$1,896
$2,404
$1,730
$2,202
$921
$2,082
$3,389
$1,035
$2,686
$1,286
$1,480
$644

$1,501
$2,514
$3,124
$1,817
$/%C02
-$53
-$56
-$43
-$53
-$26
-$46
-$84
-$29
-$65
-$34
-$38
-$20

-$38
-$69
-$69
-$46
          Note:  Values include use of A/C and off-cycle credits described in Table 12.6 and their costs.
                                                 12-83

-------
                                     EPA's Analysis of the MY2022-2025 GHG Standards
Table 12.102 CCh and Cost Changes in MY2025 using the 2014 Standards as the Reference Case and the
    2025 Standards as the Control Case for the Combined Fleet (CCh in g/mi, dollar values in 2013$)
Manufacturer
BMW
FCA
Ford
GM
Honda
Hyundai/Kia
JLR
Mazda
Mercedes-Benz
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volkswagen
Volvo
All
Base CO2
252
326
314
308
247
250
329
226
295
228
260
238
0
264
276
302
281
Final CO2
157
189
197
189
165
150
185
156
165
148
165
168
0
170
160
170
174
Delta CO2
-96
-138
-117
-119
-82
-100
-144
-70
-131
-80
-94
-71
0
-94
-116
-132
-106
% Delta CO2
-38%
-42%
-37%
-39%
-33%
-40%
-44%
-31%
-44%
-35%
-36%
-30%
0
-36%
-42%
-44%
-38%
Cost delta
$1,723
$2,200
$1,385
$1,653
$847
$1,475
$3,728
$813
$2,524
$1,180
$1,244
$636
$0
$1,130
$2,626
$2,723
$1,512
$/%C02
-$46
-$52
-$37
-$43
-$26
-$37
-$85
-$26
-$57
-$34
-$34
-$21
$0
-$32
-$63
-$62
-$40
        Note: Values include use of A/C and off-cycle credits described in Table 12.6 and their costs.
                                               12-84

-------
                                         EPA's Analysis of the MY2022-2025 GHG Standards
References
1 A DVD has been placed in the docket with the name "OMEGA_TAR2016."
2 Previous OMEGA documentation for versions used inMYs 2012-2016 Final Rule (EPA-420-B-09-035), Interim
Joint Draft TAR (EPA-420-B-10-042). Docket Nos. EPA-HQ-OAR-2010-0799-1108 and EPA-HQ-OAR-2010-
0799-1109.
3 http://www.epa.gov/oms/climate/models.htm.
4EPA-420-R-09-016, September 2009. (Docket No. EPA-HQ-OAR-2010-0799-1135).
5 EPA. Proposed Rulemaking for Greenhouse Gas Emissions and Fuel Efficiency Standards for Medium- and
Heavy-Duty Engines and Vehicles-Phase 2. EPA-420-D-15-900, June 2015.
6 EPA. National Emission Inventory Data. httBS^/wwTjy;^^
                 2011.
7 EPA. Regulatory Impact Analysis:  Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards, Chapter 4.2, 4.6, EPA-420-R-12-016, August
2012.
8 EPA.  Regulatory Impact Analysis:  Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards, Chapter 4.6.3., EPA-420-R-12-016, August
2012, Chapter 4.6.3.
9 77 FR 62821, October 15, 2012.
10 See EPA's final RIA in support of the 2012 FRM (EPA-420-R-12-016) at page 4-131.
11 Argonne National Laboratory's The Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation
(GREET) Model, Version l.Sc.O, available at http://www.transportation.anl.gov/modeling_simulation/GREET/).
EPA Docket EPA-HQ-OAR-2009-0472. (Docket No. EPA-HQ-OAR-2010-0799-1105).
12 EPA.  eGrid 2010, http://www.epa.gov/cleanenergy/energy-resources/egrid/index.html (Docket No. (EPA-HQ-
OAR-2010-0799-0832).
1340CFR86.1866-12(a).
                                                  12-85

-------
                                                         Analysis of Augural CAFE Standards
Table of Contents

Chapter 13: Analysis of Augural CAFE Standards	13-1
   13.1    Significant Assumptions and Inputs to theNHTSA Analysis	13-2
     13.1.1   MY2015 Analysis Fleet	13-2
     13.1.2   Assumptions about Product Cadence	13-5
     13.1.3   Assumptions about Consumer Behavior	13-8
     13.1.4   Updated Mileage Accumulation Schedules for the Draft TAR	13-11
        13.1.4.1  Updated Schedules	13-11
        13.1.4.2  Data Description	13-16
        13.1.4.3  Estimation	13-19
        13.1.4.4  Comparison to previous schedules	13-20
        13.1.4.5  Future direction	13-21
     13.1.5   Other Assumptions of Note	13-21
   13.2    CAFE Model (aka "Volpe Model") Overview and Updates Since the 2012 Final Rule
          13-23
     13.2.1   Updates to 2012 Final Rule Version of the CAFE Model	13-24
        13.2.1.1  Integrating Vehicle Simulation Results into the CAFE Model	13-28
     13.2.2   Overview and Technology Application	13-37
     13.2.3   Simulating Manufacturer Compliance with Standards	13-48
     13.2.4   Simulating the Economic and Environmental Effects of CAFE Standards	13-55
   13.3    Simulation Results for Augural MY2022 - 2025 Standards	13-56
     13.3.1   Industry Impacts	13-57
     13.3.2   Consumer Impacts	13-93
     13.3.3   Social and Environmental Impacts	13-99
     13.3.4   Overall Benefits and Costs	13-102

Table of Figures

Figure 13.1 CAFE and Standard from 2010 Fleet Simulations vs. 2015 Observed Fleet	13-3
Figure 13.2 Sales from 2010 Fleet Simulations vs. 2015 Observed Fleet	13-5
Figure 13.3 Share of Manufacturer Sales Redesigned In Each Model Year  2016 -2030	13-7
Figure 13.4 Industry Average CAFE and Standard 1990 - 2014	13-10
Figure 13.5 A Comparison of the Current and Previous Passenger Car Schedules	13-12
Figure 13.6 Total VMT and Share of Population by Ownership Type for Passenger Cars	13-13
Figure 13.7 A Comparison of the Current and Previous SUV/Van Schedules	13-13
Figure 13.8 A Comparison of the Current and Previous Pickup Truck Schedules	13-14
Figure 13.9 A Comparison of the Current and Previous MD Pickup/Van Schedules	13-14
Figure 13.10 Total VMT and Share of Population by Ownership Type for MD Pickups/Vans	13-15
Figure 13.11 Distribution of the Ratio of Sample Size to Population Size (by Make/Model/MY)	13-17
Figure 13.12 Percentage of Total Vehicle Population with No Odometer Readings across Model Years	13-17
Figure 13.13 Percentage of Vehicle Models  with Fewer than 5% of the Population in Odometer Readings Data (by
           Class)	13-18
Figure 13.14 Difference in Share of Each Vehicle Model in Population vs. Odometer Sample (by Class)	13-19
Figure 13.15 Retrospective Analysis of EIA Fuel Price Projections	13-22
Figure 13.16 Comparison of Fuel Price Estimates in Draft TAR and 2012 Final Rule Analysis	13-23
Figure 13.17 Fuel Economy of Simulated Vehicles before (Red) and after (Blue) Application of Level 1
           Turbocharging	13-30
Figure 13.18 Fuel Economy Improvement to Vehicles That Acquire Level 1 Turbocharging In Simulation	13-31
Figure 13.19 Midsize Vehicles in the Database Eligible to Receive TURBO1	13-33
Figure 13.20 Technology Tree Used to Map Autonomie Simulations to Draft TAR Technology Set	13-34

-------
                                                              Analysis of Augural CAFE Standards
Figure 13.21 Engine-Level Paths	13-43
Figure 13.22 Transmission-Level Paths	13-44
Figure 13.23 Platform-Level Paths	13-45
Figure 13.24 Vehicle-Level Paths	13-45
Figure 13.25 Technology Pathways Diagram	13-47
Figure 13.26 Compliance Simulation Diagram	13-50
Figure 13.27 Selection of "Next Best" Technology within CAFE Compliance Simulation	13-52
Figure 13.28 FCA Compliance Example	13-55
Figure 13.29 CAFE and Standard from 2010 Fleet Simulations vs. 2015 Observed Fleet (miles per gallon)	13-60
Figure 13.30 Passenger Car Engine Technology Penetration Rates By Manufacturer (sales weighted share of fleet)
             	13-62
Figure 13.31 Passenger Car Transmission Penetration Rates By Manufacturer (sales weighted share of fleet).. 13-63
Figure 13.32 Passenger Car Electrification Technology Penetration Rates By Manufacturer (sales weighted share of
            fleet)	13-64
Figure 13.33 Passenger Car Load Reduction Technology Penetration Rates By Manufacturer (sales weighted share
            of fleet)	13-65
Figure 13.3 4 Light Truck Engine Technology Penetration Rates By Manufacturer (sales weighted share of fleet). 13-
            68
Figure 13.35 Light Truck Transmission Technology Penetration Rates By Manufacturer (sales weighted share of
            fleet)	13-69
Figure 13.36 Light Truck Electrification Technology Penetration Rates By Manufacturer (sales weighted share of
            fleet)	13-70
Figure 13.37 Light Truck Load Reduction Technology Penetration Rates By Manufacturer (sales weighted share of
            fleet)	13-71
Figure 13.38 Industry-Wide Combined Average Fuel Economy Levels and Average Costs (2015 $)	13-74
Figure 13.39 Sensitivity of Incremental Regulatory Costs (MY2016 - MY2030) to Alternative Assumptions.. 13-91
Figure 13.40 Sensitivity of Total Regulatory Costs (MY2016 - MY2030) to Alternative Assumptions	13-92
Figure 13.41 Payback Periods for the Baseline Standards, Augural Standards, and Total over the Period	13-97
Figure 13.42 Comparison of Environmental And Physical Effects, Draft TAR and 2012 Final Rule	13-100
Figure 13.43 Societal Safety Effects for the Augural Standards (relative to MY2021 standards)	13-101
Figure 13.44 Influence of Alternative Assumptions on Net Benefits Attributable to Augural Standards	13-104


Table of Tables

Table 13.1  Summary Comparison of Lifetime VMT for Current and Previous Schedules	13-15
Table 13.2  CAFE Credits Estimated to be Available from 2010-2014 (1 vehicle x 0.1 mpg =  1 credit)	13-28
Table 13.3  CAFE Model Technologies (1)	13-38
Table 13.4  CAFE Model Technologies (2)	13-39
Table 13.5  Vehicle Technology Classes	13-40
Table 13.6  Engine Technology Classes	13-41
Table 13.7  Technology Pathways	13-42
Table 13.8  Expected Manufacturer Standards and Expected CAFE levels with Augural Standards through MY2030
             	13-59
Table 13.9  Average Per Vehicle Cost for Primary Analysis Using RPE  to Mark Up Direct Costs	13-73
Table 13.10 Draft TAR Average Per Vehicle Cost and Production Volume in MY 2025 for Primary Analysis Using
            RPE to Mark Up Direct Costs	13-75
Table 13.11  Estimated MY2028 CAFE Levels and Average Fuel Consumption Improvement	13 -77
Table 13.12 Estimated Technology Cost per Percent Fuel Consumption Improvement in MY2028	13-78
Table 13.13  Passenger Cars: Total Cost and Average  per Vehicle Costs for the No-Action Alternative for the
            Primary Analysis Using RPE to  Mark Up Direct Costs	13-80
Table 13.14 Light Trucks: Total Cost and Average per Vehicle Costs for the No-Action Alternative for the Primary
            Analysis Using RPE to Mark Up Direct  Costs	13-81
Table 13.15  AllVehicles:  Total Cost and Average per Vehicle Costs for the No-Action Alternative for the Primary
            Analysis Using RPE to Mark Up Direct  Costs	13-82

-------
                                                              Analysis of Augural CAFE Standards
Table 13.16 Passenger Cars Additional Total Cost and Average per Vehicle Costs for the MYs 2022-2025 Augural
             Standards for the Primary Analysis Using RPE to Mark Up Direct Costs	13-84
Table 13.17 Light Trucks: Additional Total Cost and Average per Vehicle Costs for the MYs 2022-2025 Augural
             Standards for the Primary Analysis Using RPE to Mark Up Direct Costs	13-85
Table 13.18 AllVehicles: Additional Total Costs and Average per Vehicle Costs for the MYs 2022-2025 Augural
             Standards for the Primary Analysis Using RPE to Mark Up Direct Costs	13-86
Table 13.19 Passenger Costs: Total Cost and Total Average per Vehicle Costs for the MYs 2022-2025 Augural
             Standards for the Primary Analysis Using RPE to Mark Up Direct Costs	13-87
Table 13.20 Light Trucks: Total Cost and Total Average per Vehicle Costs for the MYs 2022-2025 Augural
             Standards for the Primary Analysis Using RPE to Mark Up Direct Costs	13-88
Table 13.21 All Vehicles: Total Cost and Total Average per Vehicle Costs for the MYs 2022-2025 Augural
             Standards for the Primary Analysis Using RPE to Mark Up Direct Costs	13-89
Table 13.22 Definition of Sensitivity Cases Considered For Draft TAR	13-90
Table 13.23 Comparison of Cost Estimates Using Retail Price Equivalent and Indirect Cost Multiplier Mark Up. 13-
             93
Table 13.24 Average Regulatory Cost per Vehicle by Model Year, 2015 - 2028	13-95
Table 13.25 Estimated Present Value of Costs, Benefits and Net Benefits ($b) Over the Lifetimes of MYs 2016-
             2028 Vehicles Using 3 Percent Discount Rate (2013$)	13-103

-------
                                                    Analysis of Augural CAFE Standards
Chapter 13: Analysis of Augural CAFE Standards

   The purpose of this chapter is to describe the inputs, assumptions, and tools that form the
foundation of NHTSA's analysis in the Draft TAR. The results of the analysis that uses all of
these assumptions and tools are summarized in Section 13.3. While many of the inputs to this
analysis have been summarized elsewhere in the Draft TAR, this chapter provides more detailed
descriptions of assumptions that either differ in important ways from the last Final Rule
(covering MYs 2017-2021) or have the ability to significantly impact the evaluation of program
impacts. The chapter takes a close look at a range of important factors that influence the impact
of CAFE standards, such as variations in fuel price and the ways that consumer demand
influence technology integration on the supply side.

   NHTSA's analysis illustrates the impact of these and other technical assumptions by
modeling the Augural Standards for 2022-2025 as a point of comparison relative to NHTSA's
final CAFE standards through 2021. As noted in the executive summary, the Draft TAR does not
present alternatives to the Augural Standards because, as the first stage of the Midterm
Evaluation process, the TAR is principally an exploration of technical issues - including
assumptions about the effectiveness and cost of specific technologies, as well as other inputs,
methodologies and approaches for accounting for these issues. The agencies seek comment from
stakeholders to further inform the analyses, which will inform subsequent development of
stringency alternatives.

   To conduct today's analysis, NHTSA has made use of NHTSA's Corporate Average Fuel
Economy (CAFE) Modeling System (sometimes referred to as "the CAFE model" or "the Volpe
model"), which DOT's Volpe National Transportation Systems Center (Volpe Center)
continuously develops, maintains,  and applies to support NHTSA CAFE analyses and
rulemakings. The Volpe Center has supported the CAFE program since USDOT first established
fuel economy rules beginning with MY 1978, following the initial  authorization of the CAFE
program in the Energy Policy and Conservation Act of 1975. NHTSA developed the first version
of the model in 2002 to support the 2003 issuance of CAFE standards for MYs 2005-2007 light
trucks. NHTSA has since significantly expanded and refined the model, and has applied the
model to support every ensuing CAFE rulemaking, including:

   2006: MYs 2008-2011 light trucks

   2008: MYs 2011-2015 passenger cars and light trucks

   2009: MY 2011 passenger cars and light trucks

   2010: MYs 2012-2016 passenger cars and light trucks

   2012: MYs 2017-2021 passenger cars and light trucks

   2015: MYs 2021-2027 heavy-duty pickups and vans (NPRM)

   Past analyses conducted using the CAFE model have been subjected to extensive and detailed
review and comment, much of which has informed the model's expansion and refinement.
NHTSA's use of the model was considered and supported in 2007  litigation (CBD v. NHTSA),
and the model has been subjected to formal peer review and review by the General Accounting
Office (GAO) and National Research Council (NRC).  NHTSA makes public the model, source
                                            13-1

-------
                                                      Analysis of Augural CAFE Standards
code, and—except insofar as doing so will compromise confidential business information (CBI)
manufacturers have provided to NHTSA—all model inputs and outputs underlying published
rulemaking analyses.1

   Although the CAFE model can also be used for more aggregated analysis (e.g., involving
"representative vehicles," single-year snapshots, etc.), NHTSA designed the model with a view
toward (a) detailed simulation of manufacturers' potential actions given a defined set of
standards, followed by (b) calculation of resultant impacts and economic costs and benefits.  The
model is intended to describe actions manufacturers could take in light of defined standards,
estimated production constraints, and other input assumptions and estimates, not to predict
actions manufacturers will take.  While a more detailed description of the model appears in the
model documentation, Section 13.2 of this chapter provides an overview of important model
logic and new developments since the last public release accompanying the 2012 Final Rule.

13.1  Significant Assumptions and Inputs to the NHTSA Analysis

13.1.1  MY2015 Analysis Fleet

   For the CAFE model, the "analysis fleet" is the foundation of the analysis. The characteristics
of the analysis fleet have important implications both for the simulation of what standard
manufacturers are required to meet, and for what technologies are applicable within the
compliance simulation. The 2017-2021 Final Rule used all MY2010 vehicles available for sale in
the U.S. market as its analysis fleet, holding vehicle characteristics constant at MY2010  levels
but using other information  sources to estimate future production volumes. As discussed above
in Chapter 4, for the Draft TAR we have opted to use the MY2015 fleet, being the most  current
available at the time of the analysis. The sales volumes, which determine achieved CAFE levels,
are based on projections submitted by manufacturers and may differ from final end-of-year
compliance submissions.

   The standards are calculated from the sales-weighted, harmonic average of individual vehicle
targets and these targets are determined from the footprint and regulatory class  of a vehicle. For
this reason, changes to an individual vehicle which alter either of these characteristics may result
in different standards for the manufacturer fleet of that vehicle. The CAFE model currently does
not attempt to estimate changes in vehicle footprint or changes in characteristics which would
shift a given light-duty vehicle's fuel economy targets or even regulatory class, though the model
does provide means to estimate the impact of mass reduction  on fuel consumption targets for
heavy-duty pickups and vans regulated separately from light-duty vehicles, and future analyses
may consider allowing the footprint of individual vehicle models to change and thereby  alter a
given light-duty vehicle model's fuel economy target under the standards (although doing so
would likely also entail a fuel  economy change to be balanced against the change in the target).

   A manufacturer's individual average requirement under the standard may also change based
on its decision to introduce or discontinue vehicles from a fleet, or through shifts in production
vehicles among existing vehicles, especially insomuch as such shifts affect the relative shares
represented by passenger cars and light trucks, respectively.  Although the CAFE model can
accommodate inputs that account for exogenously estimated shifts in product offerings, there is
no way within the CAFE model to endogenously estimate the entrance or exiting of a model
1 Analyses can be found at http://www.nhtsa.gov/fuel-economy
                                              13-2

-------
                                                      Analysis of Augural CAFE Standards
from a manufacturer's fleet, so, from the perspective of this analysis, the set of vehicles that
exists in the analysis fleet (MY2015, in this case) is the set of vehicles to which technology may
be added to achieve compliance.

   The calculation of manufacturers' estimated actual requirements in 2015, relative to earlier
predictions for that year, demonstrate how evolving production trends can impact the standards
on a year-by-year basis. For example, Figure 13.1 compares the 2015 requirements simulated
using the 2017-2021 Final Rule Analysis, based on the 2010 fleet (dotted-lines) and the
calculated 2015 requirements (solid lines). As noted above, the patterns reflected in the chart
demonstrate the impact of, among other things, changes in the ratio between passenger cars and
light trucks - responsive to the latter comprising a greater share of production than anticipated in
at the time of the Final Rule, which assumed passenger cars would represent 65 percent of the
new vehicle market (and growing). The actual value for MY2015 is closer to 58 percent, which
is reflected in the  combined requirement for each manufacturer. When passenger cars and light
trucks are separated by class, the gaps between previously-forecast and currently-estimated
actual requirements are narrower.
2 ^
t <
I" 30^
         — 20
         --- 20
0 Fleet CAFE
5 Fleet CAFE
5 Simulated Fleet CAFE
5 Fleet Standard
5 Simulated Fleet Standard
          Figure 13.1 CAFE and Standard from 2010 Fleet Simulations vs. 2015 Observed Fleet

   Figure 13.1 also shows how CAFE levels in the MY2010 fleet compared to the MY2015 fleet.
Even with shifts in the relative market shares of passenger cars and light trucks, the CAFE level
for most manufacturers' combined fleets is higher in MY2015 than it was in 2010. Exceptions
tend to reflect especially pronounced car-to-light-truck shifts. For example, Ford, which is
currently estimated to have produced a 48 percent PC fleet in MY2015 rather than the 56 percent
forecast in the 2012 analysis, had a CAFE level of 29.48 in their 2010 fleet, and 29.33 in their
2015 fleet. For fleets where a manufacturer was well below their 2015 requirement in 2010, there
                                              13-3

-------
                                                     Analysis of Augural CAFE Standards
is the most movement in the CAFE level (BMW, Daimler, Mitsubishi, Nissan, and Subaru, while
Fiat seems to be an exception to this trend). For manufacturers that were close to their 2015
requirement in 2010 (Ford, Hyundai Kia, and Toyota, while Honda seems to be an exception,
here), there is less movement in their fleets. Some manufacturers have made choices to rely on
banked credits or pay fines. Of manufacturers that did not meet their 2015 standard, BMW,
Ford, GM, Hyundai-Kia, and Volkswagen  all had credits built up, while Daimler and Fiat
already had a negative credit balance, but have historically been fine-payers. In total, the CAFE
level has increased from 30.2 to 31.8, close to the industry standard of 32.1.

   Another trend reflected in Figure 13.1 is that while the simulated and actual 2015 CAFE
requirements differ somewhat, the delta between achieved levels and requirements is narrow, and
comparable, across both. The Final Rule simulation of MY2015 showed the industry-level CAFE
at 33.2 and the requirement at 32.9; the simulation showed the industry exceeding the 2015
average requirement by the same amount that the actual  fleet was short of the requirement (.3
MPG). For most manufacturer's combined fleet, the simulated gap between the requirement and
CAFE level  achieved was fairly close to the observed gap. For BMW, Daimler, Honda, Mazda,
Nissan, and  Subaru, the Final Rule simulation underestimated their CAFE levels in 2015 relative
to their observed CAFE: all of these manufacturers performed better in reality than their
simulated fleets. And this is true even accounting for differences between the expected 2015
requirement and the actual  requirement. For example, Nissan's requirement is very close to the
simulated requirement, but they performed much better than the simulated fleet; possibly due to
the entrance of the Leaf in 2014 to their passenger car fleet, which was not present in the
MY2010 fleet. Daimler, BMW and Honda all had slightly lower requirements, but the majority
of their improvement against their requirements can be attributed to their CAFE levels being
higher than predicted with the 2010 fleet. The Subaru simulated CAFE level is the same as their
achieved level, but their requirement is significantly less stringent than the simulated
requirement (this is likely attributable to the redesigns of two of their popular light truck
models—which made their footprints larger and their targets lower—the Forester and Outback in
2014 and 2015, respectively).

   Figure 13.2 shows the sales for all manufacturers by regulatory class in 2010, 2015, and
simulated 2015 sales from the 2010 fleet. The simulated 2015 fleet was fairly indicative of how
many vehicles were sold in each manufacturer's fleet. Sales for the Big 3 (Ford, Fiat, and GM)
were not predicted as well, possibly due to the time of the U.S. automobile industry volatility
(2008-2009) or the assumption that passenger cars would gain share for manufacturers that sell
large volumes of pickup trucks. Fiat sold slightly more passenger cars and significantly more
light trucks than shown in the 2015 simulated fleet, Ford sold  slightly fewer light trucks and
significantly fewer passenger cars, and General Motors sold more light trucks and fewer
passenger cars. Only Nissan sold more passenger cars than predicted, though they sold more
trucks as well.  The total sales were not included because they  would significantly skew the scale
of the figure, but the 2015 industry fleet was made up of 1 million more light trucks and 800
thousand fewer passenger cars than the simulated industry fleet.
                                             13-4

-------
                                                      Analysis of Augural CAFE Standards
       3M
           Source
           | 2010 fleet
           • 2015 Fleet
           • 201S Fleet Simulatic
   OJ n
   ?*
                    0

                    Q
                                f)
                I
- £
jl
                                                                  1
                Figure 13.2  Sales from 2010 Fleet Simulations vs. 2015 Observed Fleet
   The preceding discussion illustrates that compliance simulations with the CAFE model can do
a reasonably good job of estimating future CAFE levels and requirements, but that dynamic
economic trends, including consumer choice, can affect production trends over time and affect
actual requirements. Predictive modeling will consistently reflect the best available forecasts of
macro-economic trends like energy prices and overall growth, which in turn tend to inform
consumer choices and trends in driving habits.

13.1.2  Assumptions about Product Cadence

   Past comments on the CAFE model have stressed the importance of product cadence—i.e.,
the development and periodic redesign and freshening of vehicles—in terms of involving
technical, financial, and other practical constraints on applying new technologies, and NHTSA
has steadily made changes to both the CAFE model and its inputs with a view toward accounting
for these considerations. For example, early versions of the model added explicit "carrying
forward" of applied technologies between model years, subsequent versions applied assumptions
that most technologies will be applied when vehicles are freshened or redesigned, and more
recent versions applied assumptions that manufacturers would sometimes apply technology
earlier than "necessary" in order to facilitate compliance with standards in ensuing model years.
Thus, for example, if a manufacturer is expected to redesign many of its products in model years
2018 and 2023,  and the standard's  stringency increases significantly in model year 2021, the
CAFE model will estimate the potential that the manufacturer will add more technology than
necessary for compliance in MY 2018, in order to carry those product  changes forward through
the next redesign and contribute to compliance with the MY 2021 standard.  This explicit
                                             13-5

-------
                                                      Analysis of Augural CAFE Standards
simulation of multiyear planning plays an important role in determining year-by-year analytical
results.

   As in previous iterations of CAFE rulemaking analysis, the NHTSA's simulation of
compliance actions that manufacturers might take is constrained by the pace at which new
technologies can be applied in the new vehicle market. Operating at the Make/Model level (e.g.,
Toyota Camry) allows NHTSA to explicitly account for the fact that individual vehicle models
undergo significant redesigns relatively infrequently. Many popular models are only redesigned
every six years or so, with some larger/legacy platforms (the old Ford Econoline Vans, for
example) stretching more than a decade between significant redesigns. Engines, which are often
shared among many different models and platforms for a single manufacturer, can last even
longer - eight to ten years in most cases.

   Understanding manufacturers' redesign schedules, albeit subject to change, is valuable for
planning purposes,  including anticipating redesign schedules, as well as predicting  when and
how manufacturers may make use of crediting options.  However, while manufacturers'
characterizations of product cadence are important to any evaluation of the impacts of CAFE
standards, they are not known with certainty - even by the manufacturers themselves over time
horizons as long as those covered by this analysis. For example, the Honda Civic, which was
typically redesigned on a 4-6 year cycle, underwent a significant,  and unprecedented, change for
the 2013 model year to address feedback from the MY2012 redesign. Even in that case, the
engines and transmissions offered on the Civic did not change between MY2012 and MY2013,
suggesting that either Honda considered the feedback was entirely due to other characteristics of
the vehicle or that changing the powertrains so quickly was too costly.

   Indeed, when NHTSA staff meets with manufacturers to discuss manufacturers' plans vis-a-
vis CAFE requirements, manufacturers' staff typically present specific and detailed year-by-year
information that explicitly accounts for anticipated redesigns.  Such year-by-year analysis is also
essential to manufacturers' plans to make use  of statutory provisions allowing CAFE credits to
be carried forward to future model years, carried back from future model years, transferred
between regulated fleets, and traded with other manufacturers. Manufacturers are never certain
about future plans, but they spend considerable effort developing  them. For every model that
appears in the MY2015 analysis fleet, NHTSA has  estimated the model years in which future
redesigns (and less  significant "freshening." which  offer manufacturers the opportunity to make
less significant changes to models) will occur. These appear in the market data file for each
model. Figure  13.3  gives a summary of the  share of each manufacturer's sales expected to be
redesigned in a given model year. It is worth noting that every manufacturer has at least one
model year in which no significant portion of its models (by sales) is redesigned. Mid-cycle
freshening may provide additional opportunities to add  some technologies in these cases. In
addition, NHTSA's analysis accounts for multiyear planning-that is, the potential that
manufacturers may apply "extra" technology in an early model year with many planned
redesigns in order to carry technology forward to facilitate compliance in a later model year with
fewer planned redesigns. So, for example, Figure 13.3  suggests FCA might be expected to apply
more technology than required in MY2018  in  order to carry that technology forward to MY2019.
Further, NHTSA's analysis accounts for the potential that manufacturers could earn CAFE
credits in some model years and use those credits in later model years, thereby providing another
compliance option in years with few planned redesigns. Finally, it should be noted that neither
Figure 13.3 nor today's analysis account for future new products (or discontinued products) -
                                              13-6

-------
                                                      Analysis of Augural CAFE Standards
past trends suggest that some years in which an OEM had few redesigns may have been years
when that OEM introduced significant new products. Such changes in product offerings can
obviously be important to manufacturers' compliance positions, but cannot be systematically and
transparently accounted for with a fleet forecast extrapolated forward ten or more years from a
largely-known fleet.
Manufacturer
BMW
Daimler
FCA
Ford
General Motors
Honda
Hyundai Kia
JLR
Mazda
Mitsubishi
Nissan
Subaru
Toyota
Volvo
VWA
2016
7%
11%
0%
10%
2%
27%
25%
13%
15%
16%
4%
3%
22%
0%
8%
2017
1%
7%
24%
0%
22%
36%
26%
9%
0%
0%
4%
26%
5%
5%
15%
2018
27%
28%
48%
3%
19%
21%
14%
0%
58%
0%
33%
0%
31%
0%
43%
2019
37%
10%
5%
31%
27%
5%
9%
30%
0%
11%
21%
0%
16%
96%
4%
2020
11%
0%
19%
17%
24%
3%
19%
19%
30%
75%
0%
3%
23%
0%
18%
2021
10%
27%
12%
42%
20%
27%
17%
27%
0%
0%
30%
69%
16%
0%
4%
2022
13%
32%
7%
6%
2%
44%
49%
13%
15%
23%
14%
28%
4%
0%
18%
2023
10%
19%
4%
0%
29%
21%
6%
12%
55%
0%
25%
1%
15%
0%
15%
2024
37%
20%
0%
2%
19%
11%
8%
1%
4%
0%
14%
2%
14%
0%
21%
2025
16%
0%
11%
16%
32%
19%
26%
22%
0%
0%
9%
0%
20%
4%
7%
2026
9%
24%
39%
54%
9%
9%
28%
18%
44%
26%
13%
33%
34%
0%
20%
2027
11%
9%
18%
20%
34%
59%
32%
30%
0%
21%
20%
63%
14%
95%
19%
2028
17%
43%
3%
0%
17%
6%
3%
17%
44%
0%
28%
3%
14%
0%
32%
2029
26%
10%
4%
1%
24%
25%
15%
13%
0%
48%
23%
0%
3%
0%
1%
2030
20%
8%
4%
7%
3%
6%
33%
0%
11%
0%
13%
0%
28%
0%
9%
         Figure 13.3 Share of Manufacturer Sales Redesigned In Each Model Year 2016 - 2030

Additionally, each technology considered for application by the CAFE model is assigned to
either a "refresh" or "redesign" that dictates when it can be applied to a vehicle. Technologies
that are assigned to "refresh" can be applied at either a refresh or redesign, while technologies
that are assigned to "redesign" can only be applied during a significant vehicle  redesign. Table
13.3 and Table 13.4 (in the Technology section of the CAFE model, below) show the
technologies available to manufacturers in the compliance simulation, the level at which they are
applied (described in greater detail in both the CAFE model documentation and in Section 13.2
below), whether they available outside of a vehicle redesign, and a short description of each. A
brief examination of the tables shows that most technologies are only assumed to be available
during a vehicle redesign - and nearly all engine and transmission improvements are assumed to
be available only during redesign.  While there are past and recent examples of mid-cycle
product changes, NHTSA expects that manufacturers will tend to attempt to keep engineering
and other costs down by applying most major changes mainly during vehicle redesigns, and
some mostly modest changes during product freshening.  As mentioned below, NHTSA seeks
comment on its approach to accounting for product cadence.

   The assumptions about product cadence determine the extent to which manufacturers can
respond to increasingly stringent standards in a given a model year. When a sufficiently small
percentage of a manufacturer's sales volume is redesigned in a given model year, the
opportunities to increase its CAFE level may not be sufficient to achieve compliance. In these
situations, of which there are many (based on Figure 13.3), actions taken in earlier model years
and carried forward will have a much greater impact than actions taken in that single year. In
order to account for both the constraint of infrequent vehicle redesigns, and the accumulation and
depletion of CAFE credits resulting from these multi-year planning decisions, it is critical that
                                              13-7

-------
                                                     Analysis of Augural CAFE Standards
NHTSA simulate CAFE compliance on a year-by-year basis. NHTSA seeks comment on its
approach to accounting for product cadence in CAFE analysis.

13.1.3 Assumptions about Consumer Behavior

   While all previous CAFE analyses, including the present one supporting the Draft TAR, focus
on manufacturer actions in response to the standards, there are important considerations
regarding the impact of evaluated standards on consumer demand for new vehicles. One
limitation of all CAFE analyses up to this point is a lack of dynamic demand response to the
simulated changes in vehicle attributes - importantly, fuel economy, price, electrification level,
and perhaps curb weight - that occur as manufacturers add technology to new vehicles to comply
with standards. Currently, sales volumes at the model/variant level, for all future model years,
are an input to the CAFE model and do not respond to simulated changes in vehicle attributes.
The result of this implementation is that when a range of regulatory alternatives is examined, all
alternatives are assumed to have the same total number and sales mix of vehicle models,
regardless of the stringency of the alternative considered.

   To support the Draft TAR, NHTSA purchased a commercial forecast from IHS/Polk that
necessarily includes their assumptions about decisions manufacturers will have to make in order
to comply with standards through MY2021, as does the AEO 2015, which also informed the
production volumes used in this analysis. So any changes in market share, within a
manufacturer/segment that seems likely to occur between MY2015, which forms the basis for
Draft TAR analysis, and MY2021, when NHTSA's final standards stop increasing in stringency,
should already be present in the static volume projections at the model/variant level. However,
any volume changes that would occur as a result of post-2021 standards would not be captured
by the current approach.

   NHTSA has experimented with discrete consumer choice models, fully integrated into the
CAFE model  that revise up or down the model/variant sales, based on the changing attributes of
the vehicle and the availability of other vehicles in the market with more attractive features. A
developmental version of the CAFE model used a discrete choice model that contained a
representation of households in the U.S. and explicitly considered the way demand for given
vehicle attributes differs by household type - and the sales implications of modifying those
vehicle attributes through a program like CAFE. While testing showed promise, the current
version of the model relies on the static approach described above, for a number  of reasons.

   One important implication of relying on a discrete choice model to dynamically adjust vehicle
sales is that the concept of price becomes a driving factor. While it is also an obviously important
factor in real-world decisions about new vehicle purchases, there is no obvious definition of
price that fits  all purchases. For example, the CAFE model does not consider the value of
optional vehicle content (e.g., navigation or sound systems, luxury interior options like
heated/cooled seats, or exterior options like roof racks), yet some of these options can influence
sales price to  a greater degree than NHTSA's estimates for many new powertrain technologies. It
is also true that sales price, which can vary considerably by geographic location,  is rarely equal
to the vehicle's Manufacturer's Suggested Retail Price (MSRP) - which is all NHTSA currently
observes in the analysis fleet, and on which most consumer choice models are estimated. While
the analysis fleet has some resolution at the make/model/variant level (e.g., each engine variant
of the Honda  Civic), bundled packages and model editions that do not vary by fuel economy,
footprint, or both are unlikely to be represented in the analysis fleet. As  such, even the MSRP
                                             13-8

-------
                                                       Analysis of Augural CAFE Standards
values in the analysis fleet represent an average across model variants that, while identical for the
purposes of CAFE compliance, vary in other consumer-facing attributes in ways that strongly
influence MSRP.

   Other considerations are the pricing strategies that manufacturers employ that also influence
MSRP - often cross subsidizing vehicles in one class, or at a particular stage of design life, with
more popular vehicle models or models serving market segments with less price sensitivity.
NHTSA has considered multiple technology cost allocation (i.e. pricing) models over the last
several years, but for reporting purposes, currently implements a pay-as-you-go model where the
change in price of each vehicle model reflects the amount of additional technology content it
acquires in response to the standards. NHTSA seeks comment on these and other aspects of
consumer behavior and how to account for them.

   Still another consideration involves how manufacturers apply technology that improves
energy efficiency. Manufacturers may prefer to apply technology to improve other vehicle
attributes that consumers value if their compliance position is favorable and if that affordable
technology is available. Historical evidence is sufficient to justify the existence of consumer
preferences for vehicle size, power, or both. Yet, the CAFE model does not currently attempt to
estimate the potential that manufacturers would seek to apply fuel-saving technologies with a
view toward also improving vehicle performance or utility.2 In other words, while technology-
related inputs to the CAFE model can reflect underlying assumptions about manufacturers' likely
balancing of the potential to improve fuel economy and/or performance, the model itself does not
attempt to endogenously optimize this balance when considering the potential to apply specific
technologies to specific vehicles.  With inputs that assume manufacturers would apply
technologies such that most or all of the technical potential is used to improve fuel economy, this
could lead to a consumer choice model showing a manufacturer of already-efficient vehicles
losing market share to a rival who improves fuel economy in a cost effective manner, while
preserving  already-superior levels of performance.

   One interpretation of the current approach is that NHTSA assumes manufacturers will price
vehicle models in a way that both covers the increase in technology cost attributable to the CAFE
standards and allows them to sell the mix of vehicles that makes them the most profitable.  In
that context, NHTSA need not account for prices explicitly. This characterization implicitly
assumes that manufacturers are able to cross-subsidize the sale of less profitable models with
more profitable ones - to fully recover the cost increase without affecting the mix of vehicles
sold. While this is already current practice, NHTSA recognizes the importance  of considering the
impact of potential  standards on the ability to cross-subsidize without affecting fleet mix and
other factors.

   In the absence of satisfying resolutions to these issues, NHTSA continues to us the static
volume approach it has used in the past while it continues to refine an approach to modeling the
demand response to changing prices and attributes in the new vehicle market. However, there is
2 The current CAFE analysis, which assumes manufacturers are unlikely to reduce powertrain output except at
  relatively significant levels of mass reduction, effectively assumes that some vehicles could improve in
  performance or utility, depending in part on how technologies are shared among different vehicles. This
  approach helps to preserve the size of the initial set of engines in the MY2015 fleet. The approach does not
  generate unique engines for each variant, based on NHTSA's analysis of observed trends for managing platform
  and powertrain complexity given resource and cost considerations.
                                               13-9

-------
                                                      Analysis of Augural CAFE Standards
an area where NHTSA has attempted to capture some market behavior and its interaction with
the supply of new vehicles.
     40
     35
     25
     15
                         -Pass Car Standard   	LTStandard	Pass Car CAFE FE	LTCAFEFE
                                              Model Year
                   Figure 13.4 Industry Average CAFE and Standard 1990 - 2014

   As Figure 13.4 illustrates, the industry (though not all individual manufacturers) has exceeded
the required CAFE level for both classes in the past, though by almost 5 MPG during the fuel
price spikes of the 2000s. Worth noting is that the industry average in Figure 13.4 includes a
number of manufacturers that traditionally paid CAFE fines - some of whom reached
compliance during years with high oil prices. NHTSA attempts to account for this observed
consumer preference for fuel economy, above and beyond that required by the CAFE standard,
by allowing fuel price to influence the ranking of technologies when the model applies
technology to vehicles in order to achieve compliance. In particular, the model ranks available
technology not by cost, but by "effective cost."

   While described in greater detail in the CAFE model documentation, the effective  costs
contains an assumption not about consumers' actual willingness to pay for additional  fuel
economy, but about what manufacturers believe consumers are willing to pay. The default
assumption in the model is that manufacturers will treat all technologies that pay for themselves
within the first three years of ownership (through reduced expenditures on fuel) as if the cost of
that technology were negative. This holds true up to the point at which the manufacturer achieves
compliance with the standard - after which the manufacturer treats all technologies that pay for
themselves within the first year of ownership as having a negative effective cost. This change in
                                             13-10

-------
                                                       Analysis of Augural CAFE Standards
the pre- and post-compliance effective valuate of fuel economy is intended to serve as proxy for
manufacturers' differential willingness to risk providing "too much" fuel economy.3

   One implication of this assumption is that futures with higher, or lower, fuel prices produce
different sets of attractive technologies (and at different times). In the extreme cases, where fuel
prices are above $7 or $8/gallon, many of the technologies in this analysis could pay for
themselves within a year and appear in the baseline.  Similarly, at the other extreme, almost no
additional fuel economy would be observed.

   While these assumptions about desired payback period and consumer preferences for fuel
economy may not affect the eventual level of achieved CAFE in the later years of the program,
they will affect the amount of additional technology  cost and fuel savings that are attributable to
the standard. NHTSA seeks comment on the  approach described above, the current values it
ascribes to manufacturers' belief about consumer willingness-to-pay for fuel economy, and
suggestions for future improvements and refinements.

13.1.4  Updated Mileage Accumulation Schedules for the Draft TAR

   In order to develop new mileage accumulation schedules for vehicles regulated under the
CAFE program (classes 1-3), NHTSA purchased a data set of vehicle odometer readings from
IHS/Polk (Polk). Polk collects odometer readings from registered vehicles when they encounter
maintenance facilities, state inspection programs, or  interactions with dealerships and OEMs.
The (average) odometer readings in the data set NHTSA purchased are based on over 74 million
unique odometer readings across 16 model years (2000-2015) and vehicle classes present in the
data purchase (all registered vehicles less than 14,000 Ibs. GVW).

   The Polk data provide a measure of the cumulative lifetime vehicle miles traveled (VMT) for
vehicles, at the time of measurement, aggregated by  the following parameters: make, model,
model year, fuel type, drive type, door count, and ownership type (commercial or personal).
Within each of these subcategories they provide the average odometer reading, the number of
odometer readings in the sample from which Polk  calculated the averages, and the total number
of that subcategory of vehicles in operation.

13.1.4.1       Updated Schedules

   Figure 13.5 shows the predicted total VMT by age for the  sample of passenger cars. It also
shows the previous and current schedules together. The previous schedule was developed using
self-reported odometer data in the 2009 National Household Travel Survey (NHTS), and was the
basis for estimated travel demand in the 2012 final rule. The current schedule predicts lower
annual VMT for all ages—except the first year—but the difference increases for vehicles older
than 8 years. The resulting difference in VMT over a 30-year life of a passenger car is a decrease
of 96,882 miles under the new schedule, a 32 percent decrease from the previous schedule. A
notable trend in the new passenger car schedule is  a higher annual VMT for the first year,
3 NHTSA does not endogenously model the purchase choices of individual new vehicle buyers, nor do we attempt to
  estimate the usage profiles of individual new vehicle buyers. NHTS A's analysis currently vehicle survival and
  mileage accumulation in terms of the nationwide average of vehicles—based on millions of odometer readings
  spanning both high and low usage owners—that varies by vehicle class. It is possible that the difference between
  the total estimated benefits derived from the average usage and the sum of the true individual usage could be
  either higher or lower depending upon the fleet mix and the extent to which lower and higher fuel economy
  models are driven differently than the average.


                                              13-11

-------
                                                       Analysis of Augural CAFE Standards
followed by a relatively constant annual VMT until age 6 (MY 2014 to MY 2008, for our
sample). This trend is likely a byproduct of the patterns of commercial and personal vehicle
ownership over the age of vehicles, although other factors (e.g., fuel prices, employment levels,
GDP, typical length of a new car loan) could underlie the steep decline in average annual
mileage accumulation after vehicles have been in operation for 6 years.
            Passenger Car Milage Accumulation
                   (Data and Model)
New and Old Passenger Car Schedules
                              •
                                         ••
                                                                    ••
            Figure 13.5 A Comparison of the Current and Previous Passenger Car Schedules

   Figure 13.6 shows the share of passenger cars registered between commercial and personal
fleets, and the population-weighted average odometer reading by ownership type. Commercial
vehicles are driven more than personally-owned vehicles, and make up the largest share of one-
year-old vehicles, relative to other ages. Since a model year of vehicles is sold starting in the fall
of the previous calendar year, throughout the matching calendar year, and into the succeeding
one, this initial proportion suggests that (in proportion to fleet share) more commercially-owned
vehicles are bought early. Another partial explanation is likely that commercial vehicles are sold
into the personal fleet after a short time. Regardless of the cause, this pattern of ownership likely
explains why the first year annual VMT is higher than other years: the share of more heavily-
driven commercial vehicles is highest for age one vehicles, and we weight the models by the
proportion each makes up of the total population of registered vehicles. The SUV/Van and light-
duty truck class fleets show similar patterns of more-heavily driven commercial vehicles, and the
highest share of commercial vehicles occurring for one-year-old vehicles. Unsurprisingly, the
initial peak of annual VMT occurs for these classes as well.
                                              13-12

-------
                                                       Analysis of Augural CAFE Standards

J -
o -


Ownership Type Population Share for Passenger Cars
•
•
•
•

0 5 10 15
AGE
• Commercial Fleet • Personal Fleet






-
0 -


VMT by Ownership Type for Passenger Cars
.,•••"
• •
• •
• •
• •
• •
•
:
•
•
:

0 5 10 15
AGE
• Commercial Fleet • Personal Fleet
         Figure 13.6 Total VMT and Share of Population by Ownership Type for Passenger Cars
   The old SUV and van schedules are very similar (Figure 13.7). Since the Polk data is already
aggregated to the model-level, there are 38 categories of vans in 2014. For all other classes there
are at least three times as many model-level classifications. For these reasons, we determined
that vans and SUVs were sufficiently similar, and merged them into a single class for VMT
purposes.  The new SUV/Van schedule shows a peak average annual VMT (16,035) occurring at
age one. It predicts lower annual VMT for all ages (except the first year, which is slightly higher
than the old SUV schedule, though still predicts lower annual VMT than the old van schedule).
The new schedule predicts a total of 101,023 (30 percent) fewer miles driven over a 30-year
lifespan than the old SUV schedule, and a total of 124,859 (34 percent) fewer miles driven over a
30-year lifespan than the old van schedule.
                SUV/Van Milage Accumulation
                     (Data and Model)
New and Old SUV/Van Schedules
                                 10
                                             15
                          Age
                                                                                     JO
              Figure 13.7 A Comparison of the Current and Previous SUV/Van Schedules
   The new light-duty pickup schedule predicts a peak annual VMT of 17,436 miles at age one.
Figure 13.8 shows that the new light-duty pickup VMT schedule predicts higher annual VMT for
ages one through five, and lower annual VMT for all other ages. Even considering this, the new
                                              13-13

-------
                                                       Analysis of Augural CAFE Standards
schedule for light pickups predicts a total 30-year lifetime decrease of 95,133 (26 percent) from
the old schedule for light trucks.
             Light Pickup Milage Accumulation
                    (Data and Model)
     New and Old Light Pickup Schedules
                                                                                       30
            Figure 13.8 A Comparison of the Current and Previous Pickup Truck Schedules
   The new medium-duty van/pickup schedule in Figure 13.9 predicts higher annual VMT for
vehicles between ages one through five years, and lower annual VMT for all other vehicle ages,
than the old schedule. Over the first 30-year span, the new schedule predicts that medium-duty
vans/pickups drive 24,249 (9 percent) fewer miles than the old  schedule. We predict the
maximum average annual VMT for medium-duty vehicles (23,307 miles) at age two. The pattern
of the share of commercially and personally owned vehicles (see Figure 6) is qualitatively
different than the other classes, and offers a potential explanation for the maximum annual VMT
occurring at age two.
         Medium-Duty Pickup/Van Milage Accumulation
                    (Data and Model)
                               10
                                           15
New and Old Medium-Duty Pickups/Vans Schedules
                         Age
                                                                                        30
           Figure 13.9 A Comparison of the Current and Previous MD Pickup/Van Schedules
   Figure 13.10 shows that while the maximum share of commercially-owned vehicles occurs at
age one, the registration population-weighted average odometer reading for personally and
commercially owned vehicles are almost identical for this age. However, the share of
                                              13-14

-------
                                                      Analysis of Augural CAFE Standards
commercially-owned vehicles is higher for age two vehicles than all older ages, and there is a
larger spread between the average odometer readings of the two ownership types for this age of
vehicle (while the spread between the average odometer readings for age three is even larger, the
share of commercially-owned vehicles is smaller, and likely counteracts this effect in the
registration population-weighted models). This increase in the difference between the average
odometer reading of the ownership types can explain the peak annual VMT at age two.
0 -
VMT by Ownership Type for Medium-duty Vans/Pickups
./""'
0 5 10 15
AGE
• Commercial Fleet • Personal Fleet

Owr
CO -
CD
CM
lership Type Population Share for Medium-duty Vans/Pickups
9 9
t
9 •
0 5 10 15
AGE
. Commercial Fleet • Personal Fleet
       Figure 13.10 Total VMT and Share of Population by Ownership Type for MD Pickups/Vans
   Table 13.1  Summary Comparison of Lifetime VMT for Current and Previous Schedules
offers a summary of the comparison of lifetime VMT (by class) under the new schedule,
compared with lifetime VMT under the old schedule. In addition to the total lifetime VMT
expected under each schedule for vehicles that survive to their full expected life, Table 13.1 also
shows the survival-weighted lifetime VMT for both schedules.  This represents the average
lifetime VMT  for all vehicles, not only those that survive to their full  expected life. The
percentage difference between the two schedules is not as stark for the survival-weighted
schedules: the  percentage decrease of survival-weighted lifetime VMT under the new schedules
range from 6.5 percent (for medium-duty trucks and vans) to 21.2 percent (for passenger vans).
         Table 13.1 Summary Comparison of Lifetime VMT for Current and Previous Schedules


Car
Van
SUV
Pickup
2b/3
Lifetime VMT
Current
204,233
237,623
237,623
265,849
246,413
Previous
301,115
362,482
338,646
360,982
270,662
% difference
32.2%
34.4%
29.8%
26.4%
9.0%
Survival-Weighted
Lifetime VMT
Current
142,119
155,115
155,115
157,991
176,807
Previous
179,399
196,725
193,115
188,634
189,020
% difference
20.8%
21.2%
19.7%
16.2%
6.5%
                                             13-15

-------
                                                      Analysis of Augural CAFE Standards
13.1.4.2      Data Description

  While the Polk data set contains model-level average odometer readings, the CAFE model
assigns lifetime VMT schedules at a lower resolution based on vehicle body style. For the
purposes of VMT accounting, the CAFE model classifies every vehicle in the analysis fleet as
being one of the following: passenger car, SUV, pickup truck, passenger van, or medium-duty
pickup/van. In order to use the Polk data to develop VMT schedules for each of the (VMT)
classes in the CAFE model, we constructed a mapping between the classification of each model
in the Polk data and the classes in the CAFE  model. The only difference between the mapping
for the VMT schedules and the rest of the CAFE model is that we merged the SUV and van body
styles into one class (for reasons described in our discussion of the SUV/van schedule above).
This mapping allowed us to predict the lifetime miles traveled, by the age of a vehicle, for the
categories in the CAFE model.

  In estimating the VMT models, we weighted each data point (make/model classification) by
the share of each make/model in the total population of the corresponding CAFE class. This
weighting ensures that the predicted odometer readings, by class and model year, represent each
of vehicle classification among observed vehicles (i.e., the vehicles for which Polk has odometer
readings), based on each vehicles' representation in the registered vehicle population of its class.
Implicit in this weighting  scheme, is the assumption that the samples used to calculate each
average odometer reading by make, model, and model  year are representative of the total
population of vehicles of that type. Several indicators suggest that this is a reasonable
assumption.

  First, the majority of each vehicle make/model is well-represented in the sample. Histograms
and empirical cumulative  distribution functions (CDF's) of the ratio of the number of odometer
readings to the total population of those makes/models by each class (Figure 13.11, below), show
that for more than 85 percent  of make/model combinations, the average odometer readings are
collected for 20 percent or more of the total  population. Most make/model observations have
sufficient sample sizes, relative to their representation in the vehicle population, to produce
meaningful average odometer totals at that level4.
4 We developed similar figures, stratified by each vehicle class, but these were no more revealing than the figures for
  all vehicles.
                                             13-16

-------
                                                        Analysis of Augural CAFE Standards
      Ratio of the Number Odometer Readings to Population
Cumulative Representation of Odometer Readings
                                                          .2       .4       .6       .8       1
                                                       Ratio of Odometer Readings to Population [Representation]
      Figure 13.11  Distribution of the Ratio of Sample Size to Population Size (by Make/Model/MY)
   We also considered whether the representativeness of the odometer sample varies by vehicle
age, since VMT schedules in the CAFE model are specific to each age. To investigate, we
calculated the percentage of vehicle types (by make, model, and model year) that did not have
odometer readings. Figure 13.12 shows that all model years, apart from 2015, have odometer
readings for 96 percent or more of the total types of vehicles observed in the fleet.

                           Population with No Odometer Readings by Model Year
                          2000
                                       2005           2010
                                            Model Year
                                                                  2015
   Figure 13.12 Percentage of Total Vehicle Population with No Odometer Readings across Model Years
   While the preceding discussion supports the coverage of the odometer sample across
makes/models by each model year, it is possible that, for some of those models, an insufficient
number of odometer readings is recorded to create an average that is likely to be representative
of all of those models in operation for a given year. Figure 13.13 below shows the percentage of
all vehicle types for which the number of odometer readings is less than 5 of the total population
(for that model). Again, for all model years other than 2015, about 95 percent or more of vehicles
                                               13-17

-------
                                                       Analysis of Augural CAFE Standards
types are represented by at least 5 percent of their population. For this reason, we included
observations from all model years, other than 2015, in the estimation of the new VMT schedules.

                          Population with Fewer than 5% of Odometer Readings
                         2000
                                       2005            2010
                                            Model Year
                                                                   2015
  Figure 13.13 Percentage of Vehicle Models with Fewer than 5% of the Population in Odometer Readings
                                       Data (by Class)
   It is possible that the odometer sample is biased. If certain vehicles are over-represented in the
sample of odometer readings relative to the registered vehicle population, a simple average, or
even one weighted by the number of odometer observations will be biased.  However, while
weighting by the share of each vehicle in the population will account for this bias, it would not
correct for a sample that entirely omits a large number of makes/models within a model year. We
tested for this by computing the proportion of the count of odometer readings for each individual
vehicle type—within a class and model year—to the total count of readings for that class and
model year. We also compared the population of each make/model—within each class and
model year—to the population of the corresponding class and model year. The difference of
these two ratios shows the difference of the representation of a vehicle type—in its respective
class and model year—in the sample versus the population (summarized in Figure 13.14, below).
All vehicle types are represented in the sample within 10 percent of their representation in the
population, and the variance between the two representations is normally distributed.  This
suggests that, on average, the likelihood that a vehicle is in the sample is comparable  to its
proportion in the relevant population, and that there is little under or over sampling of certain
vehicle makes/models.5
5 We produced similar figures, stratified by class, but these were no more revealing; the only difference being that
  cars are represented in the sample within 5% of their representation in the population (with a distribution range of
  .05 on either side).
                                              13-18

-------
                                                      Analysis of Augural CAFE Standards
                            Difference in Representation in Sample vs. Population
                         S -
                                    -.05       0       .05
                                       Difference in Representation
   Figure 13.14 Difference in Share of Each Vehicle Model in Population vs. Odometer Sample (by Class)
13.1.4.3
Estimation
   Since model years are sold in the fall of the previous calendar year, throughout the same
calendar year, and even into the following calendar year—not all registered vehicles of a
make/model/model year will have been registered for at least a year (or more) until age 3. The
result is that some MY2014 vehicles may have been driven for longer than one year, and some
less, at the time the odometer was observed. In order to consider this in our definition of age, we
assign the age of a vehicle to be the difference between the average reading date of a
make/model and the average first registration date of that make/model. The result is that the
continuous age variable reflects the amount of time that a car has been registered at the time of
odometer reading, and presumably the time span that the car has accumulated the miles.

   After creating the "Age" variable, we fit the make/model lifetime VMT data points to a
weighted quartic polynomial regression of the age of the vehicle (stratified by class). The
predicted values of the quartic regressions are used to calculate the marginal annual VMT by age
for each class by calculating differences in estimated lifetime mileage accumulation by age.
However, the Polk data acquired by NHTSA only contains observations for vehicles newer than
16 years of age. In order to estimate the schedule for vehicles older than the age 15 vehicles in
the Polk data, we combined information about that portion of the schedule from the VMT
schedules used in both the 2017-2021 Final Light Duty Rule and 2019-2025 Medium-Duty
NPRM. The light-duty schedules were derived from the survey data contained in the 2009
National Household Travel Survey (NHTS) and the 2001 Vehicle In Use Survey  (VIUS), for
medium-duty trucks.

   Based on the vehicle ages for which we have data (from the Polk purchase), the newly
estimated annual schedules differ from the previous version in important ways. Perhaps most
significantly, the annual  mileage associated with ages beyond age 8 begin to, and continue to,
trend much lower. The approach taken here attempts to preserve the results obtained through
estimation on the Polk observations, while leveraging the existing (NHTS-based) schedules to
support estimation of the higher ages (age 16  and beyond). Since the two schedules are so far
apart, simply splicing them together would have created not only a discontinuity, but also
                                             13-19

-------
                                                      Analysis of Augural CAFE Standards
precluded the possibility of a monotonically decreasing scale with age (which is consistent with
previous schedules, the data acquired from Polk, and common sense).

   In the 2009 NHTS survey, VMT per vehicle decreases steadily as household vehicles age,
though with declining samples sizes for the oldest vehicles.  The Polk data show an annual VMT
increase for the oldest vehicles. In order to force the expected monotonicity, we perform a
triangular smoothing algorithm until the schedule is monotonic. This performs a weighted
average which weights the observations close to the observation more than those farther from it.
The result is a monotonic function, which predicts similar lifetime VMT for the sample span as
the original function. Since the Polk data does not show vehicles greater thanlS years of age, we
are not able to correctly capture that part of the annual VMT curve using only the new dataset.
For this reason, we use trends in the old data to extrapolate the new schedule for ages beyond the
sample range.

   In order to use the VMT information from the newer data source for ages outside of the
sample, we use the final in-sample age (15 years) as a seed and then apply the annual VMT
decline from the old schedules to extrapolate the new schedules out to age 30. To do this, we
calculated the annual percentage difference in VMT of the old schedule for ages 15-30. The
same annual percentage difference in VMT is applied to the new schedule to extend beyond the
final in-sample value. This assumes that the overall proportional trend in the outer years is
correctly modeled in the old VMT schedule, and imposes this same trend for the outer years of
the new schedule. The extrapolated schedules are the final input for the VMT schedules in the
CAFE model.

   Older vehicles are not well represented, even in the NHTS, where sample sizes for these
vehicles are very small.  This is an area that would benefit from further research.

13.1.4.4      Comparison to previous schedules

   New VMT data suggest lower lifetime mileage accumulation rates than the VMT schedule
used in the last Light-Duty CAFE Final Rule, particularly for higher vehicle ages.  The previous
schedules are based on self-reported odometer readings that were acquired during  a period of
economic and fuel price volatility, while the observations from Polk are between 5 and 7 years
newer than those in the NHTS and represent observed  odometer readings (rather than self-
reported information).

   Additionally, NHTSA finds the Polk data,  which provides a much larger representative
sample of some 70 million vehicles preferable to the previous schedule, which relied on the
NHTS's representative sample of about 200,000 households. However, by properly accounting
for vehicle population weights in the new averages and models, we  corrected for this issue in the
derivation of the new schedules.

   Sample surveys have inherent limitations.  While the NHTS is carefully designed to be a
representative  sample of households, it may not be a representative  sample of vehicles.  Since the
NHTS only samples households, it does not detect the  differing driving patterns of commercially
registered vehicles, which turn out to be particularly important for new vehicles and for medium
pick-ups.  It seems likely that there is another previously undetected phenomenon: there may be
many older light duty vehicles that retain their registration but are little driven from one year to
                                             13-20

-------
                                                      Analysis of Augural CAFE Standards
the next. These vehicles, if they exist, were not detected by the NHTS survey.  This is an
uncertainty that could be clarified by further research.

   Both the previous and current schedules are limited by the nature of the data on which they
are based. Each schedule relies upon a single snapshot in time, then treats the cross-section of
vehicle ages as if it were a panel - observations about the same set of vehicles as they age. This
is done out of necessity, but can  clearly bias estimates of mileage accumulation. In the case of
the NHTS, older vehicles would have experienced nearly a decade of strong economic growth
and historically low fuel prices -perhaps inflating VMT relative to today. In the case of the Polk
sample, vehicles would have experienced prolonged periods of both fuel price instability and
economic distress (the years from 2007 - 2010, though continuing longer for certain age cohorts
that remained chronically underemployed for a longer period of time) - perhaps depressing VMT
relative to today. These biases cannot even be detected with a single year of data,  and NHTSA
intends to take steps in the future to improve the resources on which the schedules are estimated.

13.1.4.5      Future direction

   In consultation with other agencies closely involved with VMT estimation (e.g., FHWA),
NHTSA will continue to seek means to further refine estimated mileage accumulation schedules.
For example, one option under consideration would be to obtain odometer reading data from
successive calendar years, thus providing a more robust basis to consider, for example,  the
influence of changing fuel prices or economic conditions on the accumulation of miles by
vehicles of a given age.

   NHTSA seeks comment on the information and methods used to develop today's odometer-
based estimates of annual mileage accumulation schedules, recommendations regarding any
other methods to estimate such schedules,  and information that could be used to refine these
schedules or develop and implement alternative methods.

13.1.5 Other Assumptions of Note

   There are a number of additional assumptions that influence both the simulation of
manufacturers' compliance decisions and the estimated benefits and costs resulting from the
standards - among them are technology cost and effectiveness, both discussed in greater detail in
chapter 5 of the Draft TAR. One assumption that warrants additional discussion are fuel prices.

   Few inputs touch as many aspects of the analysis as fuel prices;  they are a primary driver of
the value of fuel savings (which  is the largest single benefit of the program), they influence the
projected share of light trucks in the new vehicle market, the ranking of technologies by
manufacturers in the compliance simulation (discussed more later), the amount of additional fuel
economy demanded by the market in the absence of regulatory pressure,  and the magnitude of
the rebound effect that generates additional vehicle miles traveled when fleet fuel economy
improves. Yet, over the increasingly long time horizons of recent CAFE  analyses (the Draft TAR
analysis covers the full useful lives of vehicles produced between model  years 2015 and 2032,
and the Final Rule analysis covered the full useful lives  of vehicles produced between model
years 2011 and 2025 - necessitating fuel price estimates out as far as 2060), the uncertainty in
fuel price projections becomes increasingly important. In Figure 13.15, we see a comparison of
oil price projections from the Annual Energy Outlook compared to the actual average price
                                             13-21

-------
                                                        Analysis of Augural CAFE Standards
observed in a given year. The green cells represent underestimates, while the blue cells highlight
overestimates.
Projected vs. Actual
                      1995  1996
                                                2002  2003  2004  2005  2006  2007  2008  2009  2010  2011
                                                     3   3.
                                                               -557
                                                               -444
                  Figure 13.15 Retrospective Analysis of EIA Fuel Price Projections

   As Figure 13.15 shows, projections of years farther in the future tended to be significantly
different from observed prices. Also of note is the fact that long-term underestimation continued
for a number of years after observed price increases - suggesting that the forecasting model is
slow to adapt to regime changes. In general, this stability may be advantageous; a model that is
too reactionary could produce large swings between iterations of the AEO and present
projections that are too "noisy" for planning purposes. However, if longer-term prices are
significantly different from prices over the last 8-10 years, current forecasts could overstate or
understate future oil prices.  There is  inherent uncertainty in future fuel prices, and updates to
forecasts will continue to integrate current information as it becomes available, which will
continue to impact future CAFE analysis.

   As discussed elsewhere in this document, the global oil market has experienced  a period of
rapid and dramatic change since the final rule was published in 2012. The fuel price estimates in
the AEO reflect these changes. As Figure  13.16 illustrates, the recent decline in fuel prices
represents a deviation from the projections used in the 2012 final rule analysis. However, as
discussed above, the long term trend is roughly consistent  with the older forecast but starts from
a lower point. And while these lower prices are likely to increase demand relative to a higher
price scenario, each gallon saved results in a lower value of fuel savings to consumers as a result
of the drop in per-gallon price relative to the 2012 FR analysis.
                                               13-22

-------
                                                     Analysis of Augural CAFE Standards
           3.5
         ID
         ao
         i£i
         0)
         .y 2.5
         Q.
         "> 1 r
         ro 1.5
         ID
           0,5.
               ooooooooooooooooooooooooooo
                          • 2012 FR price projection
•TAR price projection
       Figure 13.16 Comparison of Fuel Price Estimates in Draft TAR and 2012 Final Rule Analysis
13.2   CAFE Model (aka "Volpe Model") Overview and Updates Since the
2012 Final Rule

   This analysis reflects several changes made to the model since 2012, when NHTSA used the
model to estimate the effects, costs, and benefits of final CAFE standards for light-duty vehicles
produced during MYs 2017-2021, and Augural  Standards for MYs 2022-2025.  Some of these
changes specifically enable analysis of potential fuel consumption standards (and, hence, related
CCh emissions standards harmonized with fuel consumption standards) for heavy-duty pickups
and vans; other changes implement more general improvements to the model. Key changes
relevant to today's analysis include the following:

          •  Expansion of model inputs, procedures, and outputs to accommodate technologies
             not included in prior analyses.
          •  Changes to the algorithm used to apply technologies,  enabling more explicit
             accounting for shared vehicle platforms and adoption and "inheritance" of major
             engine changes.
          •  Expanded accounting for CAFE  credits carried over from years prior to those
             included in the analysis fleet (a.k.a. "banked" credits).
          •  Changes to the model's approach to estimating the effect of combinations of fuel-
             saving technologies.
                                            13-23

-------
                                                      Analysis of Augural CAFE Standards
13.2.1  Updates to 2012 Final Rule Version of the CAFE Model

   After the light-duty rulemaking analysis accompanying the 2012 final rule that finalized
NHTSA's standards through MY2021, NHTSA staff began work on changes to the CAFE model
with the intention of better reflecting constraints of product planning and cadence for which
previous analyses did not account. These changes, summarized below, interact with preexisting
model characteristics discussed above. Additionally, NHTSA fully integrated the results of a
simulation database constructed by Argonne National Laboratory and described in Chapter 5
(Section 5.4.2.4). While the technologies, assumptions, and experimental design are discussed in
chapter 5, the integration into the CAFE model is discussed below.

   Engine and Transmission Sharing and Inheritance

   In practice, manufacturers are limited in the number of engines and transmissions that they
produce.  Typically a manufacturer produces a number of engines—perhaps six or eight engines
for a large manufacturer—and tunes them for slight variants in output for a variety of car and
truck applications. Manufacturers limit complexity in their engine portfolio for much the same
reason as they limit complexity in vehicle variants: they face engineering manpower limitations,
and supplier, production and service costs that  scale with the number of parts produced.

   In previous analyses that used the CAFE model, engines and transmissions in individual
models were allowed relative freedom in technology application, potentially leading to solutions
that would, if followed, create many more unique engines and transmissions that exist in the
analysis fleet (or in the market) for a given model year. This multiplicity likely failed to
sufficiently account for costs associated with such increased complexity in the product portfolio,
and may have represented an unrealistic diffusion of products for manufacturers that are
consolidating global production to increasingly smaller numbers of shared engines and platforms
(cite NAS here).  The lack of a constraint in this area allowed the model to apply  different levels
of technology to the engine in each vehicle at the time of redesign or refresh, independent of
what was done to other vehicles using a previously identical engine.

   In the current version of the CAFE model, engines and transmissions that are shared between
vehicles must apply the same levels of technology, in all technologies, dictated by engine or
transmission inheritance.  This forced adoption is referred to as "engine inheritance" in the
model documentation.

   In practice, the model first chooses an "engine leader" among vehicles sharing the same
engine.  The leader is selected first by the vehicle with the lowest average sales across all
available model years. If there is a tie, the vehicle with the highest average MSRP across model
years is chosen. The model applies the same logic with respect to the application of transmission
changes.  The model follows this formulation due to previous market trends suggesting that
many technologies begin deployment at the high-end, low-volume  end  of the market as
manufacturers build their confidence and capability in a technology, and later expand the
technology across more mainstream product lines.

   NHTSA received comments specific to its approach to accounting for shared engines and
transmissions, although comments from  some environmental organizations cited examples of
sharing between light- and heavy-duty products. NHTSA has continued to refine its
implementation of its approach to accounting for shared engines and transmissions, and again
                                             13-24

-------
                                                      Analysis of Augural CAFE Standards
seeks comment on the approach, recommendations regarding any other approaches, and any
information that would facilitate implementation of the agency's current approach or any
alternative approaches.

   Platforms, Sharing, and Technology

   The term "platform" is used loosely in industry, but generally refers to a common structure
shared by a group of vehicle variants. The degree of commonality varies, with some platform
variants exhibiting traditional "badge engineering" where two products are differentiated by little
more than insignias, while other platforms be used to produce a broad suite of vehicles that bear
little outer resemblance to one another.

   Given the degree of commonality between variants of a single platform, manufacturers do not
have complete freedom to apply technology to a vehicle: while some technologies (e.g. low
rolling resistance tires) are very nearly "bolt-on" technologies, others involve substantial changes
to the structure and design of the vehicle, and therefore necessarily are constant among vehicles
that share a common platform. NHTSA staff has,  therefore, modified the CAFE model such that
all levels of mass reduction  and aerodynamic improvement are forced, over time, to be constant
among variants of a platform.  However, because these levels are not concretely defined in terms
of specific engineering changes, and the vehicle models in the analysis fleet are not defined in
terms of specific engineering content, this aspect of the CAFE model does not mean that every
vehicle model on a platform necessarily receives identical engineering changes to attain the same
level of aerodynamic improvement or mass reduction.  Also, with the application of these
improvements tied to vehicle redesign or freshening, some vehicle models on a shared platform
may inherit them from platform "leaders."

   Within the analysis fleet, each vehicle is associated with a specific platform.  Similar to the
application of engine and transmission technologies, the CAFE model defines a platform
"leader" as the vehicle variant of a given platform that has the highest level of observed  mass
reduction and aerodynamic technologies present in the analysis fleet. If there is a tie, the CAFE
model begins applying aerodynamic and mass reduction technology to the vehicle with the
lowest average sales across  all available model years. If there remains a tie, the model begins by
choosing the vehicle with the highest average MSRP across all available model years. As the
model applies technologies, it effectively levels up all variants on a platform to the highest level
of (mass and aerodynamic) technology on the platform.

   In the 2015 NPRM proposing new fuel consumption and GHG standards for heavy-duty
pickups and vans, NHTSA specifically requested comment on the general use of platforms
within CAFE rulemakings.  While the agency received no responses to this specific request,
comments from some environmental organizations cited examples of technology sharing
between light- and heavy-duty products. NHTSA has continued to refine its implementation of
an approach accounting for  shared platforms, and again seeks comment on the approach,
recommendations regarding any other approaches, and any information that would facilitate
implementation of the agency's current approach or any alternative approaches.

   Interactions between Regulatory Classes

   Like earlier versions, the current CAFE model provides for integrated analysis spanning
different regulatory classes, accounting both for standards that apply separately to different
                                             13-25

-------
                                                      Analysis of Augural CAFE Standards
classes and for interactions between regulatory classes. Light vehicle CAFE standards are
specified separately for passenger cars and light trucks. However, there is considerable sharing
between these two regulatory classes - where a single engine, transmission, or platform can
appear in both the passenger car and light truck regulatory class. For example, some sport-utility
vehicles are offered in 2WD versions  classified as passenger cars and 4WD versions classified as
light trucks. Integrated analysis of manufacturers' passenger car and light truck fleets provides
the ability to account for such sharing and reduce the likelihood of finding solutions that could
involve introducing impractical levels of complexity in manufacturers' product lines.
Additionally, integrated analysis provides the ability to simulate the potential that manufacturers
could earn CAFE credits by over complying with one standard  and use those credits toward
compliance with the other standard  (i.e., to simulate credit transfers between regulatory classes).
This is discussed further below.

   FID pickups and vans are regulated separately from light-duty vehicles. While manufacturers
cannot transfer credits between light-duty and MDHD classes, there is some sharing of
engineering and technology between light-duty vehicles and HD pickups and vans. For example,
some passenger vans with GVWR over 8,500 pounds are  classified as medium-duty passenger
vehicles (MDPVs) and are thus included in manufacturers' light-duty truck fleets, while cargo
vans sharing the  same nameplate are classified as HD vans.  NHTSA has also identified several
engines (across all manufacturers) that are shared between the light-truck and HD pickup and
van classes.

   Today's analysis uses an overall analysis fleet spanning both the light-duty and HD pickup
and van fleets. As discussed below, doing so shows some technology "spilling over" to HD
pickups and vans due, for example, to the application of technology in response to current light-
duty standards.  For most manufacturers, these interactions appear relatively small. For Nissan,
however, they appear considerable, because Nissan's heavy-duty vans use engines also used in
Nissan's light-duty SUVs. Daimler also exhibits  significant  levels of component sharing between
its MDHD and light-duty fleets, but is not sufficiently constrained by the upcoming MDHD
CAFE standards to expect technology migration into the light-duty fleet as a result of the
regulations.

   In the NPRM proposing new standards for heavy-duty pickups  and vans NHTSA and EPA
commented on the expansion of the analysis fleet such that the  impacts of new HD pickup and
van standards can be estimated within the context of an integrated analysis of light-duty vehicles
and HD pickups  and vans, accounting for interactions between  the fleets. As mentioned above,
some environmental organizations specifically cited commonalities and overlap between light-
and heavy-duty products.  NHTSA seeks comment on the approach it has developed to account
for such sharing, recommendations  regarding any other approaches, and any information that
would facilitate implementation of the agency's current approach or any alternative approaches.

   Phase-In Caps

   The CAFE model retains the  ability to use phase-in caps  (specified in model inputs) as
proxies for a variety of practical restrictions on technology application, including the
improvements described above.  Unlike vehicle-specific restrictions related to redesign, refreshes
or platforms/engines, phase-in caps constrain technology application at the vehicle manufacturer
level for a given  model year. Introduced in the 2006 version of the CAFE model, they were
intended to reflect a manufacturer's overall resource capacity available for implementing new
                                             13-26

-------
                                                      Analysis of Augural CAFE Standards
technologies (such as engineering research and development personnel and financial resources),
thereby ensuring that resource capacity is accounted for in the modeling process.

   Compared to prior analyses of light-duty standards, these model changes result in some
changes in the broad characteristics of the model's application of technology to manufacturers'
fleets.  Since the use of phase-in caps has been de-emphasized and manufacturer technology
deployment remains tied strongly to estimated product redesign and freshening schedules,
technology penetration rates may jump more quickly as manufacturers apply technology to high-
volume products in their portfolio.

   In previous CAFE rulemakings, redesign/refresh schedules and phase-in caps were the
primary mechanisms to reflect a manufacturer's limited pool of available resources during the
rulemaking time frame and the years preceding it, especially in years where many models may
be scheduled for refresh or redesign. The newly-introduced representation of platform-, engine-,
and transmission-related considerations discussed above augment the model's preexisting
representation of redesign cycles, and eliminate the need to rely on phase-in caps.  By design,
restrictions that enforce commonality of mass reduction and aerodynamic technologies on
variants of a platform, and those that enforce engine inheritance, will result in fewer vehicle-
technology combinations in a manufacturer's future modeled fleet. NHTSA seeks comment
regarding this shift away from relying on phase-in caps and, if greater reliance on phase-in caps
is recommended, what approach and information can be used to define and apply these caps.

   Accounting for CAFE Credits

   The changes discussed above relate specifically to the model's approach to simulating
manufacturers' potential addition of fuel-saving technology in response to CAFE standards and
fuel prices within an explicit  product planning context.  The model's approach to simulating
compliance decisions also accounts for the potential to earn and use CAFE credits, as provided
by EPCA/EISA. Like past versions, the current CAFE model can be used to simulate credit
carry-forward (a.k.a. banking) between model years and transfers between the passenger car and
light truck fleets, but not credit carry-back (a.k.a. borrowing) between model years or trading
between manufacturers.  Unlike past versions, the current CAFE model provides a basis to
specify (in model inputs) CAFE credits available from model years earlier than those being
simulated explicitly. For example, with today's analysis representing  model years 2015-2032
explicitly, credits specified as being available from model year 2014 are made available for use
through model year  2019 (given the 5-year limit on carry-forward of credits).

   As discussed in the CAFE model documentation6, the model's default logic attempts to
maximize credit carry-forward—that is to "hold on" to credits for as long as possible. Although
the model uses credits before they expire if a manufacturer needs to cover a shortfall that occurs
when insufficient opportunities exist to add technology in order to achieve compliance with a
standard, the model  will otherwise carry forward credits until they are within 2 years of
expiration, at which point it will use them before adding technology. The model always applies
expiring credits before applying technology in a given model year, but attempts to use credits
that will expire within the next three years as a means to smooth out technology application over
time to avoid both shortfalls and high levels of over-compliance that can result in a surplus of
' Available at: http://www.nhtsa.gov/fuel-economy
                                             13-27

-------
                                                      Analysis of Augural CAFE Standards
credits. As further discussed in the CAFE model documentation, model inputs can be used to
adjust this logic to shift the use of credits ahead by one or more model years.

   NHTSA recently introduced the CAFE Public Information Center (at
                     'A • r  ua 7CAFJE ..... JPTCJiomeJitm) to provide public access to a range of
information regarding the CAFE program, including manufacturers' credit balances.  Having
reviewed credit balances (as of January 23, 2016) and estimated the potential that some
manufacturers could trade credits, NHTSA developed inputs for today's analysis that make
carried-forward credit available as summarized below, after subtracting credits assumed to be
traded to other manufacturers, and adding credits assumed to be acquired from other
manufacturers through such trades.  NHTSA seeks comment regarding the model's
representation of the CAFE credit provisions, recommendations regarding any other options, and
any information that could help to refine the current approach or develop and implement an
alternative approach.
    Table 13.2 CAFE Credits Estimated to be Available from 2010-2014 (1 vehicle x 0.1 mpg = 1 credit)

BMW
Daimler
FCA
Ford
General Motors
Honda
Hyundai Kia
JLR
Mazda
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volvo
VWA
Passenger Car
2010
1,867,281

2,876,264
36,375,648
27,631,650
64,652,589
47,621,472
-
13,387,185
1,925,910
-
2,198,848

169,026,869
-
15,911,604
2011
5,484,006
3,565,752
42,336,994
33,608,823
48,958,466
18
12,088,388
731,304
504,080
1,100,080
-
118,040

18,459,036
316,089
18,824,971
2012
6,487,815
3,959,432
51,750,678
42,075,418
27,741,179
2,045,973
24,961,094
867,378
1,062,098
1,602,650
4,917,773
1,579,019
1,039,207
33,398,277
45,579
18,193,147
2013
8,653,773
4,897,035
64,726,258
72,048,358
42,650,469
9,826,880
45,456,981
1,380,529
1,380,624
2,401,174
9,551,573
4,967,329
159,008
32,011,519
818,184
32,795,905
2014
13,678,596
458,100
4,182,307
64,729,568
47,350,779
1,290,074
30,988,589
847,794
180,964
4,281,902
618,917
4,740,723
514,937
3,306,679
-
34,158,829
Light Truck
2010


-
7,587,839
23,344,950
16,271,310
6,256,961
-
3,150,208
783,180
4,247,124
11,317,086

22,424,142
-
719,074
2011
39,458
160,528
5,553,261
6,551,119
4,983,427

3,566,052
148,329


194,670
145,270

7,817,895
62,876
994,291
2012
24,674
120,002
5,088,698
1,158,854
570,140

1,192,473
108,544


88,218
-

574,879
-
294,668
2013
163,927
404,128
1,461,785
5,747,065
1,988,083

616,827
395,626

508,898
-
1,839,959

1,742,995
-
1,672,648
2014
749,703

-
4,634,359
15,118,329

1,129,148
844,612

1,282,604
-
5,211,684

-
235,285
2,783,619
13.2.1.1
Integrating Vehicle Simulation Results into the CAFE Model
   In previous versions of the CAFE Model, technology effectiveness values entered into the
model as a single number for each technology (for each of several classes), intended to represent
the incremental improvement in fuel consumption achieved by applying that technology to a
vehicle in a particular class. At a basic level, this implied that successive application of new
vehicle technologies resulted in an improvement in fuel consumption (as a percentage) that was
the product of the individual incremental effectiveness of each technology applied. Since this
construction fails to capture interactive effects - cases where a given technology either improves
or degrades the impact of subsequently applied technologies - the CAFE Model applied
"synergy factors." The synergy factors were defined for a relatively small number of technology
pairs, and were intended to represent the result of physical interactions among pairs of
technologies - attempting to account for situations where 2x2^4.

   For this analysis, the CAFE Model has been modified to accommodate the results of the
large-scale vehicle  simulation study conducted by Argonne National Laboratory (and described
above). While Autonomie, Argonne's vehicle simulation model, produces absolute fuel
                                             13-28

-------
                                                     Analysis of Augural CAFE Standards
consumption values for each simulation record, the results have been modified in a way that
preserves much of the existing structure of the CAFE Model's compliance logic, but still
faithfully reproduces the totality of the simulation outcomes present in the database.
Fundamentally, the implementation represents a translation of the absolute values in the
simulation database into incremental improvements and a substantially expanded set of synergy
factors.

   Incremental Effectiveness or Absolute Improvement?

   As it always has, the CAFE Model applies a given technology, to a given vehicle and
estimates the incremental improvement in fuel consumption from the new combination of
technologies - with the ultimate goal of estimating a manufacturer's compliance position relative
to a set of fuel economy standards. However, unlike previous versions, the notion of incremental
has more nuance. As one sees from an examination of the Argonne database, each technology
applied results in a different level of fuel consumption depending upon the existing technology
content (and mass) of the vehicle to which it is applied. In the past, the incremental effectiveness
of a given technology was represented by a single point but, as the database illustrates, the true
incremental effectiveness of a given technology is a distribution across all of the technology
combinations to which it can be applied, rather than a single point.

   For example, as Figure 13.17 shows, it is possible to apply level 1 turbocharging to vehicles
of widely varying initial fuel economies, though the bulk of the observations in the database are
between 45 and 60 MPG. There are nearly 1,200 unique technology combinations to which level
1 turbocharging and downsizing (TURBO 1) can be applied. It seems reasonable to assume that
applying the same technology to vehicles with over a thousand different technology
combinations will yield different levels of improvement for at least some of these combinations.
As Figure 13.17 illustrates, that is indeed the case.  Applying TURBO 1 to a given vehicle
changes the fuel economy of that vehicle depending upon the set of technologies already present
when turbocharging is applied. Estimating the incremental improvement of adding level one
turbocharging to an otherwise identical vehicle (i.e., identical except for the presence  of other
fuel economy improving technologies) produces a distribution of fuel economy improvements,
rather than a single value, like the graph  in Figure 13.18.
                                             13-29

-------
                                                      Analysis of Augural CAFE Standards
                         40     45    50     55     60    65     70     75

                                        Fuel economy (MPG)
   Figure 13.17 Fuel Economy of Simulated Vehicles before (Red) and after (Blue) Application of Level 1
                                      Turbocharging
   Not only does Figure 13.18 illustrate that applying TURBO 1 produces some incremental fuel
consumption improvements close to zero percent (where the red represents vehicles without
TURBO 1, blue is vehicles with TURBO 1, and purple is the overlap in the distribution of fuel
economy between the two), but that it also results in some incremental improvements greater
than 15 percent depending upon the configuration to which it is applied. While only the
distribution of incremental effectiveness for level 1 turbocharging is shown here, the
distributions of incremental effectiveness for other technologies have similar levels of variation,
if not similar shapes.

   Despite the existence of absolute fuel consumption estimates from the Autonomie
simulations, there are advantages to continuing to apply technology based on incremental
effectiveness values - complicated, though it is, to incorporate the distribution of improvement
illustrated by Figure 13.18.

   The CAFE model was designed to consider, and apply, technologies based on the resulting
incremental improvement in fuel economy. Additionally, the analysis fleet (described in Chapter
4.2), represents a wide array of technology combinations and vehicle attributes - even within a
single class. For example, within the midsize car technology class (one of five technology classes
to which vehicle models in the analysis fleet are assigned), the analysis fleet starts with over 200
unique technology combinations to which the CAFE model adds technology. Attempting to
                                             13-30

-------
                                                      Analysis of Augural CAFE Standards
capture all of those technology combinations with a single effectiveness value for each
technology (and even a limited set of synergy factors) is bound to result in distortions as more
and more technology is applied within the CAFE model's compliance simulations.
                o
                o
                CO
             8-
             c  o
             o>  o
             13  T
             CT
             (D
                o
                o
                CM
                    0.00
020
                             0.05        0.10        0.15

                                  FE improvement with TURB01

Figure 13.18 Fuel Economy Improvement to Vehicles That Acquire Level 1 Turbocharging In Simulation
   But can the absolute fuel consumption values of the database be used in the CAFE model? In
the current implementation, they are - though not directly. There is a wide variety of engine
power and fuel consumption, even for a single technology combination, in the analysis fleet.
Using the absolute fuel consumption values in the Argonne database would require mapping
each vehicle to a point in the database,  and measuring the difference between its starting fuel
economy and that of the point in the database with identical technology content. Afterward, the
improvement in fuel consumption resulting from any additional technology added to that vehicle
can be based upon the difference of the points in the database and the initial fuel economy
difference resulting from the mapping.  While our approach appears  different computationally, it
produces identical results. However, in addition to circumventing some of the initial mapping, it
allows the CAFE model to consider technologies that were not simulated as part of the Argonne
project, and thus do not appear in the database. For example, reductions in a vehicle's accessory
load  produce small improvements in fuel economy, and are assumed to scale linearly with other
technologies.

   Additionally, the current approach required that we impose the structure of the decision tree,
which describes a sequence in which technologies should be considered for application, in  order
to define incremental effectiveness. While the combinations simulated by Argonne did capture
the exclusions represented by the decision tree (prohibiting variable valve lift on an engine that
                                             13-31

-------
                                                     Analysis of Augural CAFE Standards
does not also have variable valve timing, for example), there is no innate structure to inform a
sequential technology application process. For example, consider a vehicle in the analysis fleet
that starts with a 5-speed automatic transmission. Present in the database are two points, each
with an engine identical to the one in the analysis fleet under consideration, paired with a 6-
speed automatic transmission and an 8-speed automatic transmission. Without imposing the
decision tree structure on the incremental effectiveness values, the model would simply choose
the more effective of those two combinations to implement (assuming the cost-effectiveness of
the 8-speed is more attractive). While it might do this anyway, it is important that it consider the
6-speed first - doing so preserves the perspective of minimizing both the cost of compliance and
the extent to which more advanced technologies penetrate the new vehicle market.

   However, in order to translate the database of absolute fuel consumption values into some set
of incremental improvements for each technology, it is necessary to define a reference point -
the technology state (and fuel consumption) against which subsequent levels are measured to
determine the level of improvement (specified as a percentage improvement in fuel
consumption). Incremental effectiveness implies that the next technology provides  some
improvement in fuel economy over a previous technology state, holding everything else constant.
This requires that we define a "reference vehicle" against which to compare increasing levels of
technology.

   For any given technology, there are many logical reference points. There a number of vehicles
in the simulation database that are eligible to receive turbocharging and downsizing at the next
technology application. However, as Figure 13.19 shows, there is a wide variety of power, fuel
economy, and other technology content among them.
                                             13-32

-------
                                                      Analysis of Augural CAFE Standards
   61
   60
   59
   58
   57
   56
   55
   54
   53
  -50
 S49
 UJ48
 <47
 °46
   45
   44
   43
   42
   41
   40
   39
   38
   37
       94    98    102   106   110
114   118   122   126   130   134   138   142   146
 Engine Power (kW)
Transmission.Type
• AU
• CVT
• DDCT
• DM
              Figure 13.19 Midsize Vehicles in the Database Eligible to Receive TURBO1
   Any of those points, with their variety of existing technology content, could be a logical
reference point for the incremental improvement in fuel consumption that results from applying
level 1 turbocharging. While the engines of the vehicles in Figure 13.19 all have similar levels of
technology, there is a wide variety in other vehicle attributes: different transmissions (color
coded by type), different levels of electrification, mass reduction, aerodynamic and rolling
resistance improvements. While any of these points (of which there are over 2000) could serve as
the reference point for TURBO 1 improvement based on the interaction of the existing
technologies with TURBO 1, a better approach is to consider the technology tree holistically and
define a series of reference points that are intuitive, and internally consistent.
   Defining the reference point for incremental improvement
   The technologies have always been considered as part of a tree, where a vehicle moves from
one technology state to another in order of (generally) increasing complexity. While the engine
                                              13-33

-------
                                                        Analysis of Augural CAFE Standards
technologies are (almost) all related to one another, there is no inherent connection between the
engine technologies and technologies on other paths of the tree. For example, any of the
transmissions can be combined with any of the engine technologies - so those can safely be
considered separate paths. As Figure 13.20 shows, there are about 12 distinct paths that can be
traversed by a vehicle to which the model  applies technology. However, by combining logically
sequential technologies into common paths, we are left with 6 distinct paths (which may have
more than one branch where technologies  are considered to be mutually exclusive).
  Adv. Eng. Path
  ifechnoiogies	:
                                   Transmission Paths
                              [Manual Transmission Path	

  Electrification Path
[Technologies	
Hybrid/Electric Path
technologies
1 SHEVP2

SHEVPS
                                                       Adv. Hybrid/Electric Path
                                                       Technologies
     Figure 13.20 Technology Tree Used to Map Autonomie Simulations to Draft TAR Technology Set
   Electrification technologies represent an exception to this general construction. While the
stop/start technology is defined incrementally to the initial state across all paths, both of the
integrated starter generator variants (belt, BISG, and crank, CISG) are defined relative to the
12V stop/start. The full hybrids are also different - with the power split hybrid (SHEVPS)
defined relative to the crank-integrated starter generator (CISG), and the parallel hybrid
(SHEVP2) defined relative to the belt-integrated starter generator (BISG). The 30-mile-range
plug-in hybrid electric system is defined relative to the power split hybrid, and the subsequent
electrification technologies follow the path described in the decision tree.

   The "incremental effectiveness" values that appear in the model input files, and that are used
in the fuel consumption calculations when new technology is added to a vehicle, are all based on
incremental differences over a single reference point for each technology. However, progress
along some technology paths is treated as linear (forcing consideration of 6-speed automatic
transmission prior to considering application of CVT, for example), and along others as  strictly
sequential (mass reduction levels, for example, must logically be considered in order, since one
cannot reduce the mass of a vehicle by 10 percent without first reducing it 5 percent). Thus, the
reference point for each technology's incremental effectiveness estimate is the logical preceding
                                               13-34

-------
                                                       Analysis of Augural CAFE Standards
technology along its path, and the null state along all other paths7 - where the null state is
defined as a vehicle with (only) variable valve timing (VVT), a 5-speed automatic transmission
(ATS), no electrification, mass reduction, aerodynamic improvements, or low rolling resistance
tires. For example, the reference engine for each class has only VVT. When considering the
incremental impact of applying a 8-speed automatic transmission to a vehicle, the point of
reference is the logical preceding technology on the transmission path (in this case, the 6-speed
automatic transmission), and the base engine without any electrification, no mass reduction, and
no improvements in aerodynamics or rolling resistance.

   Translating the technology tree

   In order to incorporate the results of the Argonne database, while still preserving the basic
structure of the CAFE model's technology module, it was necessary to translate the points in the
database into locations on the technology tree8, shown in Figure 13.20. By recognizing that most
of the paths on the technology tree are unrelated, or separable, it is possible to decompose the
technology tree into a small number of paths  and branches by technology type. To achieve this
level of linearity, we define technology groups - only one of which is new. They are: engine cam
configuration (CONFIG), engine technologies (ENG), transmission technologies (TRANS),
electrification  (ELEC), mass reduction levels (MR), aerodynamic improvements (AERO), and
rolling resistance (ROLL). The combination of technology levels along each of these paths
define a unique technology combination that  corresponds to a single point in the database for
each technology class. These technology state definitions are more important for defining
synergies than for determining incremental effectiveness, but the paths are incorporated into
both.

   As an example, a technology combination with a SOHC engine, variable valve timing (only),
a 6-speed automatic transmission, a belt-integrated starter generator, mass reduction (level 1),
aerodynamic improvements (level 2), and rolling resistance (level 1) would be specified as
SOHC; VVT; AT6;BISG;MR1;AERO2;ROLL1. By assigning each technology state a vector such
as the one in the example, the CAFE model assigns each vehicle in the analysis fleet an initial
state that corresponds to a point in the database. Next, the  model determines  a percentage
improvement from the database for the new combination of technologies that is applied to each
vehicle model  and that percentage improvement is applied to the fuel consumption of that
vehicle model  in the analysis fleet.

   Once a vehicle is assigned a technology state (one of the tens of thousands of unique 7-tuples,
defined as CONFIG;ENG;TRANS;ELEC;MR;AERO;ROLL), adding a new technology to the
vehicle simply represents progress from one technology state to another. The vehicle's fuel
consumption is
' There are a few exceptions to this general rule, where the decision tree merges after a fork. For example, power
  split strong hybrid is incremental to both the belt-integrated and crank-integrated starter generator (BISG and
  CISC), but is defined incrementally to the CISC. Similarly, TURBO 1 is defined relative to cylinder deactivation
  (DEAC), even though it is incremental to both the high compression ratio engine (HCR) and DEAC. These
  instances are coded into the CAFE model, and accounted for in the technology effectiveness estimates and
  synergy factors.
! The technology tree was also modified to make some branches more sequential (or at least linear) and reduce the
  number of places where distinct branches converge.
                                              13-35

-------
                                                     Analysis of Augural CAFE Standards
   Where: FCiis the fuel consumption resulting from the application of technology /', FCo is the
vehicle's fuel consumption before technology /' is applied, FCL is the incremental fuel
consumption (percentage) improvement associated with technology /', Skis the synergy factor
associated with the combination, k, of technologies the vehicle technology / is applied, and So the
synergy factor associated with the technology state that produced fuel consumption FCo. The
synergy factor is defined in a way that captures the incremental improvement of moving between
points in the database, where each point is defined uniquely as a 7-tuple describing its cam
configuration, highest engine technology, transmission, electrification type, mass reduction level,
and level of aerodynamic or rolling resistance improvement.

   Throughout successive application of technologies, the simple product of the incremental
effectiveness associated with those technologies drifts away from the magnitude of the
improvements determined by Autonomie, and represented in the database, since the simple
product inadequately captures the interactions of those technologies. The synergy values correct
for this. In the past, synergy values in the  Volpe  model were represented as pairs. However, the
new values are 7-tuples and there is one for every point in the database. The synergy factors are
based (entirely) on values in the Argonne database, producing one for each unique technology
combination for each technology class, and are calculated as
                                 "«    FC0 ' 0(1 - *i)

   Where: Skis the synergy factor for technology combination k, FCo is the fuel consumption of
the reference vehicle (in the database), xiis the fuel consumption improvement of each
technology /' represented in technology combination k (where some technologies are present in
combination k, and some are precedent technologies that were applied, incrementally, before
reaching the current state on one of the paths).

   Future direction

   Integration of the database into the CAFE model resolves one of two important challenges -
the combined impact of applying many new technologies simultaneously.  Compared to past
reliance on pairwise synergy factors, simulating all combinations explicitly provides a basis to
more fully account for the overall impacts of combinations of multiple technologies. NHTSA
will continue to consider means to address a second challenge, which is not new to the current
approach, and that involves the application of simulation results for one vehicle to a much wider
set of vehicles. Like past analyses, today's analysis assumes that improvements scale uniformly
within a technology class. However, there are important differences  between the range of
vehicle power and mass in the MY2015 fleet compared to the range explicitly simulated by
ANL, and these differences could impact the magnitude of fuel economy improvements that can
be expected for the application of any particular technology combination.  Volpe Center staff are
exploring the potential to estimate a series of functions (given the current simulation database,
likely over  3500 functions) that would control for the unique combination  of technologies (e.g. a
vehicle with VVT,VVL,SGDI, ATS, SS12V, and AERO10+ROLL10) when estimating the
impact of vehicle mass and power on fuel economy. If successful, this effort could yield a set of
estimated functions and fitted coefficients that can be used to estimate absolute fuel consumption
                                             13-36

-------
                                                      Analysis of Augural CAFE Standards
associated with a given vehicle that has initial mass and engine power levels determined by the
observed values in the analysis fleet.

   NHTSA seeks comment on all of the above revisions to the model's approach to estimating
the extent to which the addition of various combinations of technologies to specific vehicles
could improve fuel economy, in particular on the approach to integrating the results of full
vehicle simulation. The agency seeks information that could be used to further refine this aspect
of the CAFE model and the supporting model inputs, as well as information that could be used to
develop and implement any alternative approaches.

13.2.2 Overview and Technology Application

   The CAFE model is the tool that NHTSA uses to simulate each manufacturer's decisions
about how to comply with a given set of standards. The model is designed to accommodate
standards with a variety of user-defined  specifications regarding the slope of the curve that
relates footprint to fuel economy by class, locations of the flat slope regions, and rates of
increase over time that can vary by year and regulatory class. While the properties of
technologies included  in the analysis are specified by the user (e.g. fuel consumption
improvement resulting from application, cost of the technology), the set of included technologies
is part of the model itself, which contains the information about the relationships between
technologies. In particular, the CAFE model contains the information about the sequence of
technologies, the paths on which they reside, any prerequisites associated with a technology's
application,  and any exclusions that naturally follow once it is applied.

This section summarizes the representation of fuel saving technology in the CAFE model. Table
13.3 and Table 13.4 contain all of the technology assumed to be available for manufacturers in
the Draft TAR analysis. The "application level" describes the system of the vehicle to which the
technology is applied,  which in turn determines the extent to which that decision affects other
vehicles in a manufacturer's fleet. For example, if a technology is applied at the "engine" level, it
naturally affects all other vehicles that share that same engine (though not until they themselves
are redesigned, if it happens to be in a future model year). The application schedule identifies
when manufacturers are assumed to be able to apply a given technology - with most available
only during vehicle redesigns. The application schedule also accounts for which technologies the
CAFE model tracks, but does not apply. These enter as part of the analysis fleet, and while they
are necessary for accounting related to cost and incremental fuel economy improvement, they do
not represent a choice  that manufacturers make in the model.
                                             13-37

-------
                                                       Analysis of Augural CAFE Standards
                            Table 13.3 CAFE Model Technologies (1)
Technology
SOHC
DOHC
OHV
TEFRI
LUBEFR1
LUBEFR2
LUBEFR3
VVT
VVL
SGDI
DEAC
HCR
HCRP
TURBO1
SEGR
DWSP
TURBO2
CEGR1
CEGR1P
CEGR2
HCR2
CNG
ADSL
TURBO DSL
DWSP DSL
EFRDSL
CLCDSL
LPEGRDSL
DSIZEDSL
Application
Level
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Engine
Application
Schedule
Baseline Only
Baseline Only
Baseline Only
Redesign Only
Refresh/Redesign
Redesign Only
Redesign Only
Refresh/Redesign
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Baseline Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Description
Single Overhead Camshaft Engine
Double Overhead Camshaft Engine
Overhead Valve Engine
Engine Friction Reduction Improvements (time-based)
Improved Low Friction Lubricants and
Engine Friction Reduction
LUBEFR2, Level 2
LUBEFR2, Levels
Variable Valve Timing
Variable Valve Lift
Stoichiometric Gasoline Direct Injection
Cylinder Deactivation
High Compression Ratio Engine
High Compression Ratio "Plus" Engine
Turbocharging and Downsizing, Level 1 (18 bar)
Stoichiometric Exhaust Gas Recirculation
Engine Downspeeding
Turbocharging and Downsizing, Level 2 (24 bar)
Cooled Exhaust Gas Recirculation, Level 1 (24 bar)
Cooled Exhaust Gas Recirculation, Level 1 "Plus" (24 bar)
Cooled Exhaust Gas Recirculation, Level 2 (27 bar)
Advanced High Compression Ratio Engine
Compressed Natural Gas Engine
Advanced Diesel
Improved Diesel Turbocharger
Diesel Engine Downspeeding with Increased Boost
Diesel Engine Friction Reduction
Closed Loop Combustion Control
Low Pressure Exhaust Gas Recirculation
Diesel Engine Downsizing
As discussed in Chapter 4.2, the analysis fleet contains the information about each vehicle
model, engine, and transmission selected for simulation and defines the initial technology state
of the fleet relative to the sets of technologies in Table 13.3 and Table 13.4.
                                              13-38

-------
                                                       Analysis of Augural CAFE Standards
                            Table 13.4 CAFE Model Technologies (2)
Technology
MT5
MT6
MT7
TATI
ATS
AT6
AT6P
ATS
AT8P
DCT6
DCT8
CVT
EPS
IACC1
IACC2
SS12V
BISG
CISC
SHEVP2
SHEVPS
PHEV30
PHEV50
BEV200
FCV
LDB
SAX
ROLL10
ROLL20
MR1
MR2
MRS
MR4
MRS
AERO 10
AERO20
Application
Level
Transmission
Transmission
Transmission
Transmission
Transmission
Transmission
Transmission
Transmission
Transmission
Transmission
Transmission
Transmission
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Platform
Platform
Platform
Platform
Platform
Platform
Platform
Application
Schedule
Baseline Only
Redesign Only
Redesign Only
Refresh/Redesign
Baseline Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Refresh/Redesign
Refresh/Redesign
Refresh/Redesign
Refresh/Redesign
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Refresh/Redesign
Refresh/Redesign
Refresh/Redesign
Refresh/Redesign
Refresh/Redesign
Redesign Only
Redesign Only
Redesign Only
Redesign Only
Refresh/Redesign
Redesign Only
Description
5-Speed Manual Transmission
6-Speed Manual Transmission
7-Speed Manual Transmission
Automatic Transmission Improvements (time-based)
5-Speed Automatic Transmission
6-Speed Automatic Transmission
6-Speed "Plus" Automatic Transmission
8-Speed Automatic Transmission
8-Speed "Plus" Automatic Transmission
6-Speed Dual Clutch Transmission
8-Speed Dual Clutch Transmission
Continuously Variable Transmission
Electric Power Steering
Improved Accessories - Level 1
Improved Accessories - Level 2
(w/ Alternator Regen and 70% Efficient Alternator)
12V Micro-Hybrid (Stop-Start)
Belt Mounted Integrated Starter/Generator
Crank Mounted Integrated Starter/Generator
P2 Strong Hybrid/Electric Vehicle
Power Split Strong Hybrid/Electric Vehicle
30-mile Plug-In Hybrid/Electric Vehicle
50-mile Plug-In Hybrid/Electric Vehicle
200-mile Electric Vehicle
Fuel Cell Vehicle
Low Drag Brakes
Secondary Axle Disconnect
Low Rolling Resistance Tires, Level 1 (10% Reduction)
Low Rolling Resistance Tires, Level 2 (20% Reduction)
Mass Reduction, Level 1 (5% Reduction in Glider Weight)
Mass Reduction, Level 2 (7.5% Reduction in Glider Weight)
Mass Reduction, Level 3 (10% Reduction in Glider Weight)
Mass Reduction, Level 4 (15% Reduction in Glider Weight)
Mass Reduction, Level 5 (20% Reduction in Glider Weight)
Aero Drag Reduction, Level 1 (10% Reduction)
Aero Drag Reduction, Level 2 (20% Reduction)
   Vehicle technologies provide a set of possible improvements available for the vehicle fleet
within the modeling system. The input assumptions for vehicle technologies, referred to below
simply as "technologies," are defined by the user in the technology input file for the model. As
part of the technology definition, the input file includes: additional cost associated with
application of the technology, an improvement factor (in terms of percent reduction of fuel
consumption), initial year that the technology may be considered for application, whether it is
applicable to a given class of vehicle, as well as other miscellaneous assumptions outlining
additional technology characteristics.
                                              13-39

-------
                                                       Analysis of Augural CAFE Standards
   The CAFE model defines several technology classes and pathways for logically grouping all
available technologies for application on a vehicle. Technology classes provide costs and
improvement factors shared by all vehicles with similar body styles, curb weights, footprints, and
engine types, while technology pathways establish a logical progression of technologies on a
vehicle.

   The modeling system defines two types of technology classes: the vehicle technology classes
and the engine technology classes. The system utilizes vehicle technology classes as a means for
specifying common technology input assumptions for vehicles that share similar characteristics.
Predominantly, these classes signify the degree of applicability of each of the available
technologies to a specific class of vehicles, as well as determine the base improvement factors
attributed to those technologies. Furthermore, for each technology, the vehicle technology
classes also  define the amount by which the vehicle's weight may decrease (resulting from
application of mass reducing technology), and the additional cost associated with application of
non-engine-level technologies. It is up to the user to assign each vehicle in the analysis fleet to
one of these technology classes.

   The model supports seven vehicle technology classes as shown in Table 13.5.
                             Table 13.5 Vehicle Technology Classes
Class
SmallCar
MedCar
SmallSUV
MedSUV
Pickup
Truck 2b/3
Van 2b/3
Description
Small passenger cars
Medium to large passenger cars
Small sport utility vehicles and station wagons
Medium to large sport utility vehicles, minivans, and passenger vans
Light duty pickups and other vehicles with ladder frame construction
Class 2b and class 3 pickups
Class 2b and class 3 cargo vans
Since the costs attributed to application of engine-level technologies vary based upon the engine
configuration (such as number of engine cylinders or banks), the model defines separate engine
classes for specifying input costs for these technologies. The modeling system provides sixteen
engine technology classes as shown in Table 13.6. Once each vehicle is assigned a technology
and engine class, the model uses these assignments to obtain the appropriate applicability, fuel
economy improvement, and cost for each technology as appropriate for an individual vehicle.
                                              13-40

-------
                                                      Analysis of Augural CAFE Standards
                             Table 13.6 Engine Technology Classes
Class
2C1B
3C1B
4C1B
4C2B
5C1B
6C1B
6ClB_ohv
6C2B
6C2B_ohv
8C2B
8C2B_ohv
10C2B
10C2B_ohv
12C2B
12C4B
16C4B
Description
SOHC/DOHC engine with 2 cylinders and 1 bank
SOHC/DOHC engine with 3 cylinders and 1 bank
SOHC/DOHC engine with 4 cylinders and 1 bank
SOHC/DOHC engine with 4 cylinders and 2 banks
SOHC/DOHC engine with 5 cylinders and 1 bank
SOHC/DOHC engine with 6 cylinders and 1 bank
OHV engine with 6 cylinders and 1 bank
SOHC/DOHC engine with 6 cylinders and 2 banks
OHV engine with 6 cylinders and 2 banks
SOHC/DOHC engine with 8 cylinders and 2 banks
OHV engine with 8 cylinders and 2 banks
SOHC/DOHC engine with 10 cylinders and 2 banks
OHV engine with 10 cylinders and 2 banks
SOHC/DOHC engine with 12 cylinders and 2 banks
SOHC/DOHC engine with 12 cylinders and 4 banks
SOHC/DOHC engine with 16 cylinders and 4 banks
   The modeling system defines technology pathways for grouping and establishing a logical
progression of technologies on a vehicle. Each pathway (or, path) is evaluated independently and
in parallel, with technologies on these paths being considered in sequential order. As the model
traverses each path, the costs and improvement factors are accumulated on an incremental basis
with relation to the preceding technology. The system stops examining a given path once a
combination of one or more technologies results in a "best" technology solution for that path.
After evaluating all paths, the model selects a most cost-effective solution among all pathways.
This "parallel path" approach allows the modeling system to progress thorough technologies in
any given pathway without being unnecessarily prevented from considering technologies in other
paths.

   Rather than rely on a specific set of technology combinations or packages, the model
considers the universe of applicable technologies, dynamically identifying the most cost-
effective combination of technologies for each manufacturer's vehicle fleet based on the
assumptions about each technology's effectiveness, cost,  and interaction with all other
technologies both present and available.

The modeling system incorporates thirteen technology pathways  for evaluation as shown in
Table 13.7. Similar to individual technologies, each path carries an intrinsic application level that
denotes the scope of applicability of all technologies present within that path, and whether the
pathway is evaluated on one vehicle at a time, or on a collection of vehicles that share the same
platform, engine, or transmission.
                                              13-41

-------
                                                      Analysis of Augural CAFE Standards
                              Table 13.7 Technology Pathways
Technology Pathway
Basic Engine Path
Turbo Engine Path
Advanced Engine Path
Diesel Engine Path
Manual Transmission Path
Automatic Transmission Path
Electrification Path
Hybrid/Electric Path
Advanced Hybrid/Electric Path
Dynamic Load Reduction Path
Low Rolling Resistance Tires Path
Mass Reduction Path
Aerodynamic Improvements Path
Application Level
Engine
Engine
Engine
Engine
Transmission
Transmission
Vehicle
Vehicle
Vehicle
Vehicle
Vehicle
Platform
Platform
   The technologies that comprise the four Engine-Level paths available within the model are
presented in Figure 13.21 below. Note that the baseline-level technologies (SOHC, DOHC,
OHV, and CNG) are grayed out. As mentioned earlier, these technologies are used to inform the
modeling system of the input engine's configuration, and are not otherwise applicable during the
analysis.
                                             13-42

-------
                                                      Analysis of Augural CAFE Standards
Basic Engine Path
SOHC
\i












DEAC




DOHC
I
LUBEFR1
I
LUBEFR2
I
LUBEFR3
1
WT
1
WL
1
SGDI
L/\





OHV
i/ !












HCR
1 !
HCRP

                                        JTurbo Eng. Path i  jAdv. Eng. Path
| Diesel Eng. Path
                               Figure 13.21 Engine-Level Paths
   For all pathways, the technologies are evaluated and applied to a vehicle in sequential order,
as shown, from top to bottom. In some cases, however, if a technology is deemed ineffective, the
system will bypass it and skip ahead to the next technology. If the modeling system applies a
technology that resides later in the pathway, it will "backfill" anything that was previously
skipped in order to fully account for costs and improvement factors, each of which are specified
on an incremental basis. For any technology that is already present on a vehicle (either from the
input fleet or previously applied by the model), the system skips over those  technologies as well
and proceeds to the next. These skipped technologies, however, will not be  applied again during
backfill.

   The Basic Engine path begins with SOHC, DOHC, and OHV technologies defining the initial
configuration of the vehicle's engine. Since these technologies are not available during
modeling, the system evaluates this pathway starting with LUBEFR1 technology. Toward the
end of the path, the model encounters a choice between DEAC and HCR technologies.
Whenever a technology pathway forks into two or more branch points, all of the branches are
treated as mutually exclusive. The system evaluates all technologies forming the branch
simultaneously, and selects the most cost-effective for the application, while disabling the
remaining paths not chosen. In the case of the Basic Engine path, that means if a vehicle
                                             13-43

-------
                                                     Analysis of Augural CAFE Standards
continues with application of the DEAC technology, the HCR and HCRP technologies will be
disabled. Likewise, if the vehicle applies the HCR technology, the HCRP technology will still be
available for evaluation, while the DEAC technology will be disabled.

   The technologies exposed by the Advanced Engine path (HCR2 and CNG) are not
incremental over each other and do not follow a traditional progression logic present on other
paths. Consequently, these technologies are treated as mutually exclusive within the model.
Since CNG is a baseline-level technology, the only remaining choice for application within the
Advanced Engine path is HCR2.

   The technologies that make up the two Transmission-Level paths defined by the modeling
system are shown in Figure  13.22 below. The baseline-level technologies (MT5 and ATS) are
grayed and are only used to represent the initial configuration of the vehicle's transmission. For
simplicity, all manual transmissions with five forward gears or fewer have been assigned the
MT5 technology in the analysis fleet. Similarly, all automatic transmissions with  five forward
gears or fewer have been assigned the ATS technology.
M;
anual Trn. Path
MT5
1
MT6
1
MT7


                            Figure 13.22  Transmission-Level Paths

   Given the definition of incremental costs and fuel consumption improvement factors utilized
during the analysis, the system assumes that all manual transmissions with seven or more gears
are mapped to the MT7 technology. Moreover, the ATS technology should map to all automatic
transmissions with seven or more forward gears, DCT6 technology  should map to all dual-clutch
(DCT) or auto-manual (AMT) transmissions with five or six forward gears, and DCT8
technology should map to all DCT's or AMT's with seven or more forward gears. These
transmission technology utilization assignments, however, are defined within the analysis fleet,
and are not strictly enforced by the modeling system.

   As mentioned earlier, the branch points  shown in the Automatic Transmission path are
mutually exclusive. For example, if a vehicle transitions to the DCT branch, the CVT and all
automatic transmission technologies will become unavailable.
                                             13-44

-------
                                                      Analysis of Augural CAFE Standards
   The technologies that compose the two Platform-Level paths provided by the model are
displayed in Figure 13.23 below, and consist of mass reduction and aerodynamic improvements.
Ml
iPath
MR1
1
MR2
1
MR3
1
MR4
1
MRS


                                             JAEROPath
                              Figure 13.23 Platform-Level Paths
   The technologies that constitute the two Vehicle-Level paths defined by the system are
outlined in Figure 13.24 below.
Electrification Path


EPS
I
IACC1
1
IACC2
1
SS12V


L/ \l
BISG


CISG


Hybrid/Electric Path
SHEVP2 SH

EVPS


Advanced Hybrid/ EJectric Path
PHEV30
1
PHEV50
L/ \i
BEV200 F


cv

                              Figure 13.24 Vehicle-Level Paths
   The technologies on the Hybrid/Electric path (SFEVP2 and SFtEVPS) are defined as stand-
alone and mutually exclusive. These technologies are not incremental over each other and do not
follow a traditional progression logic present on other paths.
                                             13-45

-------
                                                      Analysis of Augural CAFE Standards
   Even though the model evaluates each technology path independently, some of the pathways
are interconnected to allow for additional logical progression and incremental accounting of
technologies. For example, the SHEWS (power-split strong hybrid/electric) technology on the
Hybrid/Electric path is defined as incremental over the DEAC (cylinder deactivation) technology
on the Basic Engine path, the ATS (5-speed automatic) technology on the Automatic
Transmission path, and the CISG (crank mounted integrated starter/generator) technology on the
Electrification path. For that reason, whenever the system evaluates the SHEVPS technology for
application on a vehicle, it ensures that, at a minimum, all the aforementioned technologies (as
well as their predecessors) have already been applied on that vehicle. However, if it becomes
necessary for a vehicle to progress to the power-split hybrid, the model will virtually apply the
technologies associated with the reference point in order to evaluate the attractiveness of
transitioning to the strong hybrid.

   Of the thirteen technology pathways present in the model, all Engine paths, the Automatic
Transmission path, the Electrification path, and both Hybrid/Electric paths are logically linked
for incremental technology progression. This relationship between pathways is illustrated in
Figure  13.25 below.

   Some of the technology pathways, as defined in the CAFE model  and shown in the diagram
below,  may not be compatible with a vehicle given its state at the time of evaluation. For
example, a vehicle with a 6-speed automatic transmission will not be able to get improvements
from a  Manual Transmission path. For this reason, the system implements logic to explicitly
disable certain paths whenever a constraining technology from another path is applied on a
vehicle. On occasion, not all of the technologies present within a pathway may produce
compatibility constraints with another path. In such a case, the system will selectively disable a
conflicting pathway (or part of the pathway) as required by the incompatible technology.
                                             13-46

-------
                                                        Analysis of Augural CAFE Standards
          Basic Engine
             Path
         Turbo Engine
             Path
  Automatic
 Transmission
     Path
   Manual
Transmission
    Path
 Electrification
     Path
                                          Hybrid/Electric
                                              Path
       Advanced Engine
             Path
                                            Advanced
                                          Hybrid/Electric
                                              Path
         Diesel Engine
             Path
        Mass Reduction
           (MR) Path
 Aerodynamic
Improvements
 (AERO) Path
Dynamic Load
  Reduction
  (DLR) Path
  Low Rolling
Resistance Tires
  (ROLL) Path
                           Figure 13.25 Technology Pathways Diagram
   For any interlinked technology pathways shown in Figure 13.25 above, the system also
disables all preceding technology paths whenever a vehicle transitions to a succeeding pathway.
For example, if the model applies SFLEVPS technology on a vehicle, the system disables the
Turbo, Advanced, and Diesel Engine paths (as defined above), as well as the Basic Engine, the
Automatic Transmission, and the Electrification paths (all of which precede the Hybrid/Electric
path)9. This implicitly forces vehicles to always move in the direction of increasing technological
sophistication each time they are reevaluated by the model.
9 The only notable exception to this rule occurs whenever SHEVP2 technology is applied on a vehicle. This
  technology may be present in conjunction with any engine-level technology, and as such, the Basic Engine path is
  not disabled upon application of SHEVP2 technology, even though this pathway precedes the Hybrid/Electric
  path.
                                               13-47

-------
                                                      Analysis of Augural CAFE Standards
13.2.3 Simulating Manufacturer Compliance with Standards

   In the U.S. market, the stringency of CAFE standards can influence the design of new
vehicles offered for sale by requiring manufacturers to produce increasingly fuel efficient
vehicles in order to meet program requirements. This is also true in the CAFE model simulation,
where the standards can be defined with a great deal of flexibility to examine the impact of
different program specifications on the auto industry. Standards are defined for each model year,
and can represent different slopes that relate fuel economy to footprint (or work factor, in the
case of medium-duty pickup trucks and vans), different regions of flat slopes, and different rates
of increase for each of three regulatory classes covered by the CAFE program (passenger cars,
light trucks, and medium-duty pickup trucks and vans).

   As a starting point, the model needs enough information to represent each manufacturer
covered by the program. The MY2015  analysis fleet contains information about each
manufacturer's:

      •   Vehicle models offered for sale - their current (i.e., MY2015) and future production
          volumes, prices, fuel saving  technology content (relative to the set of technologies
          described in Table 13.3 and Table 13.4 and other attributes (curb weight, drive type,
          assignment to technology class and regulatory class),
      •   Production constraints - product cadence of vehicle models (i.e., schedule of model
          redesigns and "freshening"), vehicle platform membership, degree of engine and/or
          transmission  sharing (for each model variant) with other vehicles in the fleet,
      •   Compliance constraints and flexibilities - historical preference for full compliance or
          fine payment, willingness to apply  additional cost-effective fuel saving technology in
          excess of CAFE requirements, projected applicable flexible fuel credits, and current
          CAFE credit  balance in first model year of simulation.


   Each manufacturer's CAFE requirement represents  the harmonic average of their vehicle's
sales-weighted targets. This means that no individual vehicle has a "standard," merely a target,
and each manufacturer is free to identify a compliance strategy that makes the most sense given
its unique combination  of vehicle models, consumers, and competitive position in the various
market segments. As  the CAFE model  provides flexibility when defining a set of CAFE
standards, each manufacturer's requirement is dynamically defined based on the specification of
the standards for any  simulation.

   In order to simulate a manufacturer's actions to bring its fleet into compliance with the
standards, the CAFE  model needs information about the context in which those decisions occur.
In particular, the model requires:

      •   The universe of technologies that can be used to achieve compliance, as well as
          information about the logical progression among them, and any restrictions that occur
          when applying one, or more, or them (see Section 13.2.2),
      •   The cost of each technology and its fuel economy improvement, relative to a wide
          array of starting points that span not only the set of observed technology combinations
          in the MY2015 fleet, but also the set that will exist as the fleet evolves to achieve
          compliance with CAFE standards,
                                             13-48

-------
                                                      Analysis of Augural CAFE Standards
      •   The fuel prices that consumers will face when purchasing new vehicles, and the
          number of miles they expect to travel in those vehicles.


   Given this information, the model estimates each manufacturer's potential year-by-year
application of fuel-saving technologies to each engine, transmission, and vehicle.  Subject to a
range of engineering and planning-related constraints (e.g., secondary axle disconnects can't be
applied to 2-wheel  drive vehicles, many major technologies can only be applied practicably as
part of a vehicle redesign, and applied technologies carry forward between model years), the
model attempts to apply technology to each manufacturers' fleet in a manner that minimizes
"effective costs."

   The effective cost captures more than the incremental cost of a given technology - it
represents the difference between their incremental cost and the value of fuel savings to a
potential buyer over the first three years of ownership. This construction allows the model to
choose technologies that both improve  a manufacturer's  CAFE compliance position and are most
likely to be attractive to its consumers.  This also means that different assumptions about future
fuel prices will produce different rankings of technologies when the model evaluates available
technologies for application. For example, in a high fuel  price regime, an expensive but very
efficient technology may look attractive to manufacturers because the  value of the fuel savings is
sufficiently high to both counteract the higher cost of the technology and, implicitly, satisfy
consumer demand to balance price increases with reductions in operating cost. The model
continues to add technology until a manufacturer either:  (a) reaches compliance with CAFE
standards (possibly through the accumulation and application of CAFE credits), (b) reaches a
point at which it is  more cost effective to pay fines than to add more technology, or (c) reaches a
point beyond  compliance where the manufacturer assumes its consumers will be unwilling to pay
for additional fuel saving technologies (specified as a desired "payback period," assumed to be
one year for all manufacturers in this analysis.

   A graphical depiction of the compliance  simulation loop appears in Figure  13.26, below.
Having determined the applicability of each technology to each vehicle model, platform, engine,
and transmission, the compliance simulation algorithm begins the process of applying
technologies based on the CAFE standards applicable during the current model year. This
involves repeatedly evaluating the degree of noncompliance, identifying the next "best"
technology (ranked by the effective cost discussed above) available on each of the parallel
technology paths described in Chapter 5, and applying the best of these. The algorithm combines
some of the pathways, evaluating them sequentially instead of in parallel, in order to ensure
appropriate incremental progression of technologies.

   The algorithm first finds the best next applicable technology in each of the technology
pathways, then selects the best among these. If a manufacturer is assumed to be unwilling to pay
CAFE civil penalties, then the algorithm applies the technology to the affected vehicles.
Afterwards, the algorithm reevaluates the manufacturer's degree of noncompliance and
continues application of technology. Once a manufacturer reaches compliance (i.e., the
manufacturer would no longer need to pay CAFE civil penalties), the algorithm proceeds to
apply any additional technology determined to be cost-effective (as discussed above).
Conversely, if a manufacturer is assumed to prefer to pay CAFE civil  penalties, the algorithm
only applies technology up to the point where doing so is less costly than paying fines. The
                                             13-49

-------
                                                            Analysis of Augural CAFE Standards
algorithm stops applying additional technology to this manufacturer's products once no more
cost-effective solutions are encountered. This process is repeated for each manufacturer present
in the input fleet. It is then repeated again for each modeling year. Once all modeling years have
been processed, the compliance simulation algorithm concludes.
             Begin *
               Evaluate Manufacturer's
                 bevel ef Compliance
           Evaluate All Technolofy Pathway*
             Find (Se&i N&xi Techno logy in:
             > feik. Engine- Pvnh,
             > Turbtj Engine' P-iih,
             > Adv.j nasti Engine f?«uh
             Find fk'jst Nexc Jkithrmb^y in:
             > Ok'M'l tnginfe IMih
             I ind Ik'st Nex£ Itii-hnolft&'y in.
             >Miinu;il Irimsmi.wtm Palh
             f ind Ifr-st NttJtt Iprhnulugy in:
             >Adv;inri?ri Hyhrid/tlnrtlie Path
             Find fk'S* N*fK5 T^chncilcjgy i
             > [>yn.gimir; loi-id H^ducUon P
             Find Best N^?sf TechnDto^y in:
             >• low Rolling Rpsiit, ^iri?1^ Rgth
             Find Bo'it N*?K? T^chno'kJ
             > Mass ReducHon Path
             Find B$
                       ^nts P-3th
Sclwt Mo*t Co*t-Effective
 Technology S^liitlon
  from AM Pathways
                                               T
                                             Is T«ch Solution
                                             Manufacturer
                                              Has Fines?
                             Manufacturer
                             Prcferi to Pay
                               Finos?

                                IVes
                                                                        Pay Fines
                                                                            \
                                                   Repeat for Next Manufacturer
                            Figure 13.26 Compliance Simulation Diagram
   Engine, transmission, and platform sharing represent constraints to a manufacturer as it
attempts to modify its product lines in ways that achieve CAFE requirements. The combination
of shared components and product cadence can create challenges for manufacturers in any given
year, and strongly influence both the pace and extent of new fuel saving technology application.
For example, Ford produces approximately  1,000 different model variants across the passenger
car, light truck, and medium-duty pickup/van regulatory classes (though more than 800 of these
are differently configured medium-duty pickup trucks and vans). However, all of these models
are powered by only about 25 different engines. Even ignoring all of the class 2b3 trucks, the
ratio of model variants to unique engines is  about 10:1. So when Ford changes an engine on one
                                                  13-50

-------
                                                     Analysis of Augural CAFE Standards
of its vehicles to improve its fleet fuel economy for CAFE, the changes to that engine appear on
an average of 10 other vehicles as well. Multi-year planning horizons in the CAFE model
account for this nuance, and represent the fact that building a fleet of vehicles for compliance is
different than modifying a single vehicle to exceed its fuel consumption target. Underlying the
compliance simulation loop in Figure 13.26 is the selection of the "next best" technology within
each path. In the new version of the CAFE model, "next best" incorporates both the product
cadence and component sharing discussed in this chapter.
                                             13-51

-------
                                                                           Analysis of Augural CAFE Standards
             Evaluate Potential Applications
              of Pending Technology on all
              Vehicles for a Manufacturer
                 Evaluate Next Vehicle
                  Can Pending Tech
                   be Applied on a
                     Vehicle?
Backfill any Omitted Technology
>
r
              Calculate Effective Cost
                    Manufacturer
                     Has More
                     Vehicles?
                Select Best Solution for
                 Pending Technology
              [i.e., yielding lowest eff-cost)
                   Is Tech Solution
                   Cost Effective?
                   (i.e., eff-cost <0)
         Yes
  Pathway
  Has More
Technologies?
                         No
                                                  Begin
                                                    Evaluate Next Technology
                                                     on Technology Pathway
                          Yes
        Is Pending
       Technology
       Vehicle-Level?
                                                        No
                                               Calculate Effective Cost
                               Calculate Impacts of Pending
                               Technologies on Each Affected
                               Vehicle:
                               > New FE,
                               > Change in Weight,
                               > Total Tech Cost,
                               > Change in Value of Tech
                               > Value of Fuel Saved
                               Estimate New Compliance Values
                               for Each Reg-Class
                               for a Manufacturer:
                               > Standard (required CAFE level)
                               > CAFE (achieved CAFE level)
                               > Credits
                               > Fines
                                                  Calculate Reduction in
                                                  Total Fines
                                                  Calculate Effective Cost for
                                                  Vehicles/Technologies
                                                  Combination
   Evaluate Other         \
Technology Pathways      /
                                      1
                                                 Select Best Solution from All
                                                  Technologies in Pathway
Evaluate Potential Applications
 of Pending Technology on all
Components for a Manufacturer
                                                                                         Evaluate Next Component
                                                                                      Find Component's Leader Vehicle
                                         " Can Pending Tech^-,
                                           be Applied on a      ^j>
                                         ^ Leader Vehicle? ^^
                                                                                       Find All Follower Vehicles Where
                                                                                        Pending Tech Can be Applied
                                                                                       Backfill any Omitted Technology
                                                                                       on Leader and Follower Vehicles
                                                                                          Select Best Solution for
                                                                                           Pending Technology
                                                                                        (i.e., yielding lowest eff-cost)
                                                                                             Is Tech Solution
                                                                                             Cost Effective?
                                                                                             (i.e., eff-cost <0}
                                                                                                   Jjo
                                                                                                   ,
                                                                                               Pathway
                                                                                               Has More
                                                                                             Technologies?
                                                                                            Yes
                                                                               [No
            Figure 13.27  Selection of "Next Best" Technology within CAFE Compliance Simulation
    Figure 13.27 illustrates the logic employed by the model when choosing the "next best"
technology when simulating compliance for a manufacturer. Note, in the diagram above, a
"component" is  any platform, engine, or transmission produced by a manufacturer, where
                                                               13-52

-------
                                                     Analysis of Augural CAFE Standards
application of a technology is evaluated on a vehicle designated as a leader of that component.
The model chooses a "leader" for each shared component - engine, transmission, platform -
based on existing technology level, sales volume, and MSRP. New technologies (e.g., upgrades
to engines) are first made on the component leader when it is redesigned. Because other vehicle
models that share the same engine as the leader are likely to be redesigned at different model
years, they will inherit the new component (e.g., a new engine) only when they are redesigned.
However, the leader drives technology application. When a "follower" (who shares that
component, but is not designated as the leader) is redesigned, it may not change the shared
component in ways that differ from the implementation that exists on the leader in that year. The
model accounts for this sharing among component explicitly. When  selecting technologies to add
to a component leader, any follower vehicles of the same component that are redesigned at the
same time as the leader, will also be evaluated during technology application. Conversely, since
vehicle-level technologies affect only one vehicle at a time, all technology improvements are
applied immediately to just the one vehicle model during its refresh or redesign year.

  When the model steps forward to a new model year, all vehicles that are scheduled to be
redesigned in that year inherit the most current level of any shared components. For example, if
vehicle A and vehicle B share an engine, where vehicle A is the leader, vehicle B will inherit the
same engine that vehicle A has when it is redesigned before considering additional technology
application. It is possible that a vehicle model can be the leader on one component  and a follower
on another. This means that when that vehicle is redesigned, it first inherits the current state of
all technology components on which it is a follower, before making any improvements to
components on which it is the leader. These restrictions help to preserve the size of the initial set
of engines, transmissions, and platforms that are observable in the MY2015 fleet. The approach
does not generate unique engines for each variant, based  on NHTSA's analysis of observed
trends for managing platform and powertrain complexity given resource and cost considerations.

      As shown in the figures  above, the CAFE model considers each technology path
separately within each analysis step - virtually applying each of the best technologies in each
discrete path and choosing among them. Because this is an iterative process, for any vehicle in
any single model year, the CAFE model dynamically constructs a package of technologies to
improve its fuel economy, rather than choosing a package from a pre-defined set. The integration
of the Argonne simulation study means that for each technology class, the full vehicle simulation
results for over 20 thousand unique technology combinations are available to the CAFE model in
this evaluation. Many of these combinations will not be cost-effective for a given vehicle's
starting technology state each time it is evaluated, but considering them allows the  model to
avoid applying new technology in manner that ignores the existing technology preferences
specific manufacturers have exhibited in the MY2015 fleet.

  The CAFE model also simulates compliance on a yearly basis, over the entire period -
making choices in any given year with an eye toward compliance in future years. While the
compliance simulation loop is accurately described in Figure 13.26, the first step in the  process,
"Evaluate Manufacturer's Level of Compliance," is more nuanced that the figure suggests. The
first step  in the evaluation is the application of expiring credits - any CAFE credits carried
forward from earlier model years that will expire in the model year under consideration are
applied to the manufacturer's CAFE level. Then all of the models redesigned in that year inherit
the most  current versions of shared engines, transmissions, and platforms if they are eligible to
do so. The CAFE model also considers the application of older, but not yet expiring, credits if
                                             13-53

-------
                                                     Analysis of Augural CAFE Standards
the opportunities for technology application in a given model year are limited. In this way, a
manufacturer need not apply enough technology in any given model year to exactly comply with
the CAFE standard for that year. Instead, the CAFE model allows manufacturers to apply
technology more aggressively when opportunities exist in order to generate enough credits to
comply with standards when opportunities for technology application are more limited. The
CAFE model represents all of the credits that manufacturers currently hold (and their expiration
dates) when the simulation begins. The fact that the existing credit balances are significant (and
expected to be necessary for some manufacturers to comply with standards in the short-term)
suggests that capturing this behavior in the CAFE model is important.

   The following example demonstrates how manufacturer choices with respect to product
cadence may lead to blocky improvements in fleet fuel economy, generating credits in some
years that can then be applied in future years. Figure 13.28 shows the compliance pathway
simulated for FCA under both the final CAFE standards through MY2021 and the Augural
Standards through MY2025 (and assumed to remain constant after MY2025). Figure 13.3
showed the product cadence assumed for each manufacturer in this analysis.  As that figure
shows, FCA has a number of model years where relatively little of their total sales volume is
expected to be redesigned and several years where 20 percent or more of their total volume is
expected to be redesigned. As Figure 13.28 shows, the years with the highest increases in CAFE,
MY2018 and MY2020, correspond to years with high degrees of redesigns. However, the figure
also shows that FCA is simulated to exceed the standard in MY2018  for light trucks and
MY2020 for passenger cars by a large amount. Due to limited credit trading between passenger
car and light trucks fleets, it would be necessary for FCA to increase  both fleets in order to avoid
paying fines  (rather than simply relying on the over compliance in one or the other overcome
shortfalls). While FCA exceeds the standard for a number of years, generating credits which it
then  carries forward, it also falls short of compliance around MYs 2022 - 2024, when it applies
the earned credits from previous years to account for the shortfall.

   As discussed above, these results provide an estimate, based on analysis inputs, of one way
FCA could add fuel-saving technologies to its products, and are not a prediction of what FCA
will do under these standards.
                                            13-54

-------
                                                          Analysis of Augural CAFE Standards
        56

        54


        52


        50


        48


        46

        44


        42


        40


        38

      5"
      | 36

      ¥
      | 34


      K 32

        30


        28


        26


        24


        22


        20


        18


        16


        14

          2015  2016  2017 2018 2019 2020  2021  2022 2023 2024 2025  2026  2027 2028 2029 2030  2031  2032

      Regulatory Class, Standard, CAFE
      • LightTruck, CAFE
        LightTruck, Standard
      • Light!ruck2b3l CAFE
        LightTruck2b3, Standard
      • PassengerCar, CAFE
        PassengerCar, Standard
                              Figure 13.28 FCA Compliance Example
56

54


52


50


48


46


44


42


40


38


36 £
  S_

34 S


32

30


28


26


24


22


20


18


16


14
o
13.2.4 Simulating the Economic and Environmental Effects of CAFE Standards

   In addition to simulating compliance with CAFE standards, the CAFE model also estimates
the economic and environmental impacts associated with the changes to the vehicle fleet that are
estimated to occur as a result of the standards. To this, the model requires information about the
economic and environmental impacts of fuel consumption and travel. In particular, it requires
information about:
                                                 13-55

-------
                                                     Analysis of Augural CAFE Standards
          •  All of the information related to travel demand and energy prices that influence
             compliance simulation also influence effects calculations,
          •  The fuel economy rebound effect (the extent to which reductions in operating cost
             increase demand for miles traveled),
          •  The value of refueling time saved by consumers who have to refuel more efficient
             vehicles less often,
          •  Functions that determine the safety impacts of increased travel and vehicle mass
             reduction,
          •  The social costs of increases in the amount of congestion and noise (from
             additional travel demand) and the number of crashes and fatalities,
          •  The social cost of dependence on oil and the social cost of carbon emissions,
          •  Tailpipe and upstream emission factors, fuel density and carbon content
             associated with a variety of fuels.


   Having estimated the extent to which each manufacturer might add fuel-saving technologies
under each specified regulatory alternative, the model  calculates a range of physical impacts,
such as changes in highway travel (i.e., VMT), changes in fleetwide fuel consumption, changes
in highway fatalities, and changes in vehicular and upstream greenhouse gas and criteria
pollutant emissions.  The model then uses the information supplied about economic and
environmental values to calculate economic costs and  benefits to vehicle owners and society,
based on these physical impacts. The CAFE model calculates these changes and economic
impacts for each scenario, producing differences relative to a no-action case.  The values assigned
to all of the required environmental and economic inputs can be downloaded from NHTSA's
website.

13.3  Simulation Results for Augural MY2022 - 2025 Standards

   In the results that follow, NHTSA considered the impact of implementing the Augural
Standards described  in the 2012 Final Rule for MYs 2022 - 2025 relative to the current final
standards through MY2021 as the reference point. NHTSA uses the CAFE model to evolve the
analysis fleet in order reach the point where the Augural Standards begin in MY2022. It does this
by simulating manufacturers' compliance decisions in response to the standards, discussed in
greater detail below.

   EPCA/EISA constrains how NHTSA conducts its analysis in order to inform the actual
determination of the maximum feasible stringency of CAFE standards. For example, the statute
requires NHTSA to set aside EPCA/EISA's CAFE credit carry-forward provisions from such
analysis. In  recent CAFE rulemakings, NHTSA has included both a "standard setting" analysis
and a "real world" analysis, with the latter accounting  for some of these factors, as practicable.
This draft TAR is not a rulemaking document to inform actual decisions regarding the maximum
feasible stringency of future CAFE standards; therefore, today's analysis is all conducted on a
"real world"  basis. The analysis accounts for the potential that manufacturers, as allowed by
EPC A/EISA, could transfer CAFE credits between the passenger car and light truck fleet, or
carry CAFE  credits forward for later use. Except for CAFE credits earned prior to MY2015,
today's analysis does not account for the potential that manufacturers could trade CAFE credits.
                                             13-56

-------
                                                      Analysis of Augural CAFE Standards
Today's analysis also does not attempt to simulate the potential that manufacturers could carry
CAFE credits back from future model years.

   Like both recent "standard setting" and "real world" analyses, today's analysis also accounts
for the potential that some manufacturers might, as allowed by EPCA/EISA, elect to pay civil
penalties if doing so would likely be less expensive than applying additional fuel-saving
technology (accounting for technology costs and avoided fuel expenditures).  Recent legislation
requires the civil penalty rate be increased from the current level of $5.50 per 0.1 mpg per
vehicle to a considerably higher level of $14 per 0.1 mpg per vehicle, and today's analysis uses
the updated rate.10

   As discussed in Chapter 4, today's analysis includes PHEVs and EVs estimated to be
produced after MY2015.  Today's analysis also allows that manufacturers may elect to produce
additional PHEVs or EVs in response to new CAFE standards; however, as shown below,
compared to other technologies, PHEVs and EVs are not estimated to be cost-effective responses
to the augural CAFE standards (i.e., the CAFE model identifies more cost-effective  solutions
than building additional PHEVs or EVs). Had it included more PHEVs or EVs either in the
analysis fleet or as a forced additional application of technology, today's analysis would have
shown lower application rates for some other technologies (e.g., full HEVs) in the results shown
below.

   Some of the aspects of today's analysis, such as the change in the civil penalty rate, are
considerably different from those in NHTSA's 2012 analysis supporting the final rule for MYs
2017-2021. Together with other improvements and updates to data and methods, these combine
to produce updated results from those presented in 2012.  Especially with a view toward
understanding incremental impacts, today's analysis evaluates the potential response to the
existing standards in place through MY2021, referred to here as the "No Action Alternative."
Defining the No Action Alternative aids understanding of changes in inputs and methods, and
provides a proper point of reference for understanding the estimated impacts of the Augural
Standards. NHTSA is not considering changes to the already-final CAFE standards through
MY2021.

13.3.1 Industry Impacts

The footprint-based CAFE standards finalized in 2012 will require manufacturers to improve the
average fuel economy of their fleets between now and MY2021. In the baseline case, the
standards are  assumed to remain constant at the MY 2021 level indefinitely. The analysis in this
report compares this baseline case with the augural CAFE standards for MY2022 - MY2025.
10 As a result of the Federal Civil Penalties Inflation Adjustment and Improvement Act of 2015 (Pub. L. 114-74),
  Section 701, and OMB guidance from February 2016 on how agencies should implement that Act, NHTSA is
  required to increase the $5-per-tenth-of-an-mpg civil penalty. NHTSA will publish our proposal to implement
  that increase in a forthcoming Federal Register notice; for purposes of the current analysis, we have used $14-per-
  tenth-of-an-mpg, which is consistent with the OMB guidance.


                                             13-57

-------
                                                    Analysis of Augural CAFE Standards
Table 13.8, below, summarizes the actual CAFE requirement for each manufacturer in MY2015
(based on the MY2015 analysis fleet, described in Chapter 4.2); the estimated CAFE
requirement in MY2021 through which CAFE standards are final; and the estimated CAFE
requirement in MY2030, when NHTSA modeling indicates that the Augural Standards would
produce a fully stable fleet. The Augural Standards are assumed to remain constant at the MY
2025 level through MY 2030. Due to credit carry-forward, trading between fleets, and product
cadence considerations, NHTSA estimates that some manufacturers will be taking actions to
reach compliance with MY2025 standards for several model years thereafter. Table 13.8
indicates that, between MY2015 and MY2030, manufacturers as a group will be required to
increase required vehicle fuel economy levels by more than 50 percent for passenger cars and 40
percent for light trucks. As in previous analyses, NHTSA's analysis assumes that manufacturers
who have consistently chosen to pay CAFE fines in the past may continue to do so. However,
this analysis also assumes  an increase in NHTSA's CAFE non-compliance fine rate from $55 per
MPG under the required level per vehicle sold to $140 per MPG. As a result, the modeling
indicates that many fine-paying manufacturers will respond more aggressively to CAFE
requirements than in previous analyses.
                                            13-58

-------
                                                     Analysis of Augural CAFE Standards
 Table 13.8 Expected Manufacturer Standards and Expected CAFE levels with Augural Standards through
                                        MY2030
Manufacturer
BMW
Daimler
FCA
Ford
General
Motors
Honda
Hyundai Kia
JLR
Mazda
Mitsubishi
Nissan
Subaru
Toyota
Volvo
VWA
TOTAL
Regulatory
Class
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
Passenger Car
Light Truck
2015
Standard
35.5
29.0
34.8
29.1
35.3
28.1
35.6
26.5
35.6
26.0
36.8
29.4
35.9
29.3
33.9
29.1
36.1
30.4
38.7
31.9
36.3
28.9
37.3
31.4
36.4
28.5
35.3
30.1
36.8
29.3
36.0
27.8
CAFE
33.9
29.3
33.6
26.9
32.7
25.5
35.0
25.5
33.5
24.5
41.3
31.8
35.5
27.7
26.8
25.2
42.4
31.8
41.7
35.2
41.4
29.0
38.8
37.2
40.3
26.1
35.6
26.6
36.5
27.7
37.1
26.5
2021
Standard
44.8
35.0
43.8
35.1
44.7
33.3
45.0
30.1
45.2
29.9
46.4
36.1
45.5
36.3
42.3
35.0
45.8
36.9
48.9
39.5
45.8
34.6
46.9
39.0
46.0
33.6
44.5
36.6
46.7
35.7
45.5
32.9
CAFE
39.0
32.7
41.1
33.0
49.0
35.5
49.1
33.9
49.0
31.8
44.0
36.0
46.6
36.9
32.1
31.0
46.1
36.1
51.3
44.6
49.1
37.5
52.1
46.6
48.1
36.8
41.9
33.2
40.6
32.0
47.1
35.1
2030
Standard
54.0
42.1
52.6
42.3
53.7
40.1
54.0
36.3
54.4
36.0
56.0
43.2
54.7
43.7
50.7
42.2
55.1
44.5
59.0
47.6
55.0
41.7
56.4
47.1
55.3
40.5
53.5
44.0
56.1
43.2
54.8
39.6
CAFE
48.5
42.1
50.8
42.2
54.2
40.3
56.7
37.4
54.6
36.0
58.1
43.8
55.9
43.7
35.0
41.2
55.4
44.8
63.3
55.0
57.3
42.1
57.5
47.4
56.4
40.6
48.4
33.4
51.8
38.2
55.8
39.9
As Table 13.8 shows, among those manufacturers assumed willing to pay civil penalties as
allowed under EPCA/EISA, a few (e.g., JLR, Volvo) could find that option attractive enough to
fall well short of one or both standards by MY2030. However, also by MY2030, all
manufacturers assumed to be averse to paying CAFE fines (e.g., Ford, GM, and FCA) are
estimated to be able to reach compliance without the use of credits. Among those manufacturers,
                                            13-59

-------
                                                        Analysis of Augural CAFE Standards
several exceed the standard as fuel savings technologies applied in earlier years propagate
through shared components across platforms (discussed in greater detail in Section 13.3).

   In NHTSA's modeling, manufacturer's fleets evolve from a starting point, which is generally
defined as a description, including number of vehicles sold, fuel economy, weight, footprint,
engine and transmission type, and aerodynamic drag, of each "model" built by each
manufacturer in some recent historical model year.  In the 2012 FRM, the starting point was the
MY 2010 fleet. In this analysis, the starting point has been updated to the MY2015 fleet. Figure
13.29 shows the required and achieved CAFE levels for the MY2025 fleet simulated from the
MY2010 analysis fleet in the 2012 FRM and the MY2025 fleet simulated from the MY2015 fleet
in the current analysis. Total industry average CAFE level and standard are lower using the
MY2015 fleet in the current analysis than they were using the MY2010 fleet in the FRM, largely
attributable to the shifts in sales between light trucks and passenger cars,  described earlier in this
chapter. Both simulations show manufacturers achieving CAFE levels close to the requirements,
albeit generally closer for the passenger cars than the light trucks.
        60

        50
• 2025 CAFE (Simulated fram201E Fleet)
H 2025 CAFE (Simulated from 2010 Fleet)
— 2025 Standard (Simulated from 201 E- Fleet)
•-- 2025 Standard (Simulated from2010 Fleet)
o
  Figure 13.29 CAFE and Standard from 2010 Fleet Simulations vs. 2015 Observed Fleet (miles per gallon)
                                               13-60

-------
                                                     Analysis of Augural CAFE Standards
   Technology penetration rates for passenger cars and tight trucks

   The analysis that follows explains how the CAFE model projects manufacturers could reach
Augural Standards for both passenger cars and light trucks. This analysis simulates the
application of fuel efficiency improving technologies, however does not change vehicle footprint
or mix as a compliance strategy. The analysis is not intended to be a prediction of how any
given manufacturer will actually respond to CAFE requirements, but represents a low cost
technology solution in the context of the assumptions made in this analysis. Figure 13.30 through
Figure 13.33 show passenger car technology penetration rates for engine, transmission,
electrification, and load reduction technologies, respectively.

   Figure 13.34 through present comparable analysis for light trucks. The green values in the
tables show that reliance upon a given technology is modeled at less than 50 percent of the sales
volume for a manufacturer, and the red values highlight progressively higher dependence upon a
technology within the market.  As the tables illustrate, different manufacturers apply different sets
of technologies to raise CAFE levels and achieve compliance with the standards.

   In each table, technology complexity generally increases moving left to right, though  each
group of technologies has interdependences and mutually exclusive choices so this progression
of complexity is not always strictly increasing. For example, the DCT8 appears at the far right of
the transmission table (after the DCT6), but may be less complex than the CVT. However, the
CAFE model's logic progresses to CVTs along the automatic transmission path and models that
start as DCTs remain DCTs. The ranking merely reflects this progression.

   As Figure  13.30 shows, manufacturers across the industry are projected to deploy  most of the
lower complexity engine technologies (e.g.,variable valve timing and lift, direct injection) at
levels approaching 100 percent for most manufacturers by MY2030. However, after deploying
all of these engine technologies, manufacturers choose different levels of turbocharging
technology. At the industry level, the penetration rate of level 1 turbocharging (TURBO1) drops
over time as the rates of level 2 turbocharging (TURBO2) and cooled EGR both steadily
increase. This trend is observable for individual manufacturers as well, though most pronounced
among the primarily European manufacturers (and Ford) that already rely on TURBO 1 to a
significant degree in MY2015. Some of these manufacturers continue along the engine path to
cooled EGR, though only VWA relies on advanced diesel engines to any meaningful  degree, and
at a level that is projected to decrease over time.
                                             13-61

-------
        Analysis of Augural CAFE Standards
Figure 13.30 Passenger Car Engine Technology Penetration
fleet)
OEM
Industry


BMW


Daimler


FCA


Ford


General
Motors

Honda


Hyundai
Kia

JLR


Mazda


Mitsubishi


Nissan


Subaru


Toyota


Volvo


VWA


Model
Year
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
WT
89%
90%
94%
98%
98%
98%
83%
85%
85%
96%
93%
95%
100%
100%
100%
81%
82%
82%
55%
69%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
97%
95%
94%
100%
100%
100%
98%
97%
98%
100%
100%
100%
77%
88%
88%
WL
23%
58%
62%
98%
95%
54%
0%
12%
12%
49%
67%
57%
0%
24%
39%
18%
71%
64%
98%
98%
98%
0%
79%
91%
0%
3%
3%
2%
5%
6%
0%
35%
91%
10%
64%
89%
0%
83%
89%
2%
43%
52%
0%
0%
0%
30%
35%
20%
SGDI Cyli"der
Deac
45%
67%
69%
96%
96%
54%
94%
75%
19%
0%
67%
55%
66%
69%
63%
72%
77%
67%
45%
45%
87%
82%
87%
93%
0%
3%
3%
98%
95%
94%
0%
35%
91%
3%
58%
93%
7%
90%
94%
8%
46%
55%
0%
0%
0%
76%
87%
47%
2%
23%
22%
0%
2%
0%
0%
12%
4%
3%
4%
4%
0%
0%
0%
0%
22%
3%
13%
46%
75%
1%
23%
1%
0%
3%
3%
0%
0%
0%
0%
0%
0%
0%
46%
72%
0%
79%
37%
0%
35%
37%
0%
0%
0%
0%
6%
0%
Rates By Manufacturer (sales weighted share of
High
Comp TURBO1 TURBO2
Ratio
3%
3%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
3%
2%
0%
0%
0%
0%
98%
91%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
2%
2%
0%
0%
0%
0%
0%
0%
16%
14%
5%
92%
82%
30%
77%
61%
15%
10%
28%
24%
33%
23%
4%
18%
3%
0%
0%
0%
0%
1%
2%
0%
95%
88%
43%
0%
5%
0%
0%
0%
0%
3%
8%
0%
10%
0%
0%
1%
9%
8%
100%
77%
0%
72%
28%
4%
1%
5%
14%
4%
11%
24%
1%
0%
0%
0%
0%
0%
0%
0%
0%
3%
0%
5%
0%
0%
24%
0%
0%
0%
0%
0%
33%
0%
4%
53%
5%
41%
91%
0%
5%
22%
0%
14%
47%
0%
0%
10%
0%
23%
69%
1%
52%
42%
Cooled
EGR
0%
21%
29%
0%
0%
0%
0%
0%
0%
0%
39%
30%
0%
29%
59%
0%
57%
60%
0%
0%
0%
0%
56%
93%
0%
0%
0%
0%
0%
0%
0%
0%
5%
0%
0%
0%
0%
0%
14%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Adv.
Diesel
1%
1%
1%
1%
2%
2%
1%
1%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
15%
12%
12%
13-62

-------
        Analysis of Augural CAFE Standards
Figure 13.31 Passenger Car Transmission Penetration Rates By Manufacturer (sales weighted share of fleet)
OEM
Industry


BMW


Daimler


FCA


Ford


General
Motors

Honda


Hyundai
Kia

JLR


Mazda


Mitsubishi


Nissan


Subaru


Toyota


Volvo


VWA


Model
Year
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
5-Speed 6-Speed 8-Speed
auto auto auto
1%
0%
0%
0%
0%
0%
3%
0%
0%
2%
0%
0%
0%
0%
0%
0%
0%
0%
2%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
5%
0%
0%
0%
0%
0%
0%
0%
0%
44%
25%
1%
15%
15%
0%
0%
4%
0%
18%
6%
0%
64%
42%
0%
92%
42%
0%
10%
14%
2%
92%
26%
2%
5%
3%
0%
87%
49%
0%
1%
1%
0%
0%
0%
0%
1%
2%
0%
49%
49%
4%
25%
26%
23%
0%
0%
0%
14%
30%
44%
75%
73%
34%
95%
72%
12%
65%
60%
54%
0%
14%
48%
2%
38%
60%
4%
5%
15%
4%
68%
87%
95%
88%
78%
0%
37%
86%
0%
0%
0%
9%
12%
12%
0%
0%
1%
10%
15%
59%
75%
74%
46%
0%
0%
0%
CVT
24%
21%
21%
0%
0%
0%
0%
0%
0%
4%
2%
0%
0%
0%
0%
2%
3%
3%
74%
73%
74%
0%
0%
0%
0%
0%
0%
0%
0%
0%
87%
87%
83%
85%
81%
80%
83%
82%
80%
17%
16%
16%
0%
0%
0%
1%
2%
2%
DCT6
6%
5%
4%
0%
0%
13%
0%
0%
7%
5%
0%
0%
14%
0%
0%
0%
0%
0%
0%
0%
0%
2%
2%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
72%
73%
42%
DCT8
1%
2%
1%
5%
4%
3%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
3%
2%
2%
0%
1%
3%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
17%
16%
7%
13-63

-------
                                                         Analysis of Augural CAFE Standards
Figure 13.32 Passenger Car Electrification Technology Penetration Rates By Manufacturer (sales weighted
                                         share of fleet)
OEM
Industry


BMW


Daimler


FCA


Ford


General
Motors

Honda


Hyundai
Kia

JLR


Mazda


Mitsubishi


Nissan


Subaru


Toyota


Volvo


VWA


Model
Year
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
Stop/
Start
7%
15%
29%
86%
79%
4%
88%
66%
1%
0%
0%
3%
0%
7%
0%
18%
39%
24%
0%
0%
80%
0%
18%
55%
80%
71%
0%
0%
0%
33%
0%
32%
55%
0%
0%
56%
0%
43%
82%
0%
0%
12%
0%
0%
0%
0%
8%
2%
ISG
0%
14%
24%
0%
3%
33%
0%
0%
8%
0%
51%
51%
0%
50%
48%
0%
28%
39%
0%
0%
0%
0%
2%
34%
0%
0%
58%
1%
1%
1%
0%
34%
28%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
27%
0%
1%
45%
Strong
Hybrid
4%
10%
22%
0%
2%
57%
0%
22%
85%
0%
24%
39%
5%
27%
35%
0%
11%
30%
2%
2%
2%
3%
3%
7%
0%
8%
22%
0%
0%
0%
0%
0%
4%
0%
0%
0%
0%
0%
2%
16%
19%
18%
0%
0%
31%
0%
0%
44%
PHEV
0%
1%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
2%
2%
2%
1%
2%
2%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
BEV200
1%
2%
2%
0%
0%
0%
1%
2%
2%
1%
2%
3%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
3%
5%
6%
0%
0%
0%
0%
0%
0%
0%
0%
0%
1%
1%
1%
FCV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
                                                13-64

-------
        Analysis of Augural CAFE Standards
Figure 13.33 Passenger Car Load Reduction Technology Penetration Rates By Manufacturer (sales weighted
share of fleet)

OEM

Industry


BMW


Daimler


FCA


Ford


General
Motors

Honda


Hyundai
Kia

JLR


Mazda


Mitsubishi


Nissan


Subaru


Toyota


Volvo


VWA



Model
Year
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
Mass
Reduc
7.5%
16%
42%
80%
0%
0%
68%
53%
75%
95%
0%
57%
62%
0%
18%
92%
2%
65%
68%
0%
0%
15%
42%
69%
100%
27%
30%
80%
78%
74%
100%
0%
0%
0%
56%
89%
100%
0%
94%
100%
4%
17%
100%
0%
93%
100%
0%
8%
100%
Mass
Reduc
10%
9%
12%
20%
0%
0%
0%
0%
3%
3%
0%
27%
27%
0%
0%
28%
2%
7%
6%
0%
0%
15%
42%
46%
52%
0%
0%
0%
52%
50%
80%
0%
0%
0%
13%
11%
18%
0%
0%
0%
4%
5%
5%
0%
0%
0%
0%
0%
19%
Mass
Reduc
15%
0%
3%
8%
0%
0%
0%
0%
0%
0%
0%
27%
27%
0%
0%
28%
2%
7%
6%
0%
0%
15%
0%
0%
5%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Mass
Reduc
20%
0%
3%
8%
0%
0%
0%
0%
0%
0%
0%
27%
27%
0%
0%
28%
0%
5%
4%
0%
0%
15%
0%
0%
5%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

AERO10

10%
86%
100%
18%
89%
100%
72%
100%
100%
0%
93%
100%
0%
58%
100%
1%
96%
100%
2%
42%
100%
2%
97%
100%
0%
100%
100%
0%
77%
100%
39%
100%
100%
9%
100%
100%
0%
100%
100%
30%
100%
100%
17%
100%
100%
1%
83%
100%

AERO20

3%
47%
100%
0%
49%
100%
54%
93%
100%
0%
62%
100%
0%
58%
100%
0%
96%
100%
0%
0%
100%
0%
23%
100%
0%
97%
100%
0%
45%
100%
0%
71%
100%
0%
5%
100%
0%
94%
100%
10%
50%
100%
0%
93%
100%
0%
36%
100%

ROLL10

0%
97%
98%
0%
89%
100%
0%
98%
98%
0%
95%
97%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
95%
94%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%

ROLL20

0%
75%
98%
0%
88%
100%
0%
94%
98%
0%
95%
97%
0%
100%
100%
0%
100%
1 00%
0%
20%
100%
0%
100%
100%
0%
100%
100%
0%
26%
1 00%
0%
100%
100%
0%
90%
94%
0%
62%
100%
0%
21%
100%
0%
100%
100%
0%
98%
100%
13-65

-------
                                                        Analysis of Augural CAFE Standards
   As shown in the tables above, for the passenger car fleet, the Augural Standards are projected
to result in large increases in a wide range of technologies over the 15 year period from MY
2015 through MY2030. All manufacturers are projected to exhibit consistent and heavy reliance
on dynamic load reduction technologies like aerodynamic improvements and low rolling
resistance tires, fully utilizing opportunities for improvement in those areas, as well as modest
levels of mass reduction. However, the projections show manufacturers following a range of
different technology pathways, with differences in areas like engine, transmission, and
electrification technology, and improvements in different areas.

   As Figure 13.31 shows, passenger cars are projected to displace 6-speed automatic
transmissions with 8-speed automatic transmissions over time, with the share of CVT and DCT
remaining relatively steady over the study period. However, a number of manufacturers are
projected to heavily deploy CVTs at levels considerably higher than their application in MY2015
- Honda, Mitsubishi, Nissan, and Subaru, in particular.

   As shown in Figure 13.32, the analysis projects a consistent and increasing reliance on
start/stop, integrated starter generators (ISG), and strong hybrids. While the penetration rate of
pure electric vehicles also increases over the period, only Nissan is projected to convert more
than 3 percent of its passenger car fleet to battery electric vehicles, and most manufacturers show
no significant deployment of pure EVs11.  Similarly, the CAFE simulations project that
manufacturers would be able to achieve compliance without any reliance on fuel cell vehicles
(FCV).

   In the regulatory analysis of the MY2017 - MY2021 standards, which included Augural
Standards for MYs 2022 - 2025, NHTSA concluded that compliance could be achieved
primarily through transmission improvements and technological advances to the internal
combustion engine - without significant reliance on hybridization. Compared to the 2012 final
rule, DOT's current analysis reflects a range of updates  to the CAFE model and inputs. These
include:  changes to the market forecast (involving some changes in fleet mix and technology
and fuel economy  levels, as well as changes in other vehicle and fleet characteristics); changes in
the estimated cost  and effectiveness for different technology combinations; model revisions that
improve the accuracy of the Volpe model's accounting for product cadence and shared
technologies (e.g., shared engines);12 an increase in the civil penalty  rate (from $5.50 per 0.1
mpg to $14 per 0.1 mpg); and other changes have combined to result in new estimates of
potential technology application in response to the augural standards, including wider application
of strong hybrid penetrations for this Draft TAR as shown in Figure  13.32. As in the FRM, there
11 As Tesla Motors only produces electric vehicles, the CAFE program does not represent a binding standard. The
  industry totals include the contribution of Tesla sales to the new vehicle market, but individual results for that
  manufacturer are not expected to vary as a result of CAFE standards and are omitted from the tables.
12 Note, for engine and hybrid technologies and mass reduction levels of 10 percent or more, the NHTSA analysis
  assumes manufacturers would reduce engine displacement to maintain vehicle performance, because the change
  in performance and displacement would be moderate. For other technologies and lower levels of mass reduction,
  the NHTSA analysis assumes manufacturers would not redesign engines to preserve vehicle performance because
  performance impacts and changes in engine displacement would be smaller and would not justify the engineering
  resources and costs that would be incurred to do so. Therefore, for those other technologies, some portion of the
  fuel saving potential results in an increase in vehicle performance.
                                               13-66

-------
                                                       Analysis of Augural CAFE Standards
remains a significant reliance on turbocharging and CEGR improvements to the internal
combustion engine.13

   Notably, each manufacturer is projected to move far along one or more technology pathways
where it has little engagement in MY 2015. For example:

      •   Most European manufacturers today are producing relatively few integrated start-
          generator (ISO) or strong hybrid vehicles for the U.S. market in MY2015. But, they
          are projected to deploy those technologies on more than 75 percent (combined) of
          their passenger car fleets produced for U.S.  sale by MY2030, in response to the
          Augural Standards.

      •   Some firms are not projected to have large increases in ISG or strong hybrids, but are
          expected to focus instead on advanced gasoline engines.  For example, Honda and
          Hyundai Kia have negligible levels of turbocharging in their passenger car fleets in
          MY2015, and are projected to include turbocharging in over 20 percent of their
          passenger car engines by MY2030.

      •   Ford,  GM and Fiat-Chrysler are projected to increase market share for their full hybrid
          systems from 0-5 percent in MY2015 to 30-39 percent in MY2030, and increase ISG
          systems from 0 percent in MY2015 to 39-51 percent in MY2030.

   A similar, but not identical, story emerges for light  trucks, with the biggest differences
between technology application levels for passenger cars and light trucks being greater use of
mass reduction technology in the latter, and greater use of ISG and strong hybrids in the former.
13 This section is focused on describing internal changes within the Volpe model.  As noted in the executive
  summary and elsewhere, there are differences between the DOT and EPA approaches that derive different
  penetration rates for hybrid as well as other technologies. These derive from a range of factors, including but not
  limited to different penetration rates of EVs and PHEVs in the two agencies' reference fleets, differences in
  technology effectiveness assumptions, and others.


                                              13-67

-------
        Analysis of Augural CAFE Standards
Figure 13.34
OEM
Industry


BMW


Daimler


FCA


Ford


General
Motors

Honda


Hyundai
Kia

JLR


Mazda


Mitsubishi


Nissan


Subaru


Toyota


Volvo


VWA


Light Truck Engine Technology Penetration Rates By Manufacturer (sales weighted share of
fleet)
Model
Year
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
WT
93%
94%
98%
91%
91%
91%
72%
73%
73%
96%
96%
96%
100%
100%
100%
97%
98%
98%
38%
57%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
95%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
84%
85%
85%
WL
11%
61%
82%
91%
91%
78%
0%
50%
37%
7%
85%
96%
0%
38%
49%
0%
23%
98%
100%
100%
100%
0%
67%
85%
0%
0%
0%
12%
29%
29%
33%
33%
100%
4%
71%
93%
0%
95%
94%
0%
81%
88%
0%
7%
7%
29%
30%
38%
SGDI
41%
83%
88%
91%
91%
78%
89%
91%
56%
0%
84%
96%
55%
89%
93%
97%
96%
98%
50%
54%
52%
100%
85%
85%
0%
0%
0%
63%
100%
100%
0%
33%
100%
4%
70%
93%
2%
97%
96%
1%
82%
89%
0%
7%
7%
84%
84%
84%
Cylinder
Deac
24%
37%
32%
0%
0%
0%
0%
49%
0%
16%
17%
17%
0%
0%
0%
68%
75%
60%
62%
100%
100%
0%
29%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
35%
42%
0%
91%
70%
0%
18%
19%
0%
7%
7%
0%
5%
15%
High
Comp TURBO1 TURBO2
Ratio
1%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
63%
71%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
12%
26%
21%
91%
91%
56%
40%
42%
56%
0%
18%
20%
51%
42%
35%
2%
1%
10%
0%
0%
0%
0%
0%
0%
100%
80%
45%
0%
28%
28%
0%
5%
5%
0%
35%
19%
2%
2%
2%
1%
64%
30%
90%
2%
0%
67%
49%
46%
0%
3%
14%
0%
0%
23%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
10%
12%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
66%
0%
27%
95%
0%
0%
31%
0%
4%
24%
0%
0%
41%
0%
92%
93%
0%
19%
23%
Cooled
EGR
0%
20%
27%
0%
0%
0%
0%
0%
0%
0%
48%
59%
0%
47%
59%
0%
7%
18%
0%
0%
0%
0%
38%
85%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Adv.
Diesel
1%
1%
1%
9%
9%
9%
9%
9%
9%
4%
4%
4%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
16%
15%
15%
13-68

-------
        Analysis of Augural CAFE Standards
Figure 13.35 Light Truck Transmission Technology
Penetration Rates By Manufacturer (sales weighted
share of fleet)
OEM
Industry


BMW


Daimler


FCA


Ford


General
Motors

Honda


Hyundai
Kia

JLR


Mazda


Mitsubishi


Nissan


Subaru


Toyota


Volvo


VWA


Model
Year
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
5-Speed
auto
10%
1%
0%
0%
0%
0%
0%
0%
0%
18%
0%
0%
0%
0%
0%
0%
0%
0%
18%
0%
0%
0%
0%
0%
0%
0%
0%
12%
0%
0%
0%
0%
0%
21%
18%
0%
0%
0%
0%
22%
0%
0%
0%
0%
0%
0%
0%
0%
6 -Speed
auto
59%
31%
20%
0%
0%
0%
0%
0%
0%
24%
17%
8%
100%
2%
0%
94%
85%
71%
44%
52%
52%
100%
46%
0%
2%
2%
0%
88%
100%
0%
5%
0%
0%
0%
0%
0%
0%
0%
0%
75%
29%
0%
100%
100%
100%
0%
0%
0%
8-Speed
auto
16%
50%
58%
100%
100%
39%
100%
100%
26%
56%
81%
90%
0%
98%
93%
5%
15%
29%
0%
0%
3%
0%
38%
85%
98%
78%
0%
0%
0%
100%
0%
5%
5%
4%
11%
29%
0%
0%
0%
1%
68%
98%
0%
0%
0%
0%
0%
0%
CVT
12%
14%
14%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
38%
48%
45%
0%
0%
0%
0%
0%
0%
0%
0%
0%
95%
95%
95%
73%
69%
69%
96%
95%
95%
0%
0%
0%
0%
0%
0%
0%
0%
0%
DCT6
0%
1%
3%
0%
0%
48%
0%
0%
39%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
45%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
11%
23%
43%
DCT8
1%
2%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
89%
76%
56%
13-69

-------
        Analysis of Augural CAFE Standards
Figure 13.36 Light Truck
Electrification Technology
Penetration Rates
By Manufacturer (sales weighted
share of fleet)
OEM
Industry


BMW


Daimler


FCA


Ford


General
Motors

Honda


Hyundai
Kia

JLR


Mazda


Mitsubishi


Nissan


Subaru


Toyota


Volvo


VWA


Model
Year
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
Stop/
Start
3%
28%
47%
100%
100%
39%
78%
78%
7%
0%
34%
74%
0%
75%
93%
0%
7%
40%
0%
0%
45%
0%
67%
68%
94%
74%
0%
0%
0%
0%
0%
0%
0%
0%
9%
21%
0%
0%
0%
0%
0%
16%
0%
0%
0%
0%
15%
15%
ISG
0%
1%
2%
0%
0%
0%
0%
0%
0%
0%
5%
5%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
100%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
23%
Strong
Hybrid
0%
1%
6%
0%
0%
61%
0%
0%
74%
0%
0%
0%
0%
0%
7%
0%
0%
0%
0%
0%
0%
0%
15%
15%
0%
20%
100%
0%
0%
0%
0%
0%
0%
0%
0%
0%
3%
3%
4%
1%
1%
1%
0%
0%
0%
0%
0%
16%
PHEV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
1%
0%
0%
BEV200
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
FCV
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
13-70

-------
        Analysis of Augural CAFE Standards
Figure 13.37 Light Truck Load Reduction Technology Penetration Rates By Manufacturer (sales weighted
share of fleet)

OEM

Industry


BMW


Daimler


FCA


Ford


General
Motors

Honda


Hyundai
Kia

JLR


Mazda


Mitsubishi


Nissan


Subaru


Toyota


Volvo


VWA



Model
Year
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
2015
2021
2030
Mass
Reduc
7.5%
15%
49%
84%
0%
5%
93%
17%
33%
97%
0%
22%
61%
45%
63%
100%
6%
77%
77%
56%
49%
100%
30%
17%
100%
0%
0%
100%
0%
0%
71%
0%
0%
0%
34%
87%
100%
0%
51%
48%
0%
41%
100%
0%
98%
100%
0%
0%
100%
Mass
Reduc
10%
8%
24%
61%
0%
0%
0%
0%
16%
16%
0%
22%
61%
45%
63%
100%
6%
8%
77%
0%
0%
48%
30%
17%
71%
0%
0%
100%
0%
0%
71%
0%
0%
0%
0%
24%
62%
0%
0%
0%
0%
41%
39%
0%
0%
0%
0%
0%
36%
Mass
Reduc
15%
0%
15%
42%
0%
0%
0%
0%
0%
0%
0%
22%
61%
0%
59%
95%
0%
2%
48%
0%
0%
48%
0%
0%
54%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Mass
Reduc
20%
0%
5%
29%
0%
0%
0%
0%
0%
0%
0%
22%
38%
0%
5%
42%
0%
2%
48%
0%
0%
48%
0%
0%
54%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%

AERO10

6%
87%
100%
0%
100%
100%
33%
100%
100%
1%
100%
100%
0%
70%
100%
0%
94%
100%
27%
49%
100%
0%
55%
100%
0%
96%
100%
0%
100%
100%
0%
100%
100%
0%
95%
100%
0%
97%
100%
19%
100%
100%
0%
100%
100%
0%
70%
100%

AERO20

0%
60%
97%
0%
5%
100%
0%
100%
100%
0%
76%
100%
0%
70%
100%
0%
94%
100%
0%
0%
100%
0%
0%
100%
0%
65%
100%
0%
0%
100%
0%
67%
100%
0%
24%
100%
0%
51%
48%
0%
59%
100%
0%
98%
100%
0%
12%
100%

ROLL10

0%
99%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
83%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
100%
100%
0%
90%
100%

ROLL20

0%
76%
95%
0%
100%
100%
0%
1 00%
100%
0%
66%
100%
0%
100%
100%
0%
100%
1 00%
0%
19%
100%
0%
83%
100%
0%
100%
100%
0%
0%
1 00%
0%
100%
100%
0%
94%
100%
0%
9%
14%
0%
80%
100%
0%
100%
100%
0%
90%
100%
13-71

-------
                                                      Analysis of Augural CAFE Standards
All manufacturers make increasing, and consistently high use of engine technologies such as
variable valve timing and lift (VVT and VVL, respectively) and direct injection (SGDI). As
Figure 13.34 illustrates, those technologies are already present in the MY2015 fleet at very high
levels for some manufacturers, but, industry-wide, at lower levels than they are simulated for
MY2030. Turbocharged engines, whose penetration varies by manufacturer, are expected to be
present, in some form, on over half of the light trucks offered for sale in MY2030 compared to
slightly more than 10 percent of the MY2015 fleet.

   The penetration rates of transmission technologies for light trucks are broadly similar to those
for passenger cars, with manufacturers generally projected to rely on the same mix of
technologies for both classes. As reflected in Figure 13.35, the manufacturers projected to rely
most heavily on CVTs for their passenger car fleets are projected to have similar reliance in their
light truck fleets.
As with passenger cars, dynamic load reduction technologies (aerodynamic improvements and
LRR tires) are simulated to reach high  levels of penetration in the light truck market for all
manufacturers by MY2030 (Figure 13.37). Mass reduction technologies are projected to be
deployed at higher rates for light trucks than passenger cars.  NHTSA's analysis restricts the
applicability of mass reduction technologies for passenger cars, but not for the light trucks. In the
modeling, most manufacturers make increasing use of all levels of mass reduction, with the
highest-volume pickup truck producers (Ford, GM, and FCA) deploying the highest available
level (20 percent reduction) on around  40 percent of their light trucks. Honda and Hyundai Kia
both apply mass reduction at the highest available level on 50 percent of their fleets by MY2030.

   NHTSA modeling projects a significant increase in the use of start/stop systems within the
light truck class, but, in contrast to the  passenger car fleet, comparatively little reliance on ISO or
strong hybrid systems, as reflected in Figure 13.36, as compared with Figure 13.32.
Projected compliance costs
The technology changes described above carry associated costs. In the NHTSA model,
manufacturers can only redesign a fraction of their fleet each year, and have flexibility to over-
or under comply in a particular year by banking or borrowing credits, manufacturers compliance
pathway over time can be complex. Thus, costs for compliance with current standards and the
Augural Standards are interconnected, and evolve on a year-by-year basis to reflect annual
redesign cycles and other factors.  Table 13.9 divides  aggregate annual average per vehicle
manufacturers' compliance costs into three categories: the investments manufacturers would
have to make to comply with current standards through 2016, costs to comply with current
standards through MY2021, and the cost to comply with the MY 2022-2025 Augural Standards.
                                             13-72

-------
                                                      Analysis of Augural CAFE Standards
      Table 13.9 Average Per Vehicle Cost for Primary Analysis Using RPE to Mark Up Direct Costs

2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
Average Per-Vehicle Costs (2013 $)
Costs added with
stringency increases
through 2016
200
250
340
350
370
380
380
380
370
370
370
370
360
350
350
350
350
Additional costs with
stringency increases
through 2021
40
150
280
390
560
670
680
670
670
670
680
670
670
660
660
660
660
Additional costs under
Mys 2022-2025 Augural
Standards
-
10
70
100
190
450
610
750
860
1,020
1,120
1,230
1,250
1,250
1,240
1,250
1,250
£
(/)
o
(J
"ro
+->
£
240
400
690
830
1,120
1,500
1,670
1,800
1,900
2,070
2,160
2,260
2,270
2,260
2,250
2,260
2,260
   Note that, in NHTSA's modeling, manufacturers begin investing in compliance with the
Augural Standards as early as 2017, redesigning vehicles that will continue to be built in 2022
and beyond, as well  as accumulating credits for future compliance.
                                             13-73

-------
                                                      Analysis of Augural CAFE Standards
   The chart below (Figure 13.38) shows the rate at which average regulatory costs increase
relative to the required and achieved CAFE levels for the industry. The figure combines the
passenger car and light truck fleets, and presents compliance with the Augural Standards.
Manufacturer-specific results have more variance (especially for manufacturers with relatively
limited ranges of product offerings). Those seeking more detail can download the simulation
results in full from NHTSA's website14.
                                                                              $1,400
                                                                              $800
                                                                              $600
                                                                              $400
                                                                              $200
                                                                              s-
        2015    2017     2019     2021     2023    2025     2027     2029

                       O Required CAFE   	Achieved CAFE  	Cost
                                                                                     QJ
                                                                              $1,200  55
                                                                              $1,000  o
                                                                                     L)
                                                                                     <
                                                                                     T3
                                                                                     O
                                                                                     ST
                                                                                     _Q
                                                                                     O
                                                                                     U
                                                                                     o
QJ
cm
re
     Figure 13.38 Industry-Wide Combined Average Fuel Economy Levels and Average Costs (2015 $)
14CAFE - Fuel Economy: http://www.nhtsa.gov/Laws+&+Regulations/CAFE+-+Fuel+Economy/ld-
  cafe-midterm-evaluation-2022-25
                                              13-74

-------
                                                           Analysis of Augural CAFE Standards
   Table 13.10 below provides additional information on the distribution of projected sales and
compliance (technology plus fines/credits) costs.  The current projection of manufacturer sales
volumes (described in greater detail in Section 4.2) combines a projection of total vehicle sales
and the division between passenger cars and light trucks form the AEO, a proprietary forecast of
manufacturer market shares from IHS/Polk, and sales volume projections  for MY2015 submitted
to NHTSA by the manufacturers.
 Table 13.10 Draft TAR Average Per Vehicle Cost and Production Volume in MY 2025 for Primary Analysis
                                Using RPE to Mark Up Direct Costs
Production Volumes and Average Costs for NHTSA Draft TAR Analysis

BMW
Daimler
Fiat-Chrysler
Ford
General Motors
Honda
Hyundai Kia
Jaguar Land Rover
Mazda
Mitsubishi
Nissan
Subaru
Toyota
Volvo
Volkswagen
Industry Average
MY2025 Production
for Sale in U.S. (m)
Passenger Cars
0.30
0.23
0.61
0.94
1.20
0.88
1.08
0.02
0.22
0.06
0.89
0.15
1.22
0.04
0.63
8.59
LO
^.
u
3
i_
1—
4-»
.n
DO
'_!
0.14
0.16
1.48
1.29
1.44
0.65
0.23
0.07
0.12
0.03
0.46
0.50
1.02
0.05
0.18
7.84
(D
4-»
0
0.44
0.39
2.09
2.23
2.64
1.53
1.32
0.10
0.34
0.09
1.35
0.65
2.24
0.09
0.82
16.43
Average Per-Vehicle Costs1 in MY2025
Costs added with
stringency increases
through 20162
240
400
1,070
260
260
90
200
1,010
60
90
120
50
510
400
500
370
Additional costs with
stringency increases
through 2021
1,010
960
720
1,080
550
120
1,200
1,030
380
650
440
510
460
830
950
670
Additional costs under
MYs 2022-2025 Augural
Standards
1,180
1,180
880
1,540
1,300
1,140
1,230
1,050
870
1,170
780
610
260
1,140
1,300
1,020
LO
4-»
LO
0
U
"(5
4-»
0
2,430
2,530
2,660
2,880
2,110
1,350
2,700
3,090
1,320
1,910
1,340
1,170
1,230
2,360
2,750
2,070
 2012 Final Rule3
                   10.98
                             5.47
                                      16.45
                                                    790
                                                                            1,700
2,480
 1 Draft TAR costs in 2013 $.
 2 Costs estimated to be accrued under standards through 2016 reflect different analysis fleets and credits. The 2012 Final
 Rule analysis uses a MY 2010 baseline fleet and includes costs for MYs 2011-2016, whereas the Draft TAR uses a MY 2015
 baseline fleet and includes costs for MY 2016, alone.
 32012 Final Rule costs in 2010 $. For manufacturer-specific costs in 2012 Final Rule, see 77 FR 99, at 63047-63049,
 63063-63067 and accompanying Final RIA, pp. 675-762.
                                                  13-75

-------
                                                      Analysis of Augural CAFE Standards
   A number of factors may affect the spread of costs across these different compliance periods
covered by the program. Notably, drops in overall costs for compliance through 2016, relative to
analysis in the 2012 final rule, reflect, among other things, choices that manufacturers across the
sector have made since 2010 (the model year providing the foundation for NHTSA's 2012
analysis) with respect applying technology and to achieving compliance in the early years.
Manufacturers' choices to integrate certain technological innovations first, as part of a multi-year
program, affect the range of additional technologies available for later integration and future
savings.  The CAFE model recognizes that technologies, once implemented, are no longer
available to generate additional savings. Multi-year regulatory certainty, combined with out-year
GHG regulations as well as Augural Standards, has provided companies with a framework for
planning that allows them to implement individual redesign cycles with the ability to understand
how each may affect their range of compliance options in the future. While NHTSA's analysis
fleet for today's analysis, being based on the 2015 fleet, reflects the application of fuel-saving
technologies between 2010 and 2015, our analysis does not attempt to quantify the cost of those
improvements. Additionally, unlike NHTSA's 2012 analysis, today's analysis includes CAFE
credits that manufacturers are estimated to have available to carry forward (and, in some cases,
trade) from model year 2010-2015 and apply toward compliance obligations during 2015-2019.

   Among the metrics that can be used to weigh the relative cost of fuel economy improvements,
one is the cost of added fuel-saving technology as compared to the resultant reduction of fuel
consumption. The latter could be measured in terms of expected gallons or dollars of avoided
fuel consumption, using estimates of future vehicle survival, vehicle use, the gap between
laboratory and real-world fuel economy, and future fuel prices.  Without applying such
estimates, the reduction of fuel consumption can be measured on a percentage basis, considering
the inverses of CAFE levels (e.g., such that increasing CAFE level from 20 to 30 mpg represents
a 33.3 percent reduction in average fuel consumption, from 0.05 to 0.033 gallons per mile, or
                                             13-76

-------
                                                        Analysis of Augural CAFE Standards
   Below, Table 13.11 shows estimated model year 2028 CAFE levels under the No-Action
Alternative and the Augural Standards.  On an industry-wide basis, the Augural Standards are
estimated to improve average fuel consumption by about 14 percent, with similar average
improvements for the passenger car and light truck fleets, with variance in both directions.
       Table 13.11 Estimated MY2028 CAFE Levels and Average Fuel Consumption Improvement

Manufacturer
BMW1
Daimler1
Fiat-Chrysler
Ford
General Motors
Honda
Hyundai Kia
Jaguar Land Rover1
Mazda
Mitsubishi
Nissan
Subaru
Tesla
Toyota
Volvo1
Volkswagen
Total
CAFE (mpg) under
No-Action Alternative
u
0-
44.9
44.2
45.2
48.9
46.2
46.8
46.9
35.1
48.6
49.9
45.6
49.7
282.9
48.5
44.5
47.1
47.4
b
35.0
35.1
34.7
31.5
30.6
37.5
39.7
34.6
39.5
39.5
35.8
41.8
282.9
37.2
33.4
35.4
34.4
Combined
41.4
40.0
37.4
37.2
36.2
42.5
45.6
34.7
44.9
46.4
41.7
43.5
282.9
42.7
38.0
44.0
40.3
CAFE (mpg) under
Augural Standards
u
Q.
47.0
50.8
54.0
56.7
54.4
58.1
55.9
35.1
55.8
58.1
55.7
57.6
282.9
55.5
48.2
51.4
55.4
b
35.7
42.2
40.3
37.4
35.9
44.0
43.6
38.8
44.8
46.6
42.0
47.4
282.9
40.5
33.4
38.0
39.6
Combined
42.8
47.0
43.7
43.7
42.5
51.3
53.4
37.8
51.3
54.3
50.2
49.5
282.9
47.7
39.3
47.8
46.7
Fuel Consumption
Improvement (% gpm)
u
Q.
4%
13%
16%
14%
15%
20%
16%
0%
13%
14%
18%
14%
0%
13%
8%
8%
14%
t
2%
17%
14%
16%
15%
15%
9%
11%
12%
15%
15%
12%
0%
8%
0%
7%
13%
Combined
3%
15%
14%
15%
15%
17%
15%
8%
12%
14%
17%
12%
0%
10%
3%
8%
14%
NOTE:
'Manufacturer assumed to be willing to pay civil penalties as allowed under EPCA/EISA, if doing so would be more financially
attractive than further increasing average fuel economy.
                                               13-77

-------
                                                       Analysis of Augural CAFE Standards
   Table 13.12 shows the estimated average additional cost in MY 2028 (compared to the No-
Action Alternative) of fuel-saving technologies producing these incremental fuel consumption
improvements under the Augural Standards. On an industry-wide basis, and excluding any
estimated civil penalties, these estimated incremental costs average about $1,110 for passenger
cars, $1,250 for light trucks, and $1,175 for the combined fleet. Estimated average incremental
costs vary considerably between manufacturers' respective fleets. However, after normalizing
for relative improvements in average fuel consumption, these cost differences are more tightly
distributed around the industry-wide average levels of $77 per % for passenger cars, $95 per %
for light trucks, and $86 per % for the combined fleet.
     Table 13.12 Estimated Technology Cost per Percent Fuel Consumption Improvement in MY2028

Manufacturer
BMW1
Daimler1
Fiat-Chrysler
Ford
General Motors
Honda
Hyundai Kia
Jaguar Land
Rover1- 2
Mazda
Mitsubishi
Nissan
Subaru
Tesla2
Toyota
Volvo1-2
Volkswagen1
Total
Fuel Consumption
Improvement (% gpm)
u
0-
4%
13%
16%
14%
15%
20%
16%
0%
13%
14%
18%
14%
0%
13%
8%
8%
14%
b
2%
17%
14%
16%
15%
15%
9%
11%
12%
15%
15%
12%
0%
8%
0%
7%
13%
Combined
3%
15%
14%
15%
15%
17%
15%
8%
12%
14%
17%
12%
0%
10%
3%
8%
14%
Add! Tech. Cost (20 13$)
under Augural Standards
u
Q.
431
1,224
1,288
1,374
1,637
1,208
1,443
-
923
976
902
949
-
617
764
729
1,111
b
119
1,719
1,335
1,931
1,604
1,095
1,037
1,029
866
1,468
878
497
-
652
-
650
1,246
Combined
336
1,422
1,321
1,693
1,620
1,162
1,378
769
903
1,115
894
609
-
632
376
712
1,174
Add! Tech. Cost (20 13$)
per % Improvement
u
Q.
97
94
79
100
109
62
90

72
69
50
69

49
98
87
77
b
68
101
97
123
109
75
118
96
73
96
59
42

80

94
95
Combined
96
96
92
113
109
67
94
96
72
77
53
50

61
115
89
86
       'Manufacturer assumed to be willing to pay civil penalties as allowed under EPCA/EISA, if doing so would be more
financially attractive than further increasing average fuel economy.

       2Blank entry indicates no incremental change compared to No-Action Alternative.
   Table 13.12 reports average fuel consumption improvements and technology costs on an
incremental basis. Measured relative to vehicles that continue with fuel economy and
technology at model year 2015 levels, the added fuel-saving technologies appear considerably
more cost-efficient.
                                              13-78

-------
                                                      Analysis of Augural CAFE Standards
   Table 13.13, Table 13.14, and Table 13.15 show total costs and average additional per vehicle
costs (above 2015 levels) for the baseline case, or the "No Action Alternative."  The three tables
show passenger cars, light trucks, and cars and trucks combined, respectively. The No Action
Alternative encompasses compliance with existing standards through MY 2021. The variations
in post-2021 costs have diverse causes.  Specialty manufacturers, such as Volvo or Jaguar Land
Rover, selling few models, may have very "lumpy" redesign costs, while manufacturers may
liquidate credit balances in particular years.  And all manufacturers are projected to incur
additional technology costs as redesigns to shared engines, transmissions, and platforms that
occurred in prior model years propagate through to all the models on which they are shared. In
some cases, this sharing crosses the boundary between light trucks and passenger cars - where an
engine (for example) must be updated to achieve compliance with the passenger car standard, but
then eventually filters through to all of the light trucks that share that engine when they are
redesigned. It is also the case, for some manufacturers, that credits earned in earlier model years
have been carried forward (in the simulation), provide opportunities for technology application
to bring the fleet into compliance with MY2021 standards in model years 2022 and beyond.
                                             13-79

-------
                                                         Analysis of Augural CAFE Standards
Table 13.13 Passenger Cars: Total Cost and Average per Vehicle Costs for the No-Action Alternative for the
                       Primary Analysis Using RPE to Mark Up Direct Costs
Total Costs (2013 $b)
under

-------
                                                        Analysis of Augural CAFE Standards
Table 13.14 Light Trucks:  Total Cost and Average per Vehicle Costs for the No-Action Alternative for the
                      Primary Analysis Using RPE to Mark Up Direct Costs
Total
Costs (2013 $b)
under No-Action Alternative

-------
                                                       Analysis of Augural CAFE Standards
Table 13.15 All Vehicles:  Total Cost and Average per Vehicle Costs for the No-Action Alternative for the
                      Primary Analysis Using RPE to Mark Up Direct Costs
Total
Costs (2013 $b)
under No-Action Alternative

-------
                                                      Analysis of Augural CAFE Standards
Table 13.16, Table 13.17, and Table 13.18 show the additional cost and average per vehicle cost
by manufacturer of complying with the Augural  Standards for MY2022 - 2025.  The three tables
cover passenger cars, light trucks, and all light duty vehicles, respectively.  These costs are in
addition to the costs for the "No Action" alternative, shown in Table 13.13, Table 13.14, and
Table 13.15.

   As noted above, manufacturers, as simulated by the model, begin investing in compliance
with the Augural Standards from 2016, both to get ahead of the redesign cycle and also to obtain
bankable credits by applying relatively lower cost technologies. Costs rise with time, are mostly
trivial before MY2019 and then flatten after MY2027 as manufacturers exhaust carried-forward
credits and bring both fleets into compliance. Per vehicle compliance costs are generally similar
for passenger cars and light trucks, though there  are large variations across manufacturers, based
on each manufacturer's product line and available technology choices.
                                             13-83

-------
                                                       Analysis of Augural CAFE Standards
Table 13.16 Passenger Cars Additional Total Cost and Average per Vehicle Costs for the MYs 2022-2025
          Augural Standards for the Primary Analysis Using RPE to Mark Up Direct Costs

Manufacturer
BMW
Daimler
Fiat-Chrysler
Ford
General
Motors
Honda
Hyundai Kia
Jaguar Land
Rover
Mazda
Mitsubishi
Nissan
Subaru
Toyota
Volvo
Volkswagen
Total
Total Additional


-------
                                                      Analysis of Augural CAFE Standards
Table 13.17 Light Trucks: Additional Total Cost and Average per Vehicle Costs for the MYs 2022-2025
          Augural Standards for the Primary Analysis Using RPE to Mark Up Direct Costs


Manufacturer
BMW
Daimler
Fiat-Chrysler
Ford
General
Motors
Honda
Hyundai Kia
Jaguar Land
Rover
Mazda
Mitsubishi
Nissan
Subaru
Toyota
Volvo
Volkswagen
Total
Total Additional
under
rs:
™ rs:
O
rs:
0.0
0.0
2.1 0.9
2.0 1.9

1.9 0.6
0.1 0.3
0.1

0.0
0.0
0.1 0.0
0.3 0.2
0.2 0.2
-
0.0
0.0
6.7 4.5
Costs (2013 $b) Additional Average Costs (2013 $)
Augural Standards under Augural Standards

m
rs:
o
rs:
0.1
0.1
0.9
2.1

1.0
0.3
0.1

0.0
0.1
0.0
0.2
0.3
0.0
0.0
0.1
5.4
oo
rs:
rsirsir>lrsirsirsirsirsirsirsirsi
OOinOOOOOOOO
rsirsij^jrsirsirsirsirsirsirsirsi
O
rs:
0.1 0.1 0.5 - 192 471 728 1,010 1,052 1,093 1,030
0.1 0.1 0.7 - 177 408 593 902 1,319 1,313 1,871
1.1 1.2 5.4 550 627 618 763 776 1,037 1,349 1,335
2.1 2.3 7.9 1,351 1,428 1,605 1,568 1,773 1,960 1,949 1,931

1.3 1.5 6.2 424 420 714 927 1,055 1,114 1,570 1,604
0.4 0.8 2.2 175 510 520 582 1,159 1,144 1,115 1,095
0.2 0.3 0.7 - 278 544 851 1,081 1,064 1,046 1,037

0.1 0.1 0.3 - 213 449 703 859 1,025 1,334 1,359
0.1 0.1 0.3 - 4 940 920 902 888 875 866
0.0 0.0 0.1 1,281 1,247 1,214 1,333 1,546 1,520 1,493 1,468
0.2 0.3 1.1 384 446 474 509 699 702 739 878
0.2 0.3 0.7 397 527 526 517 500 492 493 497
0.0 0.0 1.6 - - 25 26 33 314 578 652
0.0 0.1 0.2 - 279 529 796 1,085 1,084 1,107 1,114
0.2 0.2 0.7 - 236 513 1,023 1,330 1,369 1,342 1,268
6.2 7.4 28.5 477 568 695 792 939 1,081 1,264 1,289
                                             13-85

-------
                                                       Analysis of Augural CAFE Standards
  Table 13.18 All Vehicles: Additional Total Costs and Average per Vehicle Costs for the MYs 2022-2025
            Augural Standards for the Primary Analysis Using RPE to Mark Up Direct Costs

Manufacturer
BMW
Daimler
Fiat-Chrysler
Ford
General
Motors
Honda
Hyundai Kia
Jaguar Land
Rover
Mazda
Mitsubishi
Nissan
Subaru
Toyota
Volvo
Volkswagen
Total
Total Additional Costs (2013 $b) Additional Average Costs (2013 $)

lrsirsirsirsirsirsirsirN
OOOi^OOOOOOOO
rsirsirsicsirsirsirsirsirsirsirsirsi
O
rs:
0.2 0.4 0.5 1.7 - 271 561 881 1,182 1,180 1,206 1,291
0.2 0.3 0.5 1.6 - 302 583 874 1,175 1,286 1,280 1,587
1.6 1.8 1.8 7.9 653 730 778 874 878 1,097 1,311 1,321
3.0 3.0 3.4 12.0 1,122 1,262 1,359 1,345 1,538 1,718 1,708 1,693

2.6 3.0 3.4 12.1 706 723 985 1,131 1,299 1,322 1,611 1,620
0.9 1.2 1.7 5.5 127 500 624 793 1,141 1,127 1,178 1,162
1.2 1.3 1.7 5.7 221 771 895 1,025 1,295 1,406 1,395 1,378

0.1 0.1 0.1 0.4 - 248 497 759 1,050 1,177 1,379 1,384
0.3 0.3 0.3 0.9 247 373 893 887 869 925 913 903
0.1 0.1 0.1 0.3 879 1,146 1,111 1,126 1,169 1,161 1,141 1,115
0.8 0.9 1.1 3.4 393 521 606 714 783 796 817 894
0.4 0.4 0.4 1.1 403 594 614 629 613 602 610 609
0.2 0.3 0.6 3.7 36 37 84 122 264 399 578 632
0.0 0.1 0.1 0.3 - 263 541 833 1,138 1,144 1,269 1,267
0.4 0.8 1.1 3.4 - 291 587 971 1,303 1,369 1,395 1,379
12.1 13.9 16.8 59.9 450 613 749 857 1,024 1,118 1,225 1,245
   Table 13.19, Table 13.20, and Table 13.21 show total costs and average per-vehicle costs for
each manufacturer, based on compliance both with existing standards and the Augural Standards,
with the three tables showing passenger cars, light trucks, and all light duty vehicles,
respectively. These tables are the sum/average of the corresponding tables for the no-action
                                               13-86

-------
                                                       Analysis of Augural CAFE Standards
alternative and Augural Standards (above), and may be interpreted as an estimate of future costs
to be incurred by manufacturers for all current and Augural post-2015 fuel economy standards.

   Table 13.19 Passenger Costs:  Total Cost and Total Average per Vehicle Costs for the MYs 2022-2025
            Augural Standards for the Primary Analysis Using RPE to Mark Up Direct Costs


Manufacturer
BMW
Daimler
Fiat-Chrysler
Ford
General
Motors
Honda
Hyundai Kia
Jaguar Land
Rover
Mazda
Mitsubishi
Nissan
Subaru
Toyota
Volvo
Volkswagen
Total


rsi
O
rsi
LT)
O
rsi
1.0
0.8
6.1
6.0

7.4
0.7
5.8

0.2
0.3
0.2
2.1
0.6
2.4
0.2
2.2
36.1
Total Costs (2013 $b) Total Average Costs (2013 $)
under Augural Standards under Augural Standards
oo
rsi
rsi ro ^ LD ^ * — i rsi ro ^ LD {& l^ oo
rsi rsi rsi rsi f"^ rsi rsi rsi rsi rsi rsi rsi rsi
O O O O in O O O O O O O O
rsi rsi rsi rsi ,=3 rsi rsi rsi rsi rsi rsi rsi rsi
O
rsi
0.5 0.6 0.7 0.8 2.7 1,391 1,730 2,079 2,399 2,720 2,751 2,796 2,914
0.4 0.5 0.6 0.7 1.9 1,469 2,108 2,333 2,627 2,889 2,751 2,718 2,847
1.9 1.9 1.8 1.9 6.1 3,029 3,054 3,189 3,118 3,058 3,148 3,110 3,160
2.5 2.4 2.3 2.7 9.3 2,442 2,656 2,612 2,621 2,881 3,083 3,117 3,080

2.5 2.8 2.9 3.2 9.5 2,185 2,217 2,391 2,441 2,637 2,612 2,682 2,645
0.6 0.8 1.0 1.2 4.0 198 664 923 1,190 1,361 1,349 1,457 1,441
2.3 2.4 2.5 2.8 9.2 1,553 2,228 2,292 2,367 2,619 2,745 2,720 2,685

0.1 0.1 0.1 0.1 0.3 2,333 2,644 2,878 3,147 3,290 3,276 3,294 3,447
0.2 0.2 0.3 0.3 0.9 501 814 1,200 1,228 1,204 1,315 1,300 1,288
0.1 0.1 0.1 0.1 0.3 1,527 2,043 1,994 1,961 1,907 1,899 1,871 1,851
0.8 0.9 1.1 1.1 3.5 787 976 1,088 1,237 1,237 1,255 1,264 1,305
0.3 0.3 0.3 0.3 0.9 1,507 1,879 1,986 2,061 2,011 1,975 1,962 1,949
0.7 0.7 0.9 1.2 4.0 562 604 656 738 980 979 1,082 1,119
0.1 0.1 0.1 0.1 0.3 1,027 1,319 1,625 1,933 2,252 2,258 2,640 2,607
1.1 1.3 1.5 1.8 5.6 1,574 1,862 2,183 2,497 2,866 2,907 2,948 2,934
13.9 15.2 16.3 18.2 58.3 1,441 1,704 1,836 1,937 2,118 2,164 2,205 2,215
                                              13-87

-------
                                                        Analysis of Augural CAFE Standards
Table 13.20 Light Trucks: Total Cost and Total Average per Vehicle Costs for the MYs 2022-2025 Augural
               Standards for the Primary Analysis Using RPE to Mark Up Direct Costs


Manufacturer
BMW
Daimler
Fiat-Chrysler
Ford
General
Motors
Honda
Hyundai Kia
Jaguar Land
Rover
Mazda
Mitsubishi
Nissan
Subaru
Toyota
Volvo
Volkswagen
Total


rsi
O
rsi
LT)
O
rsi
0.3
0.8
13.9
7.0

5.9
0.4
1.8

0.5
0.3
0.1
2.0
0.9
7.5
0.2
0.7
42.3
Total Costs (2013 $b) Total Average Costs (2013 $)
under Augural Standards under Augural Standards
oo
rsi
rsi ro ^ LD ^ * — i rsi ro ^ LD {& l^ oo
rsi rsi rsi rsi f"^ rsi rsi rsi rsi rsi rsi rsi rsi
O O O O in O O O O O O O O
rsi rsi rsi rsi ,=3 rsi rsi rsi rsi rsi rsi rsi rsi
O
rsi
0.1 0.2 0.2 0.3 0.8 764 977 1,213 1,580 1,841 1,828 1,814 1,744
0.2 0.2 0.3 0.3 1.2 1,124 1,167 1,457 1,750 2,036 2,475 2,457 3,015
3.5 3.5 3.6 3.7 12.9 2,288 2,403 2,369 2,481 2,503 2,745 3,037 3,004
3.5 3.6 3.5 3.7 12.7 2,480 2,553 2,715 2,667 2,875 3,170 3,126 3,090

1.5 1.9 2.2 2.4 8.8 1,055 1,036 1,322 1,522 1,668 1,721 2,170 2,199
0.4 0.4 0.5 0.9 2.6 269 600 654 723 1,339 1,323 1,345 1,323
0.5 0.6 0.7 0.7 2.0 2,003 2,224 2,525 2,874 3,067 3,028 2,995 2,958

0.2 0.2 0.2 0.2 0.8 1,661 1,910 2,169 2,398 3,023 3,187 3,473 3,459
0.0 0.2 0.2 0.2 0.5 416 412 1,589 1,557 1,527 1,508 1,490 1,475
0.0 0.0 0.0 0.1 0.1 1,698 1,656 1,615 1,726 1,931 1,900 1,870 1,841
0.6 0.6 0.6 0.7 2.2 1,275 1,329 1,335 1,354 1,541 1,551 1,580 1,724
0.4 0.5 0.5 0.5 1.3 860 980 958 939 923 914 913 902
1.6 1.6 1.6 1.6 6.0 1,627 1,576 1,578 1,547 1,529 1,792 2,039 2,098
0.1 0.1 0.1 0.1 0.3 1,382 1,718 1,954 2,200 2,462 2,449 2,386 2,373
0.2 0.3 0.4 0.4 1.2 1,119 1,271 1,457 2,088 2,336 2,352 2,313 2,224
13.0 13.8 14.5 15.7 53.5 1,565 1,642 1,762 1,852 2,006 2,157 2,331 2,340
                                               13-88

-------
                                                        Analysis of Augural CAFE Standards
Table 13.21 All Vehicles:  Total Cost and Total Average per Vehicle Costs for the MYs 2022-2025 Augural
              Standards for the Primary Analysis Using RPE to Mark Up Direct Costs

Manufacturer
BMW
Daimler
Fiat-Chrysler
Ford
General
Motors
Honda
Hyundai Kia
Jaguar Land
Rover
Mazda
Mitsubishi
Nissan
Subaru
Toyota
Volvo
Volkswagen
Total



-------
                                                      Analysis of Augural CAFE Standards
Sensitivity of Cost to Key Inputs
   Just as the estimated costs and technology application rates developed in the 2012 analysis
have been demonstrated to be sensitive to changing conditions in the real world over time, so the
results of the Draft TAR analysis are as well. NHTSA examined how alternative assumptions
about critical inputs to the simulation would change outcomes of interest. Table 13.22 describes
the range of assumptions considered for each sensitivity case as well as the aspects of the CAFE
compliance and effects simulation that are impacted by the assumption. As the remainder of this
section shows, not all assumptions will impact all metrics of interest.
                Table 13.22 Definition of Sensitivity Cases Considered For Draft TAR
Sensitivity
Fuel prices
MR
restrictions
Lifetime VMT
Battery costs
MR costs
Product
Cadence
Rebound
Effect
Demand for
FE
Safety
coefficients
Description
AEO 2015 fuel price cases
Vary the PC restriction with
existing costs
Higher/Lower Lifetime VMT
than current schedule
Higher/Lower battery costs
than current
Higher/lower MR cost curves
Vary length of existing
redesigns
Span range in rebound
literature
Varies amount that OEMs
assume consumers are
willing to pay for additional
fuel economy beyond CAFE
levels (months)
5th and 95th percentile of
safety coefficients
High case
AEO 2015 high
No restrictions
35% - 55%
higher lifetime
None
NAS cost
2 years longer
30%
36 months
95th
Low case
AEO 2015 low
All PCs stop at
MR1 (unless they
already have >
MR1)
14% - 27% lower
lifetime
$100/kwh
Fraction of NAS
2 years shorter,
adds as many as
two redesigns to
study period
0%
0 months
5th
Affects
Value of fuel savings, PC/LT split,
over compliance, technology
choices, combined required CAFE,
achieved CAFE, fine payment
Tech choices, societal safety, net
benefits, achieved CAFE
Tech choices, crash exposure and
societal safety, fuel savings
Tech choices/penetration, tech
costs, fuel savings, achieved CAFE
Tech choices, tech cost, fine
payers, safety
Tech choices, tech cost, achieved
CAFE, over compliance
Fuel savings, crash exposure and
societal safety, externalities,
mobility benefit
Fuel savings, Achieved CAFE, net
benefits, over compliance, tech
choices
Societal safety, net benefits
                                             13-90

-------
                                                       Analysis of Augural CAFE Standards
   The two bar plots in Figure 13.39 and Figure 13.40, show the percentage change in regulatory
costs (technology costs plus fines) under these alternative assumptions. The first of these shows
the change in total regulatory costs under the Augural Standards over the study period (for the
industry) incremental to the continuation of the final standards through MY2021. Figure 13.39
shows that considerably lower battery costs can lower estimated compliance costs for the
industry - which also produces a different technology solution than described earlier in this
section, as higher levels of electrification become more cost competitive. Battery costs are an
important element in the cost of employing battery electric vehicles and hybrids. However, they
also  affect how these two technologies compete with each other and with other technologies.

   One result that may seem counterintuitive is the fact that longer product cadence (more years
between redesigns) actually reduces the incremental cost. However, this is a result of increasing
costs in the baseline relative to the central analysis described above - as manufacturers have
fewer opportunities to apply technology during the augural standard period, more technology is
added in earlier model years, reducing the incremental cost of the Augural Standards. Changes in
the price of oil (relative to the AEO2015 reference case that  informs the central Draft TAR
analysis) influences the share of light trucks in the new vehicle fleet, as well as consumer
preferences for fuel saving technology. Both factors influence incremental (and total) regulatory
cost  attributable to the Augural Standards.  Similarly, by assuming that consumers are willing to
pay for more fuel saving technology above and beyond the levels required by  CAFE standards,
more technology is applied in the baseline to satisfy consumer demand for fuel economy, leaving
less technology that needs to be applied under the Augural Standards and reducing the
incremental cost attributable to them. Other alternative assumptions had smaller impact on
incremental cost.
          Battery Costs
         Dem and for FE
        Product Cadence
            Fue Free
          Lifetime VMT
             MRcosts
         MR restrictions
• High Value

• Low Value
                            -0.15       -D.I        -:.:E        D        o.os
                                 Percentage Change in Incremental Regulatory Costs
 Figure 13.39 Sensitivity of Incremental Regulatory Costs (MY2016 - MY2030) to Alternative Assumptions
                                              13-91

-------
                                                      Analysis of Augural CAFE Standards
         Product Cadence
           Battery Costs
              MR costs
           Lifetime VMT
              Fuel Price
          MR restrictions
           Demand for FE
I Low Value

I High Value
                     -10
                             -5      0      5      10      15      20

                                 Percentage Change in Total Regulatory Cost
                                                                       25
                                                                              30
    Figure 13.40 Sensitivity of Total Regulatory Costs (MY2016 - MY2030) to Alternative Assumptions
   As Figure 13.40 shows, the rank ordering of importance changes in the context of total, rather
than incremental, regulatory cost over the period.  Where assumed demand for fuel economy was
of critical importance to the attribution of cost to the Augural Standards, when accounting for the
change in total cost between MY2016 and MY2030, it makes a much smaller difference -
influencing total cost only through the series of technology solutions that appear attractive to
manufacturers.  However, the assumption with the highest influence on total cost is now product
cadence - where longer design cycles limit manufacturers' choices and lead to cost increases
approaching 30 percent over the central analysis.  Battery costs, while less important than
product cadence, influences total cost in the direction one would expect (as do mass reduction
cost cases), though by less than 10 percent.

   NHTSA also conducted a sensitivity case analysis using indirect cost multiplier (ICM) in
place of retail price equivalent (RPE) which was used for the primary analysis.  In developing
cost estimates for technologies applied to new vehicles, the manufacturing cost of a particular
element,  for example,  a continuously variable transmission, is only a portion of the total cost of
placing a new technology on a vehicle.  The full cost of the part includes not just manufacturing,
but also research and development costs, overhead, future warranty costs, and other elements.
RPE and ICM methodologies for estimating indirect costs are discussed in Chapter 5.  Table
13.23 shows production volumes and average per  vehicle costs for both in the 2012 final rule
analysis and the Draft TAR analysis.
                                             13-92

-------
                                                      Analysis of Augural CAFE Standards
Table 13.23 Comparison of Cost Estimates Using Retail Price Equivalent and Indirect Cost Multiplier Mark
                                           Up







20 12 Final Rule


Draft TAR

MY2025
Production
for Sale in U.S.



b
53
O
bo


10.98

8.59



1
H
.£P


5.47

7.84



IS
£


16.45

16.43
Average Per- Vehicle Costs1 in MY2025
(using RPE to Mark Up Direct Costs)



w
Costs added with
stringency increas
through 20 16 2


785

370

,£3
•~£H M
Additional costs w
stringency increas
through 2021
Is
u &
"S 3.
Additional costs u
MYs 2022-2025 ,
Standards

1 698


671

1,024
ft
1

Total costs added
augural standards


2,483

2,065















20 12 Final Rule

Draft TAR

MY2025
Production
for Sale in U.S.




E

O
bo
S
U}
%


10.98

8.59






o
s
^
r*
• 5P


5.47

7.84








^^
+s
o
H


16.45

16.43
Average Per- Vehicle Costs1 in MY2025
(Side Case using ICM to Mark Up Direct Costs)



t/3
t/3
^ §N

^ .§ S
^ >. o
""O ^ ^
™ w ob
w S o
6 a •§

588

322


^ w

„, g
t/1 ^
o y ^
U .S 04
1 &°
0 O ^3
•S a I3
^3 S^ ^
|l|

•g
S Si
""rt ^
§<

S 
r 3
n^ Cu

1,913

1,859
   1 Note: 2012 Final Rule costs in 2010 $. Draft TAR costs in 2013 $.

   2 Costs estimated to be accrued under standards through 2016 reflect different analysis fleets
and credits.
13.3.2 Consumer Impacts

   As the stringency of CAFE standards increase over time, the average technology cost required
for manufacturers to reach compliance will generally increase as well. Cost inputs to today's
analysis reflect DOT's judgment that manufacturers are likely to pass future increases in
production costs on to consumers, recouping direct and indirect costs, and realizing profits that
reflect historical norms. To the extent that demand is elastic, manufacturers may absorb some of
the increased technology costs or elect to cross-subsidize some vehicles.  Manufacturers might
wish to cross-subsidize as a compliance strategy, and/or to respond to competitive pressures, to
build volume, to encourage particular customer classes to buy their vehicles, or as a profit-
                                              13-93

-------
                                                      Analysis of Augural CAFE Standards
maximization strategy. Since we do not have sufficient information to model the way in which
manufacturers actually price their current and future fleets, we cannot make credible assumptions
about what share of increased technology costs will be passed directly onto the buyer of a
specific vehicle, absorbed by the manufacturer, and/or subsidized by the purchase of other
vehicles. Without the information to establish representative assumptions about how each
manufacturer will allocate increased costs, we track the increase in technology costs associated
with a vehicle, but do not project the change in vehicle price to the consumer.

   However, given the uncertainty about how manufacturers will actually allocate costs across
their individual models, NHTSA uses the average per-vehicle regulatory cost15 increase as a
means of characterizing the magnitude of the impact of increased technology costs at the
manufacturer level.

   Although the CAFE model does not currently estimate a potential market response to changes
in vehicle prices, it does contain data on initial purchase cost (defined as current year (2015)
MSRP reported by the manufacturer) and pro-forma final vehicle purchase  cost (defined as 2015
MSRP plus added technology cost to meet the applicable  standard) for each specific vehicle
model and configuration. NHTSA staff have tested a variety of approaches  to allocating
regulatory costs, and the CAFE model currently applies a "pay  as you go" approach—for
example, if a given vehicle model configuration incurs $1,000 in additional technology costs
(after markup), the CAFE model currently reports that vehicle model configuration's purchase
cost increasing by $1,000 for the Augural Standards, compared with the cost of compliance with
the baseline standards.  As noted, these are not an accurate estimate of either initial production
cost or initial consumer price, nor the compliance production cost or compliance consumer price.
They do, however, provide a general indication of the price range of particular models, and also
gives some indication  of the starting point for manufacturer's consumer price optimization
decisions.

   NHTSA simulates each model year explicitly and includes several years beyond MY2025,
during which the Augural Standards are assumed to remain static at their MY2025 level. As
manufacturers use earned and traded credits to manage the degree of modification to either fleet
in a single year, it may still be necessary to apply technology after MY2025 in order to reach a
stable compliance solution where the fleet can comply with CAFE without using the credit carry-
forward provision (though trades between passenger car and truck fleets are likely, even once the
standards stabilize). Table  13.24 summarizes information that is available  in a cross-section of
tables in the section discussing industry impacts, and illustrates the industry average cost
increase projected between MY2016 and MY2028 as a result of the final and augural CAFE
standards. At the industry level,  the average cost increase is similar for passenger cars and light
trucks, though individual manufacturers can observe larger differences in average cost between
the two  classes over the course of the simulation. By the time the fleet reaches a stable
compliance level in MY2028, both classes of vehicles are projected to incur over $2,000/vehicle
in compliance  costs relative to the MY2015 vehicle (assuming RPE methodology).
 ' The combination of technology cost and fines for non-compliance.
                                             13-94

-------
                                                      Analysis of Augural CAFE Standards
             Table 13.24 Average Regulatory Cost per Vehicle by Model Year, 2015 - 2028
Light Trucks
Model
Year
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
Baseline
Standards
Through MY
2021
300
450
750
850
950
1,100
1,050
1,050
1,050
1,050
1,100
1,050
1,050
Augural
CAFE
Standards
MYs 2022-
2025
-
-
100
100
150
500
550
700
800
950
1,100
1,250
1,300
Total
300
450
800
950
1,100
1,550
1,650
1,750
1,850
2,000
2,150
2,350
2,350
Passenger Cars
Baseline
Standards
Through
MY2021
200
350
500
600
900
1,000
1,050
1,050
1,000
1,000
1,000
1,000
1,000
Augural
CAFE
Standards
MYs 2022-
2025
-
-
50
100
200
400
650
800
900
1,100
1,150
1,200
1,200
Total
200
350
550
700
1,150
1,450
1,700
1,850
1,950
2,100
2,150
2,200
2,200
   While, as noted above, initial purchase cost as measured by MSRP combined with technology
cost are not accurate indicators of either actual consumer prices, nor the increment to consumer
prices that would be induced by the Augural Standards, they do provide a general indication of
the expected costs that manufacturers would face, based on NHTSA modeling, and hence may
represent a reasonable starting point in determining incremental changes in consumer prices.

  NHTSA staff have examined how model-by-model estimates of technology costs are
distributed by a range of possible proxies for production cost or consumer prices, including
footprint (bigger vehicles might cost more), initial MSRP, MSRP plus technology cost, (higher
MSRP would generally indicate higher consumer prices), and curb weight (heavier vehicles
might cost more). Regression results indicate that there is little relationship between modeled
per-vehicle incremental technology costs and any of these indicators, as scatter plots show a
classic "cloud" with an essentially arbitrary regression line and show estimated elasticities of
around -0.01, and R2 of 0.02 to 0.04.  In other words, individual vehicle technology  costs are
rather evenly distributed across the range of vehicle cost, measured by multiple proxies for
vehicle price or cost, and almost none of the variation in technology costs is explained by these
proxies.
                                              13-95

-------
                                                      Analysis of Augural CAFE Standards
   The analysis does not attempt to account for any potential cross-subsidy by manufacturers.
Although additional analysis could explore various hypotheses about manufacturers' pricing
strategies, in the absence of proprietary information about manufacturers' actual costs, prices,
and production plans, it would not be possible to demonstrate whether any particular hypothesis
was true or false.  NHTSA modeling suggests that Augural Standards will increase average
vehicle technology costs by about $1,000 per vehicle relative to the average price of a new
vehicle under continuation of the MY2021  standard, and we can reasonably expect that
manufacturers will wish to raise vehicle prices on average. We cannot, however, predict the
extent to which each manufacturer will choose to mix price increases, other cost reductions, and
reduced margins in the aggregate, nor how  these decisions will be distributed across the vehicles
in each manufacturer's fleet.

   All of these factors come into play in considering how manufacturers might choose to set
prices for new vehicles in the lower price range of the new car market. The initial vehicle fleet
contains 8 models that have an initial MSRP under $15,000 (see Chapter 6.5 for a discussion of
affordability).  Manufacturers have historically used pricing strategies that allow them to service
both high and low margin market segments while maintaining overall profitability, often with a
view toward building enduring brand loyalty.

   Consumer response to manufacturers' pricing decisions is also likely to be heterogeneous
across consumer classes. Consumers'  are likely to place varying valuations on improvements in
fuel economy and other attributes of particular vehicles, and buyers of some vehicles are likely to
be more price sensitive than others. Manufacturers' strategies, in turn, will be based, in part, on
their a priori assessments of consumer response.

   In addition to the probability  that vehicles will have  higher costs, the deployment of some
new fuel economy technologies  is likely to be noticeable to new car buyers. While incremental
technology changes have often been transparent to new car buyers, for whom an automatic
transmission with more gears or a somewhat lighter or more aerodynamic vehicle would not
necessarily be  obvious, the pace and degree of new technology deployment estimated in the
Draft TAR analysis suggests that even casual observers will be aware that new vehicles may be
different in important ways.16 For example by MY 2030, 76 percent of passenger cars and 55
percent of light trucks are projected to have technology that shuts the engine off at idle, including
stop-start, integrated starter-generator (mild hybrid), full hybrid system, or plug in hybrid
(PHEV) technology. Turbocharged engines account for almost half of new vehicles by
MY2030. These technologies may be  perceived as positive or negative changes by consumers or
as items that provide greater or lesser value. Accordingly, this may influence consumer choices
about new vehicle purchases. (See Chapter 6.4 for more detail on consumer acceptance.)

   To the extent that new vehicle cost increases are passed on to consumers, other consumer cost
elements that scale with purchase price, including interest on car  loans, insurance, and some
taxes and fees would also increase. NHTSA's analysis includes estimates of some of these types
of impacts.
16 Compared to the 2012 final rule, DOT's current analysis reflects a range of updates to the CAFE model and
inputs. These are described further earlier in this chapter in section 13.3.1.
                                              13-96

-------
                                                       Analysis of Augural CAFE Standards
   While NHTSA modeling supports that new car buyers are likely to pay more to purchase,
register, and insure their new vehicles under the CAFE standards, they as well as subsequent
owners will definitely pay less to operate them. A commonly used approach to describing the
heuristic that consumers might use to consider the impact of a higher purchase cost offset by
reduced operating cost is the "payback period" for incremental technology. Payback period is
defined as the number of years of the accumulated dollar value of fuel savings needed to recover
the additional cost of technology included in the purchase price of a new vehicle. Payback period
is related to, but different from, an economic benefit calculated as a net present value of social
benefits and costs over the life of the vehicle. Since payback periods are used to simulate
consumer decisions, they use private costs and benefits, including any avoided excise taxes,
rather than social costs and benefits. While regulations with short payback periods will usually
have net economic benefits, regulations with long payback periods  do not necessarily have
negative economic benefits.

   Figure 13.41 shows the payback period associated with the technology cost increases for new
cars and trucks in each as a result of three regimes, using the same projected fuel prices, based on
the EIA's Annual Energy Outlook, as the rest of the analysis. The payback periods for the
baseline standards are calculated relative to the costs and fuel economy in the MY2015 fleet. The
payback periods for the Augural Standards are based on the incremental costs  and  fuel  savings
relative to the baseline (i.e., current standards through MY2021 carried forward). The "total"
case, represents the world consumers would actually see if the Augural  Standards are
implemented, and it is defined relative to the MY2015 fleet fuel economy. In the case of the total
scenario, it represents the payback period associated with cost increases in all future model years
(assuming the final standards through MY2021 and the Augural Standards from MY2022 -
MY2025, then carried forward unchanged through MY 2032) and fuel savings relative to the
MY2015 fuel economy levels.
Cars
 10.0

  35

  9.C

  B.«

  3.C

  75

  7.0

  6.5
                                             Trucks
  J.O

  = 5

  2.0

  1 5

  1.0

  05
  0.0
                    2023   2025
                     Model Year
2022   2024
   Model Year
  Figure 13.41 Payback Periods for the Baseline Standards, Augural Standards, and Total over the Period
                                              13-97

-------
                                                     Analysis of Augural CAFE Standards
   The payback periods in the early model years (prior to MY 2019 or MY 2020) are not really
meaningful for the Augural Standards, as the incremental cost associated with the Augural
Standards in those years are small (under $100/vehicle) and the resulting fuel savings nearly
trivial. As the figure shows, payback periods under all three scenarios are longer for cars than for
trucks. Passenger cars have comparable average per-vehicle costs under the total program, but
start from higher fuel economy levels in general. Improving the fuel economy of a less efficient
vehicle leads to greater savings because the same percentage improvement (say, 20 percent, for
example) represents a larger absolute savings, since the number of gallons consumed by the less
efficient vehicle was larger to start with (so, 20 percent of 600 gallons compared to 20 percent of
250 gallons). In addition, light trucks, on average, are also driven more miles annually than
passenger cars, so they accrue greater fuel  savings per year. These factors cause the trucks to
pay back faster than the cars, in general. Additionally, the trend for the augural standard payback
periods is generally downward (trending shorter in successive model years), despite representing
fewer gallons of savings relative to the baseline standards. Rising fuel prices over the study
period are sufficient to counteract the rising costs associated with increasingly stringent
standards, so that payback periods decline  even when the average cost increase for a new vehicle
is rising over successive model years.

   The payback period associated with the incremental impact of the Augural Standards is longer
than both the baseline and the combined program, for much the same reason. Fuel economy has
diminishing returns - once a vehicle becomes very efficient, improving its fuel economy further
saves progressively less fuel because the vehicle consumes so little in the first place. For
example, a vehicle that gets 60 MPG and drives 15,000 miles per year consumes 250 gallons of
fuel per year. If we improve the fuel economy of that vehicle by 20 percent, improving its fuel
economy to 72 MPG at a cost of $500, we  save 40 gallons per year. However, at a fuel price of
$3/gallon that fuel economy improvement  takes more than 4 years to pay back. If instead, we
increase the fuel economy of a vehicle that also drives  15,000 miles per year, but gets only 25
MPG, by the same 20 percent, we save  100 gallons per year. At a fuel price of $3/gallon, that
same $500 investment pays back in less than 2 years.

   The consumer effects of the standards are likely to be heterogeneous across different
consumers. The amount of the additional technology costs that manufacturers are able to pass
onto consumers, and the amount of the technology costs that are borne by the consumer of the
vehicles with these technologies, will depend on the  elasticity of demand of particular models,
the price of gasoline, and acceptance of new technologies, and the value that consumers place on
fuel economy Without this information, we are only  able to talk in terms of average costs across
the industry without making the assumption that demand is inelastic and manufacturers will not
cross-subsidize.

   Another aspect of consumer cost is depreciation, defined as the difference between the
purchase price of the vehicle and its subsequent market value as a used vehicle. NHTSA does
not attempt to model depreciation, and how depreciation would be affected by Augural  Standards
depends, in part, on how new and used car buyers value improved fuel economy, and if there is a
difference.  If new car buyers value fuel  economy, and manufacturers notice, then they will face
higher prices for fuel efficient vehicles, not necessarily at a level  related to the cost of providing
fuel economy.   If new car buyers place low or zero value on fuel economy, then manufacturers
will be less able to raise prices.
                                             13-98

-------
                                                      Analysis of Augural CAFE Standards
   If new car buyers and used car buyers have similar attitudes towards fuel economy, then
depreciation will scale with the price of the vehicle, but if used car buyers value fuel economy
but new car buyers do not, then new car buyers would get a benefit in the form of reduced
depreciation, while used car buyers have to pay extra for their fuel savings.  In the reverse case,
where used car buyers do not value fuel economy, but new car buyers do, then new car buyers
face increased depreciation, while used car buyers get a double bonus: used cars are less
expensive and have reduced operating costs.

   Car buyers might value fuel economy, but they may be willing to pay less for a more fuel-
efficient vehicle than the out-of-pocket fuel savings anticipated over the vehicle's expected life.
This could occur if fuel economy improvements are associated with decreases in other desirable
vehicle attributes. Car buyers' willingness to pay may also be less than the value of fuel savings
calculated here because buyers have a higher apparent discount rate than what is assumed in this
TAR. As discussed above, NHTSA applies a one-year payback period in its compliance and
technology application analysis (and assumes manufacturers will recoup all direct and indirect
costs and realize normal levels of profit). This one-year payback assumption attempts to address
the possible concerns with assuming either that new car and truck buyers place no value on fuel
economy or place a sufficiently high value on additional fuel economy to contradict historical
observations of preferences in the new car market (where trends toward smaller, more fuel
efficient vehicles under high fuel price scenarios have typically retreated as the fuel price fell).

13.3.3  Social and Environmental Impacts

   While the concept of incremental social benefits more appropriately used to rank a series of
alternatives, it is still possible to characterize some of the trends that NHTSA expects to see as a
result of the current final standards and Augural Standards. In addition to conserving the
nation's energy, two significant benefits of CAFE standards are the reduction in criteria
pollutants that affect individual health and the reduction in greenhouse gas emissions that affect
climate change.  And Figure 13.42, below, compares  the impact on criteria emissions and
greenhouse gas emissions of the Draft TAR analysis and the 2012 final rule analysis.

   The figure shows that the savings in emissions, fuel gallons, and fuel quads of total energy
consumption are generally larger under the Draft TAR analysis than the 2012 analysis. While the
savings attributable to passenger cars decreased for both gallons and metric tons of CCh saved,
the increases attributable to light trucks more than offset those reductions. Although the schedule
that (largely) determines lifetime mileage accumulation for each vehicle is lower in the Draft
TAR than in the 2012 analysis, the number of vehicles on the road is higher, and total VMT for
the overall fleet is higher in the Draft TAR than in the 2012 final rule.
                                             13-99

-------
                                                         Analysis of Augural CAFE Standards
     550%






     500%






     450%





     400%






     350%






   1 300%
   1C

   "cc
   c

   'I 250%






   g» 200%

   '>
   •
   M

   •- 150%
   •
   Dl
   C
   re

   O 100%




   I

   a!  50%





       0%






     -50%






    -100%






    -150%
     -200%
'111
          NOx (t)  SO2 (t)  PM (t)  VOC (t)  CO2   Quads Gallons  VMT   CO (t)

                                     (MMT)  Energy
   Reg.Class
   • LightTruck
   • PassengerCar


     Figure 13.42 Comparison of Environmental And Physical Effects, Draft TAR and 2012 Final Rule



   Of particular note in Figure 13.42is the magnitude of the difference in emissions savings for

the conventional tailpipe pollutants (NOx and PM). Since the 2012 final rule analysis was

conducted, additional tailpipe standards have been implemented that reduce the long-term

emissions of these pollutants, and the increase in total VMT relative to the 2012 analysis

increases the opportunity to reduce emissions. While the additional VMT associated with the

rebound effect does increase the emissions of conventional pollutants  from vehicle tailpipes, the
                                               13-100

-------
                                                        Analysis of Augural CAFE Standards
reduction in upstream emissions from avoided fuel consumption is significantly larger - and
produces social benefits.

   Another impact that requires consideration is the impact CAFE standards may have on
societal safety, as manufacturers reduce the mass of vehicles to improve fuel economy and
vehicle owners increase their travel demand as a result of lower operating costs. Figure 13.43
shows the additional fatalities attributable to the Augural Standards for passenger cars and light
trucks (by color). As discussed in Chapter 8 of the Draft TAR, reducing the mass of large light
trucks generally has a beneficial impact on societal safety, while the mass reduction in small
passenger cars has a negative effect. Both classes are projected to increase the number of miles
driven as fuel economy increases (compared to the MY2015 vehicle), however, for light trucks,
the increase in exposure to crashes is mitigated by the fact that reducing the mass of those
vehicles reduced the severity of the crashes.
  350

  300

  250

  200

  150

  100

1 50

|  0


| -5°

|-100

  -150

  -200

  -250

  -300

  -350

  -400

  -450
    2014 2015   2016  2017  2018   2019  2020   2021  2022  2023  2024  2025   2026  2027   2028  2029  2030  2031  2032 2033
                                            Model Year
Reg-Class
• Light! ruck
• PassertgerCar
      Figure 13.43  Societal Safety Effects for the Augural Standards (relative to MY2021 standards)
   As Figure 13.43 shows, the number of fatalities associated with passenger cars under the
Augural Standards grows over time as expected (because this figure measures the incremental
impacts of the Augural Standards), but the bars below the x-axis represent fatalities avoided by
changes to light trucks. The amount of mass reduction that can be applied to passenger cars has
been limited in the analysis to achieve overall neutral societal safety, thus showing a pathway
manufacturers could use to comply with the Augural Standards that has small net reductions in
fatalities over the period when considering both mass reduction and increased VMT.
                                              13-101

-------
                                                       Analysis of Augural CAFE Standards
13.3.4 Overall Benefits and Costs

Table 13.25 summarizes the costs and benefits associated with the implementation of the
Augural Standards for MYs 2022 - 2025, relative to the continuation of the MY 2021 standard
over the same period (through MY 2028). The social costs associated with the program are
primarily a direct result of technology applied to new vehicles to reach compliance with the
standards, and appears  in the table as technology cost and maintenance cost (resulting from the
incremental cost of maintaining more expensive and complicated technology - though in this
analysis it is mostly attributable to the cost of replacing low rolling resistance tires over a
vehicle's life). In addition to these "cash" costs, are the social costs of the additional travel that
results when the cost of driving is reduced as a result of increases in fuel efficiency. These have
been grouped together for presentation (though calculated separately), and represent the cost to
society of increased vehicular fatalities, crashes (that do not result in fatalities), congestion, and
road noise.

   The primary benefit of CAFE standards accrue as a result of avoided fuel expenditures by
new car and truck buyers. This single category of benefits is sufficient to ensure that the Augural
Standards result in net benefits, though it is not the only benefit to society that accrues primarily
to buyers of new vehicles. Like the value of fuel savings, other significant social benefits accrue
to new car and truck buyers, in particular the value of time associated with less frequent
refueling events and the value of additional travel that buyers of more efficient vehicles receive.
The latter serves to reduce fuel savings (since the additional driving consumes fuel), but the
value of that travel to the individual exceeds the value of the gallons that would have been saved
by foregoing the additional  travel. Three categories of benefits are the result of reducing
externalities that impact society as a result of vehicular travel. Energy security represents the
economic risk associated with dependence on oil and exposure to price shocks, the social cost of
carbon emissions estimate the long-term economic impact of global climate change, and the
conventional pollutant category represents the health savings from reducing exposure to
conventional pollutants emitted by vehicle tailpipes and throughout other parts of the fuel
production and supply cycle. All costs and benefits are discounted at 3 percent from the year in
which they occur.

   As the table  shows, pre-tax fuel savings are about 15 percent higher for light trucks in this
analysis. The projected market share of light trucks is closer to half the market, and trucks have
greater opportunities to save fuel both because they start from a lower level of fuel economy and
are driven more, on average. While the sum of benefits accruing to buyers of new cars and trucks
significantly exceeds the additional cost of new technology (and maintenance) borne by those
consumers, the  benefits associated with social externalities (only) do not. This was true for the
analysis supporting the 2012 final rule CAFE standards as well.
                                             13-102

-------
                                                      Analysis of Augural CAFE Standards
   Table 13.25 Estimated Present Value of Costs, Benefits and Net Benefits ($b) Over the Lifetimes of MYs
                     2016-2028 Vehicles Using 3 Percent Discount Rate (2013$)

Social Costs
Technology Cost
Maintenance Cost
Crashes, Fatalities,
Congestion, Noise
Social Benefits
Pre Tax Fuel Savings
Refueling Time Savings
Energy Security
Social Cost of Carbon
Emissions1
Increased Mobility
Conventional Pollutants
Net Benefits
MY 2022 - MY 2025
Augural Standards
Light
Truck

42
2
-3

64
3
5
14
5
6
55
Passenger
Cars

45
2
9

56
3
4
12
4
5
28
Total

88
5
6

122
6
9
27
9
11
85
              1 [Social cost of carbon to be added]
Sensitivity of Net Benefits to Key Inputs
   NHTSA examined how alternative assumptions about critical inputs to the simulation would
change outcomes of interest. Table 13.22 describes the range of assumptions considered for each
sensitivity case as well as the aspects of the CAFE compliance and effects simulation that are
impacted by the assumption. The effects on net benefits are shown below.

   Figure 13.44 is type of bar plot often referred to as a "tornado plot," due to the shape it creates
when sensitivities are ranked by their degree of influence on an outcome. It illustrates the change
in net benefits  attributable to the Augural Standards that results from using the alternative
assumptions described in Table 13.22. The end points of each bar indicate the magnitude and
direction of the change in net benefits that results from applying the alternative assumption
represented by the color of the bar, where blue represents the low value and gray the high value
described in Table 13.22for each of the assumptions listed on the left hand side. The reference
point is defined as the sum of benefits and costs over model years 2016 to 2030, relative to the
continuation of the MY2021 CAFE standards.
                                             13-103

-------
                                                       Analysis of Augural CAFE Standards
               Fuel Prize


                Rebound


             Dem and for FE


             LifetimeVMT


            Product Cadence


             MR restrictions


           Safety Coefficients


             Battery Costs


               MRccsts
                                  -::-'<     OK     20%     40%

                                      • Low Value  • High Value
                                                                     ;:=<
                                                                            ICGJb
   Figure 13.44 Influence of Alternative Assumptions on Net Benefits Attributable to Augural Standards
   As in the preceding discussion, we see that assumed fuel prices have the largest influence on
the net benefits attributable to the Augural Standards in the Draft TAR analysis. While the low
oil price case reduces net social benefits by nearly 30 percent, the high oil price case increases
net benefits by over 80 percent. In general, the sensitivity cases all move in intuitive directions.
For example, lower costs for mass reduction and battery technologies increase net benefits (while
higher mass reduction costs reduce them, but by a trivial magnitude). Like fuel price, rebound
impacts the benefits of the program in a way that would be present if we considered net benefits
relative to the 2012 final rule baseline. While assuming no rebound effect increases net benefits
by about 15 percent, assuming a high rebound effect reduces them by 30 percent.

   As we saw in the summary of industry impacts, the assumed consumer demand for fuel
economy does not significantly impact total technology cost (across both the baseline and
Augural Standards) but it does influence the  amount of additional cost, and benefit, that can be
attributed to the Augural Standards. If manufacturers assume that consumers will  continue to
value additional fuel saving technology, even after a manufacturer has reached compliance with
CAFE standards, more of that technology will appear in the baseline absent further increases in
stringency, and the fuel  savings associated with those technologies will net out of the baseline.

   As we also saw in the discussion of sensitivity to industry outcomes, product cadence may
play an  important role. The figure shows that a longer assumed cadence, which has the potential
to reduce manufacturers' opportunities to comply with an increase in standards during the year in
which it occurs, is likely to result in additional technology into products redesigned in earlier
model years. Similarly, shorter cadence increases the opportunities for manufacturers to respond
to increasingly stringent standards in the model years where the increases occur - forcing more
of the technology cost, and fuel savings benefit, into the model years covered by the Augural
Standards.
                                              13-104

-------
                                                      Analysis of Augural CAFE Standards
   The alternative assumptions about both mass reduction application and safety coefficients
have the directional impact on net benefits that one should expect. Including up to 20 percent
mass reduction for passenger cars, reduces net benefits by about 20 percent due to the impact on
overall societal safety. In contrast, allowing no mass reduction on passenger cars has a much
smaller impact on net benefits. Applying values for the safety coefficients in the 5th and 95th
percentile of their confidence interval produces the expected impact on fatalities, which results in
changes to net benefits in the 10-15 percent range. The combination of these two factors should
continue to emphasize the degree to which safety is an important consideration of the CAFE
program, and the expected social benefits associated with CAFE standards.
                                             13-105

-------