Joint Technical Support Document:
Final Rulemaking for 2017-2025
Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate
Average Fuel Economy Standards
&EPA
United States
Environmental Protection
Agency
*****
NHTSA
www.nhtsa.gov
-------
Joint Technical Support Document:
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
U.S. Environmental Protection Agency
and
National Highway Traffic Safety Administration
U.S. Department of Transportation
&EPA
United States
Environmental Protection
Agency
*****
NHTSA
www.nhtsa.gov
EPA-420-R-12-901
August 2012
-------
Contents
Executive Summary i
Chapter 1: The Baseline and Reference Vehicle Fleets 1-1
1.1 Why do the agencies establish baseline and reference vehicle fleets? 1-1
1.2 The 2008 and 2010 based vehicle fleet projections 1-2
1.2.1 Why did the agencies develop two fleet projections for the final rule? 1-2
1.3 The 2008 Based Fleet Projection 1-4
1.3.1 On what data is the MY2008 baseline vehicle fleet based? 1-4
1.3.2 The MY 2008 Based MY 2017-2025 Reference Fleet 1-13
1.3.3 What are the sales volumes and characteristics of the MY 2008 based
reference fleet? 1-25
1 4 The 2010 MY Based Fleet 1-31
1.4.1 On what data is the MY 2010 baseline vehicle fleet based? 1-31
1.4.2 The MY 2010 Based MY 2017-2025 Reference Fleet 1-40
1.4.3 What are the sales volumes and characteristics of the MY 2010 based
reference fleet? 1-48
1.5 What are the differences in the sales volumes and characteristics of the MY 2008
based and the MY 2010 based reference fleets? 1-54
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting, and
How Were They Developed? 2-1
2 1 Why are standards attribute-based and defined by a mathematical function? 2-1
22 What attribute are the agencies adopting, and why? 2-2
2.3 What mathematical functions have the agencies previously used, and why? 2-7
2.3.1 NHTSA in MY 2008 and MY 2011 CAFE (constrained logistic) 2-7
232 MYs 2012-2016 Light Duty GHG/CAFE (constrained/piecewise linear) 2-8
2.3.3 How have the agencies defined the mathematical functions for the MYs
2017-2025 standards, and why? 2-8
2.4 What did the agencies propose for the MYs 2017-2025 curves? 2-9
2.4.1 What concerns were the agencies looking to address that led them to change
from the approach used for the MYs 2012-2016 curves? 2-10
242 What methodologies and data did the agencies consider in developing the
2017-2025 curves presented in the proposal? 2-12
2.5 Once the agencies determined the appropriate slope for the sloped part, how did
the agencies determine the rest of the mathematical function? 2-49
-------
Contents
2.5.1 Cutpoints for Passenger Car curve 2-49
2.5.2 Cutpoints for Light Truck curve 2-51
253 Once the agencies determined the complete mathematical function shape,
how did the agencies adjust the curves to develop the proposed standards and
regulatory alternatives? 2-54
2.6 What does the agencies' updated analysis indicate? 2-56
Chapter 3: Technologies Considered in the Agencies' Analysis 3-2
3.1 What Technologies did the agencies consider for the final 2017-2025 standards?3-3
3.2 How did the agencies determine the costs of each of these technologies? 3-8
3.2.1 Direct Costs 3-8
322 Indirect Costs 3-13
3.2.3 Cost reduction through manufacturer learning 3-23
324 Costs Updated to 2010 Dollars 3-28
3.3 How did the agencies determine effectiveness of each of these technologies? 3-29
3.3.1 Vehicle simulation modeling 3-29
3.3.2 Lumped parameter Modeling 3-69
3.4 What cost and effectiveness estimates have the agencies used for each technology?
3-76
3.4.1 Engine technologies 3-77
3.4.2 Transmission Technologies 3-101
3.4.3 Vehicle electrification and hybrid electric vehicle technologies 3-112
3.4.4 Hardware costs for charging grid-connected vehicles 3-203
3.4.5 Other Technologies Assessed that Reduce CO2 and Improve Fuel Economy
3-208
3.5 How did the agencies consider real-world limits when defining the rate at which
technologies can be deployed? 3-249
3.5.1 Refresh and redesign schedules 3-249
3.6 How are the technologies applied in the agencies' respective models? 3-258
3.7 Maintenance and Repair Costs Associated with New Technologies 3-259
3.7.1 Maintenance Costs 3-260
372 Repair Costs 3-264
Chapter 4: Economic and Other Assumptions Used in the Agencies' Analysis 4-1
-------
Contents
4.1 How the Agencies use the economic and other assumptions in their analyses 4-1
4.2 What assumptions do the agencies use in the impact analyses? 4-2
4.2.1 The on-road fuel economy "gap" 4-2
422 Fuel prices and the value of saving fuel 4-7
4.2.3 Vehicle Lifetimes and Survival Rates 4-9
424 VMT 4-12
4.2.5 Accounting for the fuel economy rebound effect 4-18
4.2.6 Benefits from additional driving 4-26
4.2.7 Added costs from increased vehicle use 4-26
4.2.8 Petroleum and energy security impacts 4-28
4.2.9 Air pollutant emissions 4-39
4.2.10 Reductions in emissions of greenhouse gases 4-48
4.2.11 Benefits due to reduced refueling time 4-53
4.2.12 Discounting future benefits and costs 4-54
4.2.13 Additional Costs of Vehicle Ownership 4-54
Chapter 5: Air Conditioning, Off-Cycle Credits, and Other Flexibilities 5-1
5.1 Air conditioning technologies and credits 5-1
5.1.1 Overview 5-1
5.1.2 Air Conditioner Leakage 5-3
5.1.3 COi Emissions and Fuel Consumption due to Air Conditioners 5-22
5.1.4 Air Conditioner System Costs 5-58
52 Off-Cycle Technologies and Credits 5-62
5.2.1 Reducing or Offsetting Electrical Loads 5-64
522 Waste Heat Recovery 5-65
523 High Efficiency Exterior Lights 5-69
524 Solar Panels 5-73
5.2.5 Definitions for Electrical Load Offsetting and Reduction Technologies....5-81
5.2.6 Active Aerodynamic Improvements 5-81
5.2.7 Definition for Active Aerodynamic Improvements 5-84
5.2.8 Advanced Load Reductions 5-84
5.2.9 Thermal (and Solar) Control Technologies 5-101
iii
-------
Contents
5.2.10 Glazing 5-101
5.2.11 Active Seat Ventilation 5-106
5.2.12 Solar Reflective Paint 5-107
5213 Passive and Active Cabin Ventilation 5-108
5.2.14 Summary of Thermal (and Solar) Control Credits 5-109
5.2.15 Definitions for Solar Control Credit Technologies 5-110
5216 Summary of Credits 5-110
5.3 Full-Size Pickup Truck Credits 5-111
5.3.1 Full-Size Pick-up Truck Definition 5-112
5.3.2 Hybrid Pickup Truck Technology 5-113
533 Mild and Strong Hybrid Pickup Truck Definitions 5-113
5.3.4 Pickup Truck Performance Thresholds for Advanced Technology Credits .5-
118
IV
-------
Executive Summary
Executive Summary
The Environmental Protection Agency (EPA) and the National Highway Traffic
Safety Administration (NHTSA) are issuing a joint final rule to establish new standards for
light-duty highway vehicles that will reduce greenhouse gas emissions and improve fuel
economy. This joint final rulemaking is consistent with the Presidential Memorandum issued
by President Obama on May 21, 2010, requesting that NHTSA and EPA develop through
notice and comment rulemaking a coordinated National Program to reduce greenhouse gas
emissions and improve the fuel economy of light-duty vehicles for model years 2017-2025.
This final rule, consistent with the President's request, responds to the country's critical need
to address global climate change and to reduce oil consumption. EPA is regulating
greenhouse gas emissions standards under the Clean Air Act, and NHTSA is regulating
Corporate Average Fuel Economy standards under the Energy Policy and Conservation Act,
as amended. These standards apply to passenger cars, light-duty trucks, and medium-duty
passenger vehicles, covering model years 2017 through 2025. They require these vehicles to
meet an estimated combined average emissions level of 163 grams of CO2 per mile in MY
2025 under EPA's GHG program, and 49.6 mpg in MY 2025 under NHTSA's CAFE
program and represent a harmonized and consistent national program (National Program).
These standards are designed such that compliance can be achieved with a single national
vehicle fleet whose emissions and fuel economy performance improves each year from
MY2017 to 2025. This document describes the supporting technical analysis for areas of
these joint rules which are consistent between the two agencies.
NHTSA and EPA have coordinated closely to create a nationwide joint fuel economy
and GHG program based on consistent compliance structures and technical assumptions. To
the extent permitted under each Agency's statutes, NHTSA and EPA have incorporated the
same compliance flexibilities, such as averaging, banking, and trading of credits, off-cycle
credits, and the same testing protocol for determining the agencies' respective fleet-wide
average final standards. In addition, the agencies have worked together to create a common
baseline fleet and to harmonize most of the costs and benefit inputs used in the agencies'
respective modeling processes for this joint final rule.
Chapter 1 of this joint TSD provides an explanation of the agencies' methodology
used to develop the baseline and reference case vehicle fleets, including the technology
composition of these fleets, and how the agencies projected vehicle sales into the future. One
of the fundamental features of this technical analysis is the development of these fleets, which
are used by both agencies in their respective models. In order to determine technology costs
associated with this joint rulemaking, it is necessary to consider the vehicle fleet absent a
rulemaking as a "business as usual" comparison. In past CAFE rulemakings, NHTSA has
used confidential product plans submitted by vehicle manufacturers to develop the reference
case fleet. In responding to comments from these previous rulemakings that the agencies
make these fleets available for public review, the agencies created a new methodology for
creating baseline and reference fleets using data, the vast majority of which is publicly
available.
-------
Executive Summary
Chapter 2 of this document discusses how NHTSA and EPA developed the
mathematical functions which provide the bases for the final car and truck standards. NHTSA
and EPA worked together closely to develop regulatory approaches that are fundamentally the
same, and have chosen to use an attribute-based program structure based on the footprint
attribute, similar to the mathematical functions used in the MYs 2012-2016 rule. The
agencies revisited other attributes as candidates for the standard functions, but concluded that
footprint remains the best option for balancing the numerous technical and social factors.
However, the agencies did adjust the shape of the truck footprint curve, in comparison to the
MYs 2012-2016 rule. The agencies also modified the way the car and truck curves change
from year to year compared to the MYs 2012-2016 rule. In determining the shape of the
footprint curve, the agencies considered factors such as the magnitudes of CO2 reduction and
fuel savings, how much that shape may incentivize manufacturers to comply in a manner
which circumvents the overall goals of the joint program, whether the standards' stringencies
are technically attainable, the utility of vehicles, and the mathematical flexibilities inherent to
the statistical fitting of such a function.
Chapter 3 contains a detailed analysis of NHTSA and EPA's technology assumptions
on which the final regulations were based. Because the majority of technologies that reduce
GHG emissions and improve fuel economy are identical, it was crucial that NHTSA and EPA
use common assumptions for values pertaining to technology availability, cost, and
effectiveness. The agencies collaborated closely in determining which technologies would be
considered in the rulemaking, how much these technologies would cost the manufacturers
(directly) in the time frame of the final rule, how these costs will be adjusted for learning as
well as for indirect cost multipliers, and how effective the technologies are at accomplishing
the goals of improving fuel efficiency and GHG emissions.
Chapter 4 of this document provides a full description and analysis of the economic
factors considered in this joint final rule. EPA and NHTSA harmonized many inputs
capturing economic and social factors, such as the discount rates, fuel prices, social costs of
carbon, the magnitude of the rebound effect, the value of refueling time, and the social cost of
importing oil and fuel.
Chapter 5 of this joint TSD discusses adjustments and credits to reflect technologies
that improve air conditioner efficiency, that improve efficiency under other off-cycle driving
conditions, and that reduce leakage of air conditioner refrigerants that contribute to global
warming. The air conditioner credits are similar to the MYs 2012-2016 rule, with two notable
exceptions: NHTSA is allowing A/C efficiency improvements to help come into compliance
with fuel economy standards, and a new air conditioner test procedure is introduced to help
capture efficiency credits. NHTSA is now also allowing off-cycle improvements to help
manufacturers come into compliance with fuel economy standards. A list of some
technologies and their credits and a streamlined methodology is provided by the agencies to
help simplify the credit generating process. Chapter 5 also discusses adjustments to
encourage "game changing" technologies (such as hybridized powertrains) for full-size
pickup trucks.
-------
The Baseline and Reference Vehicle Fleets
Chapter 1: The 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 2017-2025 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. Recent volatility in oil prices and the
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 rules that would affect that future fleet. The agencies have examined various
publicly-available sources, 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 each of the two baseline and
reference fleets.
The agencies have 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 checkNHTSA's and EPA's modeling results, and perform their own analyses with
these datasets.
1.1 Why do the agencies establish baseline and reference vehicle fleets?
In order to calculate the impacts of the final GHG and CAFE standards, it is necessary to
estimate the composition of the future vehicle fleet absent the new standards. EPA and
NHTSA have developed a baseline/reference fleet in two parts. The first step was to develop
a "baseline" fleet. The agencies create a baseline fleet in order to track the volumes and types
of fuel economy-improving and CCVreducing technologies that are already present in the
existing vehicle fleet. Creating a baseline fleet helps to keep, to some extent, the agencies'
models 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 2017-2025. This is called the "reference" fleet, and
it represents the fleet volumes (but, until later steps, not additional levels of technology) that
the agencies believe would exist in MYs 2017-2025 absent any change due to regulation in
2017-2025.
After determining the reference fleet, a third step is needed to account for technologies
(and corresponding increases in cost and reductions in fuel consumption and CC>2 emissions)
that could be added to the baseline technology vehicles in the future, taking into account
previously-promulgated standards, and assuming MY 2016 standards apply at the same levels
through MY 2025. This step uses the OMEGA and CAFE models to add technologies to
vehicles in each of the baseline market forecasts such that each manufacturer's car and truck
CAFE and average CO2 levels reflect MY 2016 standards. The models' output, the
1-1
-------
The Baseline and Reference Vehicle Fleets
"reference case", is the light-duty fleet estimated to exist in MYs 2017-2025 without new
GHG/CAFE standards. All of the agencies' estimates of emission reductions/fuel economy
improvements, costs, and societal impacts for purposes of this final rulemaking (FRM) are
developed in relation to the agencies' reference cases. This chapter describes the first two
steps of the development of the baseline and reference fleets. The third step of technology
addition is developed separately by each agency as the outputs of the OMEGA and CAFE
models (see Chapter 3 of the TSD for an explanation of how the models apply technologies to
vehicles in order to evaluate potential paths to compliance).
1.2 The 2008 and 2010 based vehicle fleet projections
1.2.1 Why did the agencies develop two fleet projections for the final rule?
Although much of the discussion in this and following sections describes the
methodology for creating a single baseline and reference fleet, for this final rule the agencies
actually developed two baseline and reference fleets. In the NPRM, the agencies used 2008
MY CAFE certification data to establish the "2008-based fleet projection."a The agencies
noted that MY 2009 CAFE certification data was not likely to be representative since it was
so dramatically influenced by the economic recession (Joint Draft TSD section 1.2.1). The
agencies further noted that MY 2010 CAFE certification data might be available for use in the
final rulemaking for purposes of developing a baseline fleet (id.). The agencies also stated
that a copy of the MY 2010 CAFE certification data would be put in the public docket if it
became available during the comment period. The MY 2010 data was reported by the
manufacturers throughout calendar year 2011 as the final sales figures were compiled and
submitted to the EPA database. Due to the lateness of the CAFE data submissions'3, it was not
possible to submit the new 2010 data into the docket during the public comment period. As
explained below, however, consistent with the agencies' expectations at proposal, and with
the agencies' standard practice of updating relevant information as practicable between
proposals and final rules, the agencies are using these data in one of the two fleet-based
projections we are using to estimate the impacts of the final rules.
For analysis supporting the NPRM, the agencies developed a forecast of the light
vehicle market through MY 2025 based on (a) the vehicle models in the MY 2008 CAFE
certification data, (b) the AEO2011 interim projection of future fleet sales volumes, and (c)
the future fleet forecast conducted by CSM in 2009. In the proposal, the agencies stated we
planned to use MY 2010 CAFE certification data, if available, for analysis supporting the
final rule (Joint Draft TSD, p. 1-2). The agencies also indicated our intention to, for analysis
a 2008 based fleet projection is a new term that is the same as the reference fleet. The term is added to clarify
when we are using the 2008 baseline and reference fleet vs. the 2010 baseline and reference fleet.
b Partly due to the earthquake and tsunami in Japan and the significant impact this had on their facilities, some
manufacturers requested and were granted an extension on the deadline to submit their CAFE data.
1-2
-------
The Baseline and Reference Vehicle Fleets
supporting the final rule, use the most recent version of EIA's AEO, and a market forecast
updated relative to that purchased from CSM (Joint Draft TSD section 1.3.5).
For this final rulemaking, the agencies have analyzed the costs and benefits of the
standards using two different forecasts of the light vehicle fleet through MY 2025. The
agencies have concluded that the significant uncertainty associated with forecasting sales
volumes, vehicle technologies, fuel prices, consumer demand, and so forth out to MY 2025,
makes it reasonable and appropriate to evaluate the impacts of the final CAFE and GHG
standards using two baselines. One market forecast, similar to the one used for the NPRM,
uses corrected data regarding the MY2008 fleet, information from AEO 2011, and
information purchased from CSM. The agencies received comments regarding the market
forecast used in the NPRM suggesting that updates in several respects could be helpful to the
agencies' analysis of final standards; given those comments and since the agencies were
already planning to produce an updated market forecast, the final rule also contains another
market forecast using MY 2010 CAFE certification data, information from AEO 2012, and
information purchased from LMC Automotive (formerly JD Power Forecasting).
The two market forecasts contain certain differences, although as will be discussed
below, the differences are not significant enough to change the agencies' decision as to the
structure and stringency of the final standards. For example, MY 2008 certification data
represents the most recent model year for which the industry's offerings were not strongly
affected by the subsequent economic recession, which may make it reasonable to use if we
believe that the future vehicle model offerings are more likely to be reflective of pre-recession
offerings than models produced after MY 2008 (e.g., in MY 2010). Also, the MY 2010-based
fleet projection employs a future fleet forecast provided by LMC Automotive, which is more
current than the projection provided by CSM in 2009. However, the CSM forecast, utilized
for the MY2008-based fleet projection, was influenced by the recession, particularly in
predicting major declines in market share for some manufacturers (e.g., Chrysler) which the
agencies do not believe are reasonably reflective of future trends.
The MY 2010 based fleet projection, which is used in EPA's alternative analysis and
in NHTSA's co-analysis, employs a future fleet forecast provided by LMC Automotive,
which is more current than the projection provided by CSM in 2009, and which reflects the
post-proposal MY 2010 CAFE certification data. However, this MY 2010 CAFE data also
shows strong effects of the economic recession. For example, industry-wide sales were down
by 20% compared to pre-recession MY 2008 levels. For some companies like Chrysler,
Mitsubishi, and Subaru, sales were down by 30-40% from MY 2008 levels.0 For BMW,
General Motors, Jaguar/Land Rover, Porsche, and Suzuki, sales were down more than 40%
from MY 2008 levels. Employing the MY 2008 vehicle data avoids using these baseline
! These figure are arrived at using Table 1-17 and Table 1-39.
1-3
-------
The Baseline and Reference Vehicle Fleets
market shifts when projecting the future fleet. On the other hand, it also perpetuates vehicle
brands and models (and thus, their outdated fuel economy levels and engineering
characteristics) that have since been discontinued. The MY 2010 CAFE certification data
accounts for the phase-out of some brands (e.g., Saab, Pontiac, Hummer)6 and the
introduction of some technologies (e.g., Ford's Ecoboost engine), which may be more
reflective of the future fleet in this respect.
Thus, given the volume of information that goes into creating a baseline forecast and
given the significant uncertainty in any projection out to MY 2025, the agencies think that a
reasonable way to illustrate the possible impacts of that uncertainty for purposes of this
rulemaking is the approach taken here of analyzing the effects of the final standards under
both the MY 2008-based baseline and the MY 2010-based baseline. The agencies' analyses
are presented in our respective RIAs and preamble sections.
1.3 The 2008 Based Fleet Projection
Differences between the 2008 MY based fleet used in the final rule compared to that
used in the NPRM include minor corrections to some of the vehicle footprint data, and minor
corrections to technology "overrides" and technology class assignments used in DOT's
modeling system. A discussion of the changes is in the section below along with a thorough
description of how the projection was created.
1.3.1 On what data is the MY2008 baseline vehicle fleet based?
As part of the CAFE program, EPA measures vehicle CC>2 emissions and converts
them to mpg, and generates and maintains the federal fuel economy database. See 49 U.S.C
32904 and 40 CFR Part 600. Most of the information about the vehicles that make up the
2008 fleet was gathered from EPA's emission certification and fuel economy database, most
of which is available to the public. These data (by individual vehicle model produced in MY
2008) include: vehicle production volume, fuel economy rating for CAFE certification (i.e.,
on the 2-cycle city-highway test), carbon dioxide emissions (equivalent to fuel economy
rating for CAFE certification), fuel type (gasoline, diesel, and/or alternative fuel), number of
engine cylinders, displacement, valves per cylinder, engine cycle, transmission type, drive
(rear-wheel, all-wheel, etc.), hybrid type (if applicable), and engine aspiration (naturally-
aspirated, turbocharged, etc.). In addition to this information about each vehicle model
produced in MY 2008, the agencies need additional information about the fuel economy-
improving/CO2-reducing technologies already on those vehicle models in order to assess how
much and which technologies to apply to determine a path toward future compliance.
However, EPA's certification database does not include a detailed description of the types of
technologies considered in this FRM because this level of information was not reported in
e Based on our review of the CAFE certification data, the MY 2010-based fleet contains no Saabs, and compared
to the MY 2008-based fleet, about 90% fewer Hummers and about 75% fewer Pontiacs.
1-4
-------
The Baseline and Reference Vehicle Fleets
MY 2008 for emission certification or fuel economy testing. Thus, the agencies augmented
this description with publicly-available data which includes more complete technology
descriptions from Ward's Automotive Group.f'g The agencies also required information about
the footprints of MY 2008 vehicles in order to generate potential target footprint curves (as
discussed in Chapter 2 of the TSD). In a few instances when relevant vehicle information
(such as vehicle track width for footprint) was not available from these two sources, the
agencies obtained this information principally from publicly-accessible internet sites such as
Motortrend.com or Edmunds.com, and occasionally from other sources (such as articles about
specific vehicles revealed from internet search engine research). ''
Between the NPRM and the final rule, the agencies found discrepancies in footprint
values for a number of vehicles in the MY 2008 CAFE certification data. Specifically,
contractors to DOT employed to develop a market share model for incorporation into the
CAFE model noted that out of 1,302 vehicles in the MY 2008-based input file used in the
agencies' NPRM analysis, in 554 cases, the wheelbase value in the CAFE certification data
did not match wheelbase data from Ward's Automotive that the contractor had obtained
separately. While wheelbase is not a direct input to the models used in developing the
standards, it is a component of footprint, which is a key input in the modeling process.
Of the reported differences, 287 (51.8%) were less than or equal to 0.1 inch, and 115
(20.8%) were greater than 0.1 inch but less than or equal to 0.5 inch. The former set of
differences is most likely attributable to differences in the number of significant digits in the
reported raw data. The latter set of differences may also be due to reporting differences or
actual measurement differences, but would not have a significant impact on the computed
footprint value, all other things being equal. These differences were not considered further.
Of the remaining differences, 14 (2.5%) were greater than 0.5 inch but less than 1
inch. Most significantly, 138 (24.9%) of the differences were greater than 1 inch, ranging in
value from 1.1 inch to 23.8 inches.
To verify these findings, the Ward's data used by the contractor on wheelbase for the
152 vehicles with a discrepancy greater than 0.5 inches were compared to wheelbase data
from Edmunds, cars.com, Motor Trend, and product plans where available, and values
reflecting the agencies' best judgment about actual average values was selected.
Footprint for the 152 vehicles was thus recalculated based on corrected wheelbase. In
the process of validating the wheelbase data, the agencies noted that there were many
f WardsAuto.com: Used as a source for engine specifications shown in Table 1-1.
8 Note that WardsAuto.com, where this information was obtained, is a fee-based service, but all information is
public to subscribers.
Motortrend.com and Edmunds.com: Used as a source for footprint and vehicle weight data.
1 Motortrend.com and Edmunds.com are free, no-fee internet sites.
1-5
-------
The Baseline and Reference Vehicle Fleets
discrepancies in the track width values, which the agencies also corrected in the calculation of
the corrected footprints.
The affected vehicles included those of the following manufacturers:
Chrysler -4(2 large SUV, 2 small SUV)
Daimler -19(1 compact auto, 15 large auto, 1 midsize auto, 2 subcompact auto)
Ford - 4 (2 large pickup, 2 small pickup)
General Motors - 29 (18 compact auto, 7 midsize auto, 4 subcompact auto)
Honda - 17 (3 compact auto, 2 large SUV, 8 midsize auto, 1 small pickup, 3 subcompact auto)
Hyundai - 2 (2 subcompact auto)
Kia - 8 (2 compact auto, 4 midsize auto, 2 subcompact auto)
Mazda - 7 (4 midsize SUV, 2 small pickup, 1 subcompact auto)
Nissan - 11 (4 compact auto, 6 large auto, 1 minivan)
Subaru - 15 (6 midsize auto, 9 midsize SUV)
Tata - 2 (2 midsize auto)
Toyota - 29 (3 compact auto, 6 large pickup, 16 large auto, 4 midsize auto)
Volkswagen - 5 (4 large auto, 1 midsize auto)
Table 1-1 shows the change from the NPRM to the FRM in the average footprint for
all vehicles, cars, and trucks. The average change in footprint was very small, although quite
a few vehicles' footprints were updated.
Table 1-1 2008 MY Footprint changes (Final Rule Values - NPRM Values)
Average Footprint of all Vehicles
-0.1
Average Footprint Cars
-0.2
Average Footprint Trucks
0
The baseline vehicle fleet for the analysis informing these final rules is the same
except for the footprint changes to the baseline vehicle fleet used in the MYs 2012-2016
rulemaking, and like that baseline, is comprised of publicly-available data to the largest extent
possible. Some of the technology data included in the MYs 2012-2016 analysis' baseline
fleet was based on confidential product plan information about MY 2008 vehicles,
specifically, data about which vehicles already have low friction lubricants, electric power
steering, improved accessories, and low rolling resistance tires applied, the agencies no longer
consider that information as needing to be withheld, because by now all MY 2008 vehicle
models are already in the on-road fleet. As a result, the agencies are able to make public the
exact baseline used in this rulemaking analysis.
As explained in the MYs 2012-2016 TSD, creating the 2008 baseline fleet Excel file
was an extremely labor-intensive process. EPA in consultation with NHTSA first considered
using EPA's CAFE certification data, which contains most of the required information.
However, since the deadline for manufacturers to report this data did not allow enough time,
1-6
-------
The Baseline and Reference Vehicle Fleets
in the MYs 2012-2016 rulemaking, for early modeling review, the agencies began to create
the baseline fleet file using an alternative data source.
The agencies ultimately relied on a combination of EPA's vehicle emissions
certification data, data from a paid subscription to Ward's Automotive Group, and CAFE
certification data. EPA's vehicle emissions certification data contains much of the
information required for creating a baseline fleet file, but it lacked the production volumes
that are necessary for the OMEGA and Volpe models, and also contains some vehicle models
that manufacturers certified but did not produce in MY 2008. The data from Ward's
contained production volumes (which were not ultimately used, because they did not have
volumes for individual vehicles down to the resolution of the specific engine and transmission
level) and vehicle specifications, and eliminated extraneous vehicles.
The EPA vehicle emissions certification dataset came in two parts, an engine file and a
vehicle file, which the agencies combined into one spreadsheet using their common index.
The more-specific Ward's data also came in two parts, an engine file and a vehicle file, and
also required mapping, which was more difficult than combining the EPA vehicle emissions
certification dataset files because there was no common index between the Ward's files. A
new index was implanted in the engine file and a search equation in the vehicle file, which
identified most of the vehicle and engine combinations. Each vehicle and engine combination
was reviewed and corrections were made manually when the search routine failed to give the
correct engine and vehicle combination. The combined Ward's data was then mapped to the
EPA vehicle emissions certification data by creating a new index in the combined Ward's
data and using the same process that was used to combine the Ward's engine and vehicle files.
In the next step, CAFE certification data had to be merged in order to fill out the
needed production volumes. NHTSA and EPA reviewed the CAFE certification data for MY
2008 as it became available in the MYs 2012-2016 rulemaking. The CAFE certification set
could have been used with the Ward's data without the EPA vehicle emission certification
data set, but was instead appended to the combined Ward's and EPA vehicle emission
certification dataset. That combined dataset was then mapped into the CAFE dataset using
the same Excel mapping technique described above. Finally EPA and NHTSA obtained the
remaining attribute and technology data, such as footprint, curb weight, and others (for a
complete list of data with sources see Table 1-2 below) from other sources, thus completing
the baseline dataset.
Another step that was done for the first time in the NPRM (and used in this FRM
baseline as well) was to disaggregate the footprints of pickup trucks. In the MYs 2012-2016
rulemaking the agencies aggregated full-size pickup data in the baseline by using average
values to represent all variants of a given pickup line. While full-size pickups might be
offered with various combinations of cab style (e.g., regular, extended, crew) and box length
(e.g., 5 l/2\ 6 l/2 , 8'), and therefore multiple footprint sizes, CAFE compliance data for MY
2008 did not contain footprint information, and therefore could not reliably be used to identify
which pickup entries correspond to footprint values estimable from public or commercial
sources. Therefore, the agencies used the known production levels of average values to
1-7
-------
The Baseline and Reference Vehicle Fleets
represent all variants of a given pickup line (e.g., all variants of the F-150, or all variants of
the Sierra/Silverado) in order to calculate the sales-weighted average footprint/fuel economy
value for each pickup family. In retrospect, this may have affected how we fit the light truck
target curve, among other things, so the agencies have since created an expanded version of
the fleet to account for the variation in footprint/wheelbase for the large pickups of Chrysler,
Ford, GM, Nissan and Toyota. In MY 2008, large pickups were available from Nissan with 2,
Chrysler and Toyota with 3, and Ford and GM with 5 wheelbase/footprint combinations. The
agencies got this footprint data from MY 2008 product plans submitted by the various
manufacturers, which can be made public at this time because by now all MY 2008 vehicle
models are already in production, which makes footprint data about them essentially public
information.
The agencies created the expanded fleet by replicating original records from a single
pickup footprint model into multiple pickup models with distinct footprint values, in order to
reflect the additional pickup model footprints just noted. For example, an F-150 in the MY
2008 baseline used in the MYs 2012-2016 rulemaking analysis with a footprint value of 67
square feet, is disaggregated by replicating 2 times in all respects, except with footprint values
of 58, 67, and 73 square feet. Sales volumes of these pickups from the original record were
distributed to each of the "58 square feet" and "73 square feet" duplicates based on the
distribution of MY 2008 sales by these pickups' wheelbase/footprint, which the agencies took
from product plan data submitted by the manufacturers in 2008/2009 in response to requests
to support the MYs 2012-2016 rulemaking analysis. The agencies were able to distribute the
sales for each of the original pickups by wheelbase/footprint by matching each of the pickups
in the baseline fleet with pickups in the product plans on the basis of drive type, transmission
type, and engine displacement, cylinders/configuration and HP, and then sorting and summing
the sales of the matched pickups in the product plans by wheelbase/footprint.
Both agencies used this fleet forecast to populate input files for the agencies'
respective modeling systems. The structure of the market forecast input file used for the
Volpe model is described the model documentation.2 To help readers who wish to directly
examine the baseline fleet file for EPA's OMEGA model, and to provide some idea of its
contents for those readers who do not, Table 1-2 shows the columns of the complete fleet file,
which includes the MY 2008 baseline data that was compiled. Each column has its name,
definition (description) and source. Most elements shown in Table 1-2 also appear in the
market forecast input file for DOT's modeling system, which also accommodates some
additional data elements discussed in the model documentation.
Table 1-2 2008 MY Data, Definitions, and Sources
Data Item
Index
Manufacturer
CERT
Manufacturer Name
Name Plate
Definition
Index Used to link EPA and NHTSA baselines
Common name of company that manufactured
vehicle. May include more name plates than
Cert Manufacturer Name.
Certification name of company that
manufactured vehicle
Name of Division
Where The Data is From
Created
Certification data
Certification data
Certification data
1-8
-------
The Baseline and Reference Vehicle Fleets
Model
Reg Class
Our Class
CSM Class
Vehicle Type
Number
Vehicle Index From
Sum Page
Traditional
Car/Truck
NHTSA Defined
New Car/Truck
Total Production
Volume
Fuel Econ.
(mpg)
CO2
Area (sf)
Fuel
Fuel Type
Disp
(lit)
Effective Cyl
Actual Cylinders
Valves Per Cylinder
Valve Type
Valve Actuation
VVT
VVLT
Deac
Fuel injection
system
Boost
Engine Cycle
Name of Vehicle
EPA Fuel Economy Class Name
If a car's Footprint<43 then "SubCmpctAuto"
If a car's 43<=Footprint<46 then
"CompactAuto"
If a car's 46<=Footprint<53 then
"MidSizeAuto"
If a car's Footprint >=53 then "LargeAuto"
If a S.U.V.'s Footprint < 43 then "SmallSuv"
If a S.U.V.'s 43<=Footprint<46 then
"MidSizeSuv"
If a S.U. V's Footprint >=46 then "LargeSuv"
If a Truck's Footprint < 50 then "SmallPickup"
If a Truck's Footprint>=50 then "LargPickup"
If a Van's Structure is Ladder then
"CargoVan"
If a Van's Structure is Unibody then
"Minivan"
CSM Worldwide' s class for the vehicle. Used
to weight vehicles based on CSM data.
Vehicle Type Number assigned to a vehicle
based on its number of cylinders, valves per
cylinder, and valve actuation technology
Number to be used as a cross reference with
the Sum Pages.
Traditional Car Truck value for reference.
New NHTSA Car Truck value as defined in
201 1 Fuel economy regulations. Used in
calculations.
Total number of vehicles produced for that
model.
EPA Unadjusted Fuel Economy
CO2 calculated from MPG. CO2 weighted
1.15 times higher for diesel vehicles.
Average Track x Wheelbase
Gas or Diesel
Gas or Diesel or Electric
Engine Cylinder Displacement Size in Liters
Number of Cylinder + 2 if the engine has a
turbo or super charger.
Actual Number of Engine Cylinders
Number of Valves Per Actual Cylinder
Type of valve actuation.
Type of valve actuation with values compatible
with the package file.
Type of valve timing with values compatible
with the package file.
Type of valve lift with values compatible with
the package file.
Cylinder Deactivation with a value that is
compatible with the package file.
Type of fuel injection.
Type of Boost if any.
As Defined by EPA Cert. Definition
Certification data
Certification data
Derived From Certification data and Footprint
CSM Worldwide
Defined by EPA staff
NA
Certification data
NHTSA
Certification data
Certification data
Certification data
Calculated from track width and wheel base
Wards
Certification data
Wards/Certification data
Derived From Certification data.
Certification data
Certification data
Wards (Note: Type E is from Cert Data)
Wards
Wards
Wards
Wards
Wards
Wards
Wards
1-9
-------
The Baseline and Reference Vehicle Fleets
Horsepower
Torque
Trans Type
Trans
Num of Gears
Transmission
Structure
Drive
Drive with AWD
Wheelbase
Track Width
(front)
Track Width
(rear)
Footprint: PU
Average
Threshold Footprint
Curb
Weight
GVWR
Stop-
Start/Hybrid/Full
EV
Import Car
Towing Capacity
(Maximum)
Engine Oil
Viscosity
Volume 2009
Volume 2010
Volume 20 11
Volume 20 12
Volume 20 13
Max. Horsepower of the Engine
Max. Torque of the Engine
A=Auto AMT= Automated Manual M=Manual
CVT= Continuously Variable Transmission
Type Code with number of Gears
Number of Gears
Transmission definition. Matches the cost
definition.
Ladder or Unibody
Fwd, Rwd, 4wd
Fwd, Rwd, Awd, 4wd
Length of Wheelbase
Length of Track Width in inches
Length of Track Width in inches
Car and Large Truck Footprints are normal
(Average Track x Wheelbase). Medium and
Small Truck footprints are the production
weighted average for each vehicle.
Footprint valve that will be set to 41 for values
less than 41 , Will be set to 56 for car values >
56, and will be set to 74 for truck values >74
Curb Weight of the Vehicle
Gross Vehicle Weight Rating of the Vehicle
Type of Electrification if any. Blank = None
Cars Imported
Weight a vehicle is rated to tow.
Ratio between the applied shear stress and the
rate of shear, which measures the resistance of
flow of the engine oil (as per SAE Glossary of
Automotive Terms)
Projected Production Volume for 2009
Projected Production Volume for 2010
Projected Production Volume for 201 1
Projected Production Volume for 2012
Projected Production Volume for 2013
Wards
Wards
Certification data
Certification data
Certification data
Certification data
General Internet Searches
Certification data
Certification data
Some from Edmunds.com or Motortrend.com,
Others from product plans with a subset verified
with Edmunds.com or Motortrend.com for
accuracy.
Some from Edmunds.com or Motortrend.com,
Others from product plans with a subset verified
with Edmunds.com or Motortrend.com for
accuracy.
Some from Edmunds.com or Motortrend.com,
Others from product plans with a subset verified
with Edmunds.com or Motortrend.com for
accuracy.
Derived from data from Edmunds.com or
Motortrend.com. Production volumes or specific
footprints from product plans.
Derived from data from Edmunds.com or
Motortrend.com. Production volumes or specific
footprints from product plans.
Some from Edmunds.com or Motortrend.com,
Others from product plans with a subset verified
with Edmunds.com or Motortrend.com for
accuracy.
Some from Edmunds.com or Motortrend.com,
Others from product plans with a subset verified
with Edmunds.com or Motortrend.com for
accuracy.
Certification data
Certification data
Volpe Input File
Volpe Input File
Calculated based on 2008 volume and Annual
Energy Outlook and CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
1-10
-------
The Baseline and Reference Vehicle Fleets
Volume 20 14
Volume 20 15
Volume 20 16
Volume 20 17
Volume 20 18
Volume 20 19
Volume 2020
Volume 2021
Volume 2022
Volume 2023
Volume 2024
Volume 2025
Low drag brakes
Electric Power
steering
Volpe Index
Projected Production Volume for 2014
Projected Production Volume for 2015
Projected Production Volume for 2016
Projected Production Volume for 2017
Projected Production Volume for 2018
Projected Production Volume for 2019
Projected Production Volume for 2020
Projected Production Volume for 2021
Projected Production Volume for 2022
Projected Production Volume for 2023
Projected Production Volume for 2024
Projected Production Volume for 2025
See Volpe Documentation
See Volpe Documentation
Number used to reorder the vehicles in the
EPA baseline in the same order as the Volpe
input file.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Calculated based on 2008 volume and AEO and
CSM adjustment factors.
Volpe Input File
Volpe Input File
Volpe Input File
Notes:
1. For engines not available in the WardsAuto.com
engine file, an internet search was done to find this
information.
2. These data were obtained from manufacturer's product
plans. They were used to block (where possible) the
model from adding technology that was already on a
vehicle.
3. Ward's Automotive Group data obtained from "2008
Light Vehicle Engines."
DOT's CAFE model also uses a series of inputs—referred to as "overrides"—to
specify baseline technology content of specific vehicle models (and specific engines and
transmissions) and to indicate cases where specific technologies are not applicable to specific
vehicle models. In the MY 2008-based market forecast, DOT has corrected some of these
settings to indicate that micro-hybrid technology (or more advanced hybrid) is already present
on hybrid versions of the Altima, Aura, Civic, Camry, Escape, Highlander, Lexus GS and LS,
Lexus RX, Mariner, Malibu, Prius, Tahoe, Tribute, Vue, and Yukon. The CAFE model also
uses inputs to assign vehicles to specific "technology classes," where technology-related
inputs define the applicability, efficacy, and cost of each technology for vehicles in each
technology class. In the MY 2008-based market forecast, DOT has reassigned the Altima
1-11
-------
The Baseline and Reference Vehicle Fleets
(coupe), Audi A4, Corolla, Impala, Matrix, Passat, and Jetta to technology classes that better
represent these vehicles' size and performance characteristics.
The sales volumes for the MY 2008 baseline fleet are included in the section below on
reference fleet under the MY 2008 columns. Table 1-3 displays the engine technologies
present in the baseline fleet. Again, the engine technologies for the vehicles manufactured by
these manufacturers in MY 2008 were largely obtained from Ward's Auto online.
Table 1-3 2008 Engine Technology Percentages
Manufacturer
All
All
All
Aston Martin
Aston Martin
BMW
BMW
Chrysler/Fiat
Chrysler/Fiat
Daimler
Daimler
Ferrari
Ferrari
Ford
Ford
Geely/Volvo
Geely/Volvo
GM
GM
Honda
Honda
Hyundai
Hyundai
Kia
Kia
Lotus
Lotus
Mazda
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
T3
1)
I
o
1
£
3%
4%
1%
0%
0%
33%
5%
1%
0%
2%
16%
0%
0%
0%
0%
0%
49%
0%
1%
0%
4%
0%
0%
0%
0%
0%
0%
11%
Super Charged
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
1%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
77%
0%
0%
Single Overhead Cam
20%
17%
24%
0%
0%
14%
0%
21%
39%
55%
36%
0%
0%
15%
65%
0%
0%
0%
0%
57%
64%
0%
0%
0%
0%
0%
0%
0%
Dual Overhead Cam
63%
73%
48%
100%
0%
86%
100%
72%
4%
45%
64%
100%
0%
85%
32%
100%
100%
31%
56%
43%
36%
100%
100%
100%
100%
100%
0%
99%
Overhead Cam
17%
9%
29%
0%
0%
0%
0%
8%
57%
0%
0%
0%
0%
0%
3%
0%
0%
69%
44%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Variable Valve Timing
Continuous
8%
9%
6%
0%
0%
14%
0%
0%
0%
72%
35%
0%
0%
4%
28%
0%
0%
5%
29%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Variable Valve Timing
Discrete
22%
24%
19%
100%
0%
86%
100%
42%
4%
4%
17%
100%
0%
0%
1%
100%
100%
17%
31%
27%
4%
0%
0%
0%
0%
100%
0%
7%
Variable Valve Timing
Intake Only
30%
35%
23%
0%
0%
0%
0%
0%
0%
13%
47%
0%
0%
47%
9%
0%
0%
14%
1%
20%
28%
100%
100%
10%
17%
0%
0%
92%
Variable Valve Lift and
Timing Continuous
0%
0%
0%
24%
0%
0%
0%
0%
0%
0%
0%
29%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Variable Valve Lift and
Timing Discrete
12%
13%
10%
0%
0%
13%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
100%
0%
0%
0%
0%
0%
0%
0%
Cylinder Deactivation
6%
3%
11%
0%
0%
0%
0%
5%
4%
0%
0%
0%
0%
0%
0%
0%
0%
40%
4%
11%
0%
0%
0%
0%
0%
0%
0%
0%
a
.0
o
.°
Q
5%
7%
3%
0%
0%
33%
6%
0%
0%
2%
16%
0%
0%
0%
0%
0%
0%
0%
6%
0%
4%
0%
0%
0%
0%
0%
0%
11%
1-12
-------
The Baseline and Reference Vehicle Fleets
Mazda
Mitsubishi
Mitsubishi
Nissan
Nissan
Porsche
Porsche
Spyker/Saab
Spyker/Saab
Subaru
Subaru
Suzuki
Suzuki
Tata/JLR
Tata/JLR
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
24%
6%
0%
0%
0%
17%
12%
100%
0%
15%
3%
0%
0%
0%
0%
0%
0%
0%
0%
43%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
20%
0%
0%
0%
0%
0%
0%
1%
100%
100%
0%
0%
0%
0%
0%
0%
69%
70%
0%
0%
0%
0%
0%
0%
0%
0%
85%
0%
99%
0%
0%
100%
100%
100%
100%
100%
62%
31%
30%
100%
100%
100%
100%
0%
0%
100%
100%
15%
100%
0%
0%
0%
0%
0%
0%
0%
0%
38%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
38%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
13%
0%
0%
4%
0%
100%
100%
17%
0%
0%
23%
0%
0%
76%
0%
0%
0%
29%
61%
48%
99%
87%
0%
0%
96%
100%
0%
0%
0%
62%
31%
7%
0%
0%
24%
100%
0%
0%
71%
39%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
100%
0%
0%
1%
27%
0%
0%
0%
0%
0%
0%
0%
0%
1%
79%
0%
0%
0%
0%
0%
0%
0%
0%
28%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
24%
0%
0%
0%
0%
17%
100%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
8%
6%
89%
100%
The data in Table 1-3 indicates that manufacturers had already begun implementing a
number of fuel economy/GHG reduction technologies in the baseline (2008) fleet. For
example, VW stands out as having a significant number of turbocharged direct injection
engines, though it is uncertain whether their engines are also downsized. Some of the valve
and cam technologies are quite common in the baseline fleet: for example, nearly half the
baseline fleet already has dual cam phasing, while Honda and GM have considerable levels of
engines with cylinder deactivation. Honda also has already implemented continuously
variable valve lift on a majority of their engines. Part of the implication of these technologies
already being present in the baseline is that if manufacturers have already implemented them,
they are therefore not available in the rulemaking analysis for improving fuel economy and
reducing CC>2 emissions further, requiring the agencies to look toward increasing penetration
of these and other technologies and increasingly advanced technologies to project continued
improvements in stringency over time.
The section below provides further detail on the conversion of the MY 2008 baseline
into the MYs 2017-2025 reference fleet. It also describes more of the data contained in the
baseline spreadsheet.
1.3.2
The MY 2008 Based MY 2017-2025 Reference Fleet
The reference fleet aims to reflect the current market conditions and expectations
about conditions of the vehicle fleet during the model years to which the agencies' rules
1-13
-------
The Baseline and Reference Vehicle Fleets
apply. Fundamentally, constructing this fleet involved projecting the MY 2008 baseline fleet
into the MYs 2017-2025 model years. It also included the assumption that none of the vehicle
models had changes during this period. Projecting this future fleet is a process that is
necessarily uncertain. NHTSA and EPA therefore relied on many sources of reputable
information to make these projections.
1.3.2.1 On what data is the reference vehicle fleet based (using the 2008 baseline)?
For the MY 2008-based reference fleet, EPA and NHTSA have based the projection of
total car and light truck sales on the 2011 projections made by the Energy Information
Administration (EIA). EIA publishes a projection of national energy use annually called the
Annual Energy Outlook (AEO).3 EIA issued an early release version of AEO2012 January
2012. The agencies are continuing to use this AEO data for the MY 2008 baseline consistent
with the NPRM. EPA and NHTSA are employing the newer version of AEO in projecting the
reference fleet for the 2010 MY based baseline and reference fleet projection as discussed in
section 1.4.2.1.
As in the NPRM, the agencies used the Energy Information Administration's (EIA's)
National Energy Modeling System (NEMS) to estimate the future relative market shares of
passenger cars and light trucks. However, NEMS shifts the market toward passenger cars in
order to ensure compliance with EISA's requirement that CAFE standards cause the fleet to
achieve 35 mpg by 2020. Because we use our market projection as a baseline relative to
which we measure the effects of new standards, and we attempt to estimate the industry's
ability to comply with new standards without changing product mix (i.e., we analyze the
effects of the final rules assuming manufacturers will not change fleet composition as a
compliance strategy), using AEO 2011-projected shift in passenger car market share as
provided by EIA would cause the agencies to understate the cost of achieving compliance
through additional technology alone. Therefore, for analyses supporting today's final rule, the
agencies developed a new projection of passenger car and light truck sales shares by using
NEMS to run scenarios from the AEO 2011 reference case, after first deactivating the above-
mentioned sales-volume shifting methodology and holding post-2017 CAFE standards
constant at MY 2016 levels. Incorporating these changes reduced the projected passenger car
share of the light vehicle market by an average of about 5% during 2017-2025. This case is
referred to as the "Unforced Reference Case," and the values are shown below in Table 1-5.
Table 1-4 AEO 2011 Reference Case Volumes
Model Year
2017
2018
2019
2020
2021
2022
Cars
8,984,200
8,998,200
9,170,900
9,553,600
9,801,100
10,056,600
Trucks
6,812,000
6,552,200
6,391,300
6,336,200
6,380,000
6,384,600
Total Vehicles
15,796,100
15,550,400
15,562,200
15,889,800
16,181,100
16,441,200
1-14
-------
The Baseline and Reference Vehicle Fleets
2023
2024
2025
10,244,500
10,483,400
10,739,600
6,396,500
6,407,700
6,470,200
16,641,000
16,891,100
17,209,800
Table 1-5 AEO 2011 Interim Unforced Reference Case Volumes
Model Year
2017
2018
2019
2020
2021
2022
2023
2024
2025
Cars
8,440,703
8,376,192
8,464,457
8,725,709
8,911,173
9,123,436
9,344,051
9,580,693
9,836,330
Trucks
7,365,619
7,200,218
7,114,201
7,170,230
7,277,894
7,316,337
7,311,438
7,353,394
7,414,129
Total Vehicles
15,806,322
15,576,410
15,578,658
15,895,939
16,189,066
16,439,772
16,655,489
16,934,087
17,250,459
In this 2017 projection, car and light truck sales are expected to get up to 8.4 and 7.4
million units, respectively. While the total level of sales of 15.8 million units is similar to
pre-2008 levels, the fraction of car sales in 2017 and beyond is projected to be higher than in
the 2000-2007 time frame. Note that EIA's definition of cars and trucks follows that used by
NHTSA prior to the MY 2011 CAFE final rule. The MY 2011 CAFE final rule reclassified
approximately 1 million 2-wheel drive sport utility vehicles from the truck fleet to the car
fleet. EIA's sales projections of cars and trucks for the 2017-2025 model years under the old
NHTSA truck definition are shown above in Table 1-4 and Table 1-5.
In addition to a shift towards more car sales, sales of segments within both the car and
truck markets have also been changing and are expected to continue to change in the future.
Manufacturers are continuing to introduce more crossover models which offer much of the
utility of SUVs but use more car-like designs and unibody structures. In order to reflect these
changes in fleet makeup, EPA and NHTSA used a custom long range forecast purchased from
CSM Worldwide (CSM). CSMjk is a well-known industry analyst that provided the forecast
used by the agencies for the 2012-2016 final rule. NHTSA and EPA decided to use the
forecast from CSM in the MY 2008 baseline reference fleet for several reasons. One, CSM
J CSM World Wide is a paid service provider.
k As with any long range forecast, CSM World Wide's forecast out to 2025 has uncertainties since many
manufacturers do not have full future product plans out that far.
1-15
-------
The Baseline and Reference Vehicle Fleets
uses a ground up approach (e.g., looking at the number of plants and capacity for specific
engines, transmissions, and vehicles) for their forecast, which the agencies believe is a robust
forecasting approach1. Two, CSM agreed to allow us to publish their high level data, on
which the forecast is based, in the public domain. Three, the CSM forecast covered all the
timeframe of greatest relevance to this analysis (2017-2025 model years). Four, it provided
projections of vehicle sales both by manufacturer and by market segment. And five, it
utilized market segments similar to those used in the EPA emission certification program and
fuel economy guide, such that the agencies could include only the vehicle types covered by
the final standards. The agencies note that CSM developed the forecast during a period when
the United States economy was undergoing significant stress and some automobile
manufacturers were experiencing a high degree of financial uncertainty. In the time since
CSM developed its forecast,industry sales and in particular the sales for some individual
manufacturers have turned out differently than in the CSM forecast. Because forecasting the
market out to MY 2025 has uncertainties, the agencies believe there are benefits from using
the CSM forecast for one of the two analyses cases to reflect some level of uncertainty in the
final rule analysis. It is feasible that the CSM forecast could represent what might happen in
the future.
CSM created a forecast that covered model years 2017-2025. Since the agencies used
this forecast to generate the reference fleet (i.e., the fleet expected to be sold absent any
increases in the stringency regulations after the 2016 model year), it is important for the
forecast not to reflect changes in fleet composition during 2017-2025 attributed to CAFE/
GHG standards. However, CSM assumed that CAFE and GHG standards would continue to
increase in stringency after 2016, although CSM did not use specific future standards as
quantitative inputs to its model. In its quantitative analysis, CSM used fuel price, industry
demand, consumer demand and other economic factors to project the composition of the
future fleet. In response to question by the agencies, CSM indicated that their assumption of
future standards had a negligible (non-discernible) impact on their forecast since it was not a
direct quantitative input to the model such that CSM's forecast would have been essentially
the same had CSM assumed no stringency increases after 2016.
The agencies combined the CSM forecast with data from other sources to create the
reference fleet projections. This process is discussed in sections that follow.
1 There are other forecasting groups that do similar projections and meet all these criteria. LMC Automotive
(formerly JD Power Forecasting) is another, and this was used for the alternate reference case projection as
described below.
1-16
-------
The Baseline and Reference Vehicle Fleets
1.3.2.2 How do the agencies develop the reference vehicle fleet?
The process of producing the 2017-2025 reference fleet involved combining the
baseline fleet with the projection data described above. This was a complex multistep
procedure, which is described in this section.
1.3.2.3 How was the 2008 baseline data merged with the CSM data?
For the NPRM, EPA and NHTSA employed the same methodology as in the 2012-16
rule for mapping certification vehicles to CSM vehicles; the results were used again for
analysis supporting today's final rule. Merging the 2008 baseline data with the 2017-2025
CSM data required a thorough mapping of certification vehicles to CSM vehicles by
individual make and model. One challenge that the agencies faced when determining a
reference case fleet was that the sales data projected by CSM had different market
segmentation than the data contained in EPA's database. In order to create a common
segmentation between the two databases, the agencies performed a side-by-side comparison
of each vehicle model in both datasets, and created an additional "CSM segment" modifier in
the spreadsheet to map the two datasets. The reference fleet sales based on the "CSM
segmentation" 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.4 The spreadsheet provides
specific details on the sources and definitions for the data. The Excel file contains several
tabs. They are: "Data", "Data Tech Definitions", "SUM", "SUM Tech Definitions", "Truck
Vehicle Type Map", and "Car Vehicle Type Map". "Data" is the tab with the raw data.
"Data Tech Definitions" is the tab where each column is defined and its data source named.
"SUM" is the tab where the raw data is processed to be used in the OMEGA and Volpe
models. The "SUM" tab minus columns A-F and minus the Generic vehicles is the input file
for the models. The "Generic" manufacturer (shown in the "SUM" tab) is the sum of all
manufacturers and is calculated as a reference, and for data verification purposes. It is used to
validate the manufacturers' totals. It also gives an overview of the fleet.
Table 1-6 shows the sum of the models chosen. The number of models is determined
by the number of unique segment and vehicle type combinations. These combinations of
segment and vehicle type (the vehicle type number is the same as the technology package
number) are determined by the technology packages discussed in the EPA RIA. "SUM Tech
Definitions" is the tab where the columns of the "SUM" tab are defined. The "Truck Vehicle
Type Map" and "Car Vehicle Type Map" map the number of cylinder and valve actuation
technology to the "tech package" vehicle type number.
Table 1-6 Models from the SUM Tab Model
Model
Car Like LargeSuv >=V8 Vehicle Type: 13
Car Like LargeSuv V6 Vehicle Type: 16
Car Like LargeSuv V6 Vehicle Type: 12
1-17
-------
The Baseline and Reference Vehicle Fleets
Car Like LargeSuv V6 Vehicle Type: 9
Car Like LargeSuv 14 and 15 Vehicle Type: 7
Car Like MidSizeSuv V6 Vehicle Type: 8
Car Like MidSizeSuv V6 Vehicle Type: 5
Car Like MidSizeSuv 14 Vehicle Type: 7
Car Like SmallSuv V6 Vehicle Type: 12
Car Like SmallSuv V6 Vehicle Type: 4
Car Like SmallSuv 14 Vehicle Type: 3
LargeAuto >=V8 Vehicle Type: 13
LargeAuto >=V8 Vehicle Type: 10
LargeAuto >=V8 Vehicle Type: 6
LargeAuto V6 Vehicle Type: 12
LargeAuto V6 Vehicle Type: 5
MidSizeAuto >=V8 Vehicle Type: 13
MidSizeAuto >=V8 Vehicle Type: 10
MidSizeAuto >=V8 (7 or >) Vehicle Type: 6
MidSizeAuto V6 Vehicle Type: 12
MidSizeAuto V6 Vehicle Type: 8
MidSizeAuto V6 Vehicle Type: 5
MidSizeAuto 14 Vehicle Type: 3
In the combined EPA certification and CSM database, all 2008 vehicle models were
assumed to continue out to 2025, though their volumes changed in proportion to CSM
projections. Also, any new models expected to be introduced within the 2009-2025
timeframe are not included in the data. These volumes are reassigned to the existing models
to keep the overall fleet volume the same. All MYs 2017-2025 vehicles are mapped to the
existing vehicles by a process of mapping to manufacturer market share and overall segment
distribution. The mappings are discussed in the next section. Further discussion of this
limitation is discussed below in section 1.3.2.4. The statistics of this fleet will be presented
below since further modifications were required to the volumes as the next section describes.
1.3.2.4 How were the CSM forecasts normalized to the AEO forecasts for the 2008-
based fleet?
The next step in the agencies' generation of the reference fleet is one of the more
complicated steps to explain. Here, the projected CSM forecasts for relative sales of cars and
trucks by manufacturer and by market segment was normalized (set equal) to the total sales
estimates of the Early Release of the 2011 Annual Energy Outlook (AEO). NHTSA and EPA
used projected car and truck volumes for this period from Early AEO 2011. However, the
AEO projects sales only at the car and truck level, not at the manufacturer and model-specific
level, and the agencies' analysis requires this further level of detail. The CSM data provided
year-by-year percentages of cars and trucks sold by each manufacturer as well as the
percentages of each vehicle segment. Using these percentages normalized to the AEO-
projected volumes then provided the manufacturer-specific market share and model-specific
sales for model years 2017-2025 (it is worth clarifying that the agencies are not using the
1-18
-------
The Baseline and Reference Vehicle Fleets
model-specific sales volumes from CSM, only the higher-level volumes by manufacturer and
segment). This process is described in greater detail in the following paragraphs.
In order to determine future production volumes, the agencies developed multipliers
by manufacturer and vehicle segment that could be applied to MY 2008 volumes. The
process for developing the multipliers is complicated, but is easiest to explain as a three-step
process, though the first step is combined with both the second and third step, so only one
multiplier per manufacturer and vehicle segment is developed.
The three steps are:
1. Adjust total car and truck sales to match AEO projections.
2. Adjust car sales to match CSM market share projections for each manufacturer
and car segment.
3. Adjust truck sales to match CSM market share projections for each
manufacturer and truck segment.
The first step is the adjustment of total car and truck sales in 2008 to match AEO
projections of total car and truck sales in 2017-2025. The volumes for all of the trucks in
2008 were added up (TruckSum2008), and so were the volumes of all the cars (CarSum2008).
A multiplier was developed to scale the volumes in 2008 to the AEO projections. The
example equation below shows the general form of how to calculate a car or truck multiplier.
The AEO projections are shown above in Table 1-4.
Example Equation :
TruckMultiplier(Year X) = AEOProjectionforTrucks(Year X) / TruckSum2008
CarMultiplier(Year X) = AEOProjectionforCars(Year X) / CarSum2008
Where: Year X is the model year of the multiplier.
The AEO projection is different for each model year. Therefore, the multipliers are
different for each model year. The multipliers can be applied to each 2008 vehicle as a first
adjustment, but multipliers based solely on AEO have limited value since those multipliers
can only give an adjustment that will give the correct total numbers of cars and trucks without
the correct market share or vehicle mix. A correction factor based on the CSM data, which
does contain market share and vehicle segment mix, is therefore necessary, so combining the
AEO multiplier with CSM multipliers (one per manufacturer, segment, and model year) will
give the best multipliers.
There were several steps in developing an adjustment for Cars based on the CSM data.
CSM provided data on the market share and vehicle segment distribution. The first step in
determining the adjustment for Cars was to total the number of Cars in each vehicle segment
by manufacturer in MY 2008. A total for all manufacturers in each segment was also
calculated. The next step was to multiply the volume of each segment for each manufacturer
1-19
-------
The Baseline and Reference Vehicle Fleets
by the CSM market share. The AEO multiplier was also applied at this time. This gave
projected volumes with AEO total volumes and market share correction for Cars. This is
shown in the "Adjusted for 2017AEO and Manufacturer Market Share" column of Table 1-7.
The next step is to adjust the sales volumes for CSM vehicle segment distribution.
The process for adjusting for vehicle segment is more complicated than a simple one step
multiplication. In order to keep manufacturers' volumes constant and still have the correct
vehicle segment distribution, vehicles need to move from segment to segment while
maintaining constant manufacturers' totals. Six rules and one assumption were applied to
accomplish the shift. The assumption (based on the shift in vehicle sales in 2008 and 2009) is
that people are moving to smaller vehicles in the rulemaking time frame independently of
regulatory requirements. A higher-level (less detailed) example of this procedure is provided
in Section II of the preamble.
Vehicles from CSM's "Luxury Car," "Specialty Car," and "Other Car" segments, if
reduced, will be equally distributed to the remaining four categories ("Full-Size Car," "Mid-
Size Car," "Small Car," "Mini Car"). If these sales increased, they were taken from the
remaining four categories so that the relative sales in these four categories remained constant.
Vehicles from CSM's "Luxury Car," "Specialty Car," and "Other Car" segments, if
increased will take equally from the remaining categories ("Full-Size Car," "Mid-Size Car,"
"Small Car," "Mini Car").
All manufacturers have the same multiplier for a given segment shift based on moving
all vehicles in that segment to achieve the CSM distribution. Table 1-7 shows how the 2017
vehicles moved and the multipliers that were created for each adjustment. This does not mean
that new vehicle segments will be added (except for Generic Mini Car described in the next
step) to manufacturers that do not produce them. Vehicles within each manufacturer will be
shifted as close to the distribution as possible given the other rules. Table 1-8 has the
percentages of Cars per CSM segment. These percentages are multiplied by the total number
of vehicles in a given year to get the total sales in the segment. Table 1-7 shows the totals for
2017 in the "2017 AEO-CSM Sales Goal" column.
When "Full-Size Car," "Mid-Size Car," "Small Car" are processed, if vehicles need to
move in or out of the segment, they will move into or out of the next smaller segment. So, if
Mid-Size Cars are being processed they can only move to or be taken from Small Cars. Note:
In order to accomplish this, a "Generic Mini Car" segment was added to manufacturers who
did not have a Mini (type) Car in production in 2008, but needed to shift down vehicles from
the Small Car segment.
The data must be processed in the following order: "Luxury Car," "Specialty Car,"
"Other Car," "Full-Size Car," "Mid-Size Car," "Small Car." The "Mini Car" does not need to
be processed separately. By using this order, it works out that vehicles will always move
toward the correct distribution. There are two exceptions, BMW and Porsche only have
1-20
-------
The Baseline and Reference Vehicle Fleets
"Luxury Car," "Specialty Car," and "Other Car" vehicles, so their volumes were not changed
or shifted since these rules did not apply to them.
When an individual manufacturer multiplier is applied for a segment, the vehicles
move to or from the appropriate segments as specified in the previous rules and as shown in
Table 1-7.
Table 1-7 2017 Model Year Volume Shift*
CSM Segment
All Full-Size Car
All Luxury Car
All Mid-Size Car
All Mini Car
All Small Car
All Specialty Car
All Others
2008 MY
Sales
829,896
1,048,341
2,103,108
617,902
1,912,736
469,324
0
Adjusted for
2017AEOand
Manufacturer
Market Share
830,832
1,408,104
2,500,723
868,339
2,548,393
627,425
0
Luxury,
Specialty,
Other
Adjustment
818,226
1,423,691
2,475,267
851,234
2,513,350
702,048
0
Full Size
Adjustment
347,034
1,423,691
2,946,459
851,234
2,513,350
702,048
0
Midsize
Adjustment
347,034
1,423,691
2,431,715
851,234
3,028,094
702,048
0
Small Car
Adjustment
347,034
1,423,691
2,431,715
1,439,985
2,439,343
702,048
0
2017
AEO-
CSM
Sales Goal
347,034
1,423,691
2,431,715
1,439,985
2,439,343
702,048
0
Number Vehicles that shift and Where
All Full-Size Car
All Luxury Car
All Mid-Size Car
All Mini Car
All Small Car
All Specialty Car
All Others
(12,606)
15,587
(25,456)
(17,105)
(35,043)
74,623
0
(471,192)
0
471,192
0
0
0
0
0
0
(514,744)
0
514,744
0
0
0
0
0
588,751
(588,751)
0
0
Individual Manufacturer Multiplier
All Full-Size Car
All Luxury Car
All Mid-Size Car
All Mini Car
All Small Car
All Specialty Car
All Others
0.973
0.963
1
0.42
0.97
1.55
0.96
1-21
-------
The Baseline and Reference Vehicle Fleets
Table 1-8 CSM - Percent of Cars per Segment*
CSM Segment
Compact Car
Full-Size Car
Luxury Car
Mid-Size Car
Mini Car
Small Car
Specialty Car
Others
2017
0.00%
3.95%
16.70%
27.68%
15.33%
27.77%
8.56%
0.00%
2018
0.00%
3.56%
16.87%
27.77%
15.46%
27.57%
8.76%
0.00%
2019
0.00%
3.35%
17.14%
27.47%
15.45%
27.74%
8.84%
0.00%
2020
0.00%
4.10%
17.23%
26.94%
15.46%
27.99%
8.27%
0.00%
2021
0.00%
3.59%
17.05%
27.18%
15.59%
28.29%
8.29%
0.00%
2022
0.00%
3.03%
17.02%
27.82%
15.67%
28.43%
8.03%
0.00%
2023
0.00%
2.97%
17.10%
28.51%
15.47%
28.18%
7.77%
0.00%
2024
0.00%
2.46%
17.40%
28.11%
15.23%
28.49%
8.31%
0.00%
2025
0.00%
2.46%
17.40%
28.11%
15.23%
28.49%
8.31%
0.00%
Mathematically, an individual manufacturer multiplier is calculated by making the
segment the goal and dividing by the previous total for the segment (shown in Table 1-8). If
the number is greater than 1, the vehicles are entering the segment, and if the number is less
than 1, the vehicles are leaving the segment. So, for example, if Luxury Cars have an
adjustment of 1.5, then for a specific manufacturer who has Luxury Cars, a multiplier of 1.5 is
applied to its luxury car volume, and the total number of vehicles that shifted into the Luxury
segment is subtracted from the remaining segments to maintain that company's market share.
On the other hand, if Large Cars have an adjustment of 0.7, then for a specific manufacturer
who has Large Cars, a multiplier of 0.7 is applied to its Large Cars, and the total number of
vehicles leaving that segment is transferred into that manufacturer's Mid-Size Cars.
After the vehicle volumes are shifted using the above rules, a total for each
manufacturer and vehicle segment is maintained. The total for each manufacturer segment for
a specific model year (e.g., 2017 General Motors Luxury Cars) divided by the MY 2008 total
for that manufacturer segment (e.g., 2008 General Motors Luxury Cars) is the new multiplier
used to determine the future vehicle volume for each vehicle model. This is done by taking
the multiplier (which is for a specific manufacturer and segment) times the MY 2008 volume
for the specific vehicle model (e.g., 2008 General Motors Luxury Car Cadillac CTS). This
process is repeated for each model year (2017-2025).
The method used to adjust CSM Trucks to the AEO market share was different than
the method used for Cars. The process for Cars is different than Trucks because it is not
possible to predict how vehicles would shift between segments based on current market
trends. This is because of the added utility of some trucks that makes their sales more
insensitive to factors like fuel price. Again, CSM provided data on the market share and
vehicle segment distribution. The process for having the fleet match CSM's market share and
vehicle segment distribution was iterative.
The following totals were determined:
1-22
-------
The Baseline and Reference Vehicle Fleets
The total number of trucks for each manufacturer in 2008 model year.
The total number of trucks in each truck segment in 2008 model year.
The total number of truck in each segment for each manufacturer in 2008 model
year.
The total number of trucks for each manufacturer in a specific future model year
based on the AEO and CSM data. This is the goal for market share.
The total number of trucks in each truck segment in a specific future model year
based on the AEO and CSM data. This is the goal for vehicle segment
distribution. Table 1-9 has the percentages of Trucks per CSM segment.
Table 1-9 CSM - Percent of Trucks per Segment
CSM Segment
Full-Size CUV
Full-Size Pickup
Full-Size SUV
Full-Size Van
Mid-Size CUV
Mid-Size MAV
Mid-Size Pickup
Mid-Size SUV
Mid-Size Van
Small CUV
Small MAV
Small SUV
2017
5.9%
16.8%
1.9%
1.2%
18.0%
4.5%
6.1%
4.1%
11.6%
26.0%
2.5%
1.3%
2018
6.3%
16.5%
1.5%
1.2%
17.4%
4.6%
6.1%
4.8%
11.9%
25.9%
2.6%
1.2%
2019
6.8%
15.9%
1.3%
1.1%
17.6%
4.9%
6.1%
4.8%
11.9%
25.7%
2.8%
1.1%
2020
7.5%
16.1%
1.0%
1.4%
17.2%
5.4%
5.6%
4.5%
11.7%
25.6%
2.9%
1.2%
2021
8.3%
15.4%
0.9%
1.3%
16.9%
5.9%
5.7%
4.7%
11.6%
25.1%
3.0%
1.1%
2022
8.8%
15.1%
0.8%
1.3%
16.8%
6.2%
5.7%
4.8%
11.6%
24.9%
3.1%
1.1%
2023
9.5%
14.3%
0.5%
1.3%
16.8%
6.5%
5.8%
4.8%
11.6%
24.7%
3.1%
1.1%
2024
9.2%
13.8%
0.5%
1.2%
17.0%
7.1%
5.9%
4.6%
11.3%
25.3%
3.2%
1.0%
2025
9.1%
13.5%
0.6%
1.2%
17.0%
7.4%
5.8%
4.6%
11.3%
25.3%
3.2%
1.0%
To start, the agencies created two different types of tables. One table had each
manufacturer with its total sales for 2008 (similar to Table 1-11). This table will have the
goal for each manufacturer, and a column added for each iteration with the current total. The
second table has a truck segment total by manufacturer. The second table starts out with a
"Generic" manufacturer (Table 1-11) which is the table where the goal resides. Each
manufacturer (BMW for example is shown in Table 1-12) is then listed below the "Generic"
manufacturer. With each iteration, a new total is added for each segment that is calculated
and added to the table. This is not shown in the tables below. The agencies then engaged in a
process of first adjusting the numbers in the tables to the goal for market share distribution.
This was followed by adjusting to the goal for vehicle segment distribution. Each time an
adjustment was done a new column was added. An adjustment was done by creating a
multiplier (either segment distribution-based or manufacturer distribution-based) and applying
it to each vehicle segment total in the current iteration. A manufacturer-based multiplier is
calculated by taking the goal total for a manufacturer and dividing by the current total
(starting with 2008 model year volumes) for a manufacturer. A segment distribution-based
1-23
-------
The Baseline and Reference Vehicle Fleets
multiplier is calculated by taking the goal distribution volumes in the Generic manufacturer
set and dividing them by the current volume. Table 1-10, Table 1-11, and Table 1-12 below
illustrates two iterations using BMW as an example.
Table 1-10 Manufacturer Truck Totals
BMW
2008 Model Year Sales
61,324
Manufacturer Distribution 2017 Volume Goal
138.053
Multiplier for Iteration 1
138,053/61324=2.25
Table 1-11 Segment Specific Truck Totals for All Manufacturers
Manufacturer
Generic**
Generic
Generic
Generic
Generic
Generic
Generic
Generic
Generic
Generic
Generic
Generic
CSM Segment
Full-Size Pickup
Mid-Size Pickup
Full-Size Van
Mid-Size Van
Mid-Size MAV
Small MAV
Full-Size SUV
Mid-Size SUV
Small SUV
Full-Size CUV
Mid-Size CUV
Small CUV
2008 Model Year Sales
1,332,335
452,013
33,384
719,529
110,353
231,265
559,160
436,080
196,424
264,717
923,165
1,612,029
Segment Distribution 2017
Volume Goal
1,240,844
452,017
85,381
855,022
331,829
186,637
138,821
305,382
94,657
433,683
1,327,905
1,913,439
Multipliers
0.931
1.000
2.558
1.188
3.007
0.807
0.248
0.700
0.482
1.638
1.438
1.187
** Generic means all manufacturers.
1-24
-------
The Baseline and Reference Vehicle Fleets
Table 1-12 Segment Specific Truck Totals for BMW
Manufacturer
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
BMW
CSM Segment
Full-Size Pickup
Mid-Size Pickup
Full-Size Van
Mid-Size Van
Mid-Size MAV
Small MAV
Full-Size SUV
Mid-Size SUV
Small SUV
Full-Size CUV
Mid-Size CUV
Small CUV
Total BMW Vehicles
2008 Model Year
Sales
3,882
36,409
21,033
61,324
Iteration 1 Adjust for
Market Share
2.25*3,882=8,739
2.25*36,409=81,964
2.25*21,033=47,350
138,053
Iteration 2 Adjust for Segment
Distribution
2.85*8,739=24,907
1.1*81,964=90,134
1.. 02*47,350=48,306
163,347
Using this process, the numbers will get closer to the goal of matching CSM's market
share for each manufacturer and distribution for each vehicle segment after each of the
iterations. The iterative process is carried out until the totals nearly match the goals.
After 19 iterations, all numbers were within 0.01% of CSM's distributions. The
calculation iterations could have been stopped sooner, but they were continued to observe
how the numbers would converge.
After the market share and segment distribution were complete, the totals need to be
used to create multipliers that could be applied to the original individual 2008 model year
vehicle volumes (each unique manufacture models volume). The total for each manufacturer
segment divided by the 2008 model year total for each manufacturer segment gives a
multiplier that can be applied to each vehicle based on its manufacturer and segment.
The above process is done for each model year needed (2017-2025). The multipliers
are then applied to each vehicle in 2008 model year, which gives a volume for each vehicle in
2017 through 2025 model year.
1.3.3 What are the sales volumes and characteristics of the MY 2008 based
reference fleet?
Table 1-13 and Table 1-15 below contain the sales volumes that result from the
process above for MY 2008 and 2017-2020. Table 1-14 and Table 1-16 below contain the
sales volumes that result from the process above for MY 2021-2025.
Table 1-13 Vehicle Segment Volumes"
Reference Class Segment
Actual and Projected Sales Volume
2008
2017
2018
2019
2020
1-25
-------
The Baseline and Reference Vehicle Fleets
LargeAuto
MidSizeAuto
CompactAuto
SubCmpctAuto
LargePickup
SmallPickup
LargeSUV
MidSizeSUV
Small SUV
MiniVan
Cargo Van
562,240
3,098,927
1,979,461
1,365,833
1,582,226
177,497
2,783,949
1,263,360
285,355
642,055
110,858
376,107
3,311,268
2,347,980
2,458,222
1,514,619
156,227
3,194,489
1,358,755
148,251
754,562
185,841
356,768
3,290,408
2,325,393
2,454,112
1,443,766
157,932
3,150,101
1,309,212
149,933
739,551
199,234
353,609
3,303,621
2,369,301
2,489,208
1,383,190
160,752
3,177,868
1,267,394
154,675
717,065
201,974
394,864
3,381,785
2,448,021
2,553,350
1,386,195
146,029
3,203,244
1,285,822
162,677
714,323
219,628
a Volumes in this table are based on the pre-2011 NHTSA definition of Cars and Trucks.
Table 1-14 Vehicle Segment Volumes"
Reference Class Segment
LargeAuto
MidSizeAuto
CompactAuto
SubCmpctAuto
LargePickup
SmallPickup
LargeSUV
MidSizeSUV
Small SUV
MiniVan
Cargo Van
Projected Sales Volume
2021
380,192
3,442,116
2,520,977
2,626,364
1,368,301
150,123
3,312,914
1,281,240
167,223
729,078
210,539
2022
358,295
3,548,263
2,592,199
2,687,167
1,349,421
147,138
3,362,608
1,283,244
169,643
738,982
202,812
2023
362,672
3,692,533
2,632,926
2,721,102
1,301,293
151,315
3,412,753
1,268,288
170,239
740,785
201,585
2024
356,173
3,751,496
2,744,634
2,796,061
1,271,751
154,627
3,475,873
1,292,662
173,191
720,720
196,900
2025
368,843
3,814,941
2,843,069
2,878,288
1,260,389
154,838
3,520,992
1,305,362
175,713
726,256
201,768
a Volumes in this table are based on the pre-2011 NHTSA definition of Cars and Trucks.
Table 1-15 2011+ NHTSA Car and Truck Definition Based Volumes
Vehicle Type
Trucks
Cars
Cars and Trucks
Actual and Projected Sales Volume
2008
5,621,193
8,230,568
13,851,761
2017
5,818,655
9,987,667
15,806,322
2018
5,671,046
9,905,364
15,576,410
2019
5,582,962
9,995,696
15,578,658
2020
5,604,377
10,291,562
15,895,939
Table 1-16 2011+ NHTSA Car and Truck Definition Based Volumes
Vehicle Type
Projected Sales Volume
2021
2022
2023
2024
2025
1-26
-------
The Baseline and Reference Vehicle Fleets
Trucks
Cars
Cars and Trucks
5,683,902
10,505,165
16,189,066
5,703,996
10,735,777
16,439,772
5,687,486
10,968,003
16,655,489
5,675,949
11,258,138
16,934,087
5,708,899
11,541,560
17,250,459
Table 1-17 and Table 1-18 below contain the sales volumes by manufacturer and
vehicle type for MY 2008 and 2017-2025.
Table 1-17 NHTSA Car and Truck Definition Manufacturer Volumes
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
Chrysler/Fiat
Chrysler/Fiat
Daimler
Daimler
Ferrari
Ferrari
Ford
Ford
Geely/Volvo
Geely/Volvo
GM
GM
HONDA
HONDA
HYUNDAI
HYUNDAI
Kia
Kia
Lotus
Lotus
Mazda
Vehicle 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
2008
Baseline
Sales
13,851,761
8,230,568
5,621,193
1,370
-
291,796
61,324
703,158
956,792
208,195
79,135
1,450
-
956,699
814,194
32,748
65,649
1,507,797
1,587,391
1,006,639
505,140
337,869
53,158
221,980
59,472
252
-
246,661
2017
Projected
Volume
15,806,322
9,987,667
5,818,655
1,035
-
313,022
138,053
418,763
409,702
284,847
86,913
6,676
-
1,299,899
763,549
41,887
88,234
1,362,761
1,462,204
1,154,600
596,481
592,027
152,885
322,044
98,702
240
-
253,540
2018
Projected
Volume
15,576,410
9,905,364
5,671,046
1,051
-
322,939
131,942
397,538
387,858
276,409
83,651
6,700
-
1,311,467
748,829
42,187
89,394
1,438,355
1,474,076
1,138,087
544,619
578,373
151,461
312,370
98,280
243
-
262,512
2019
Projected
Volume
15,578,658
9,995,696
5,582,962
1,072
-
346,075
131,373
391,689
366,447
281,425
88,188
6,794
-
1,332,039
717,773
43,125
91,575
1,505,025
1,493,511
1,144,639
527,535
582,971
155,642
314,879
100,679
250
-
266,951
2020
Projected
Volume
15,895,939
10,291,562
5,604,377
1,034
-
357,942
128,339
415,319
360,677
290,989
92,919
6,916
-
1,378,789
717,037
42,615
93,003
1,530,755
1,544,983
1,163,666
525,089
598,283
154,173
323,676
96,535
266
-
270,078
1-27
-------
The Baseline and Reference Vehicle Fleets
Mazda
Mitsubishi
Mitsubishi
Nissan
Nissan
PORSCHE
PORSCHE
Spyker/Saab
Spyker/Saab
Subaru
Subaru
Suzuki
Suzuki
Tata/JLR
Tata/JLR
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
55,885
85,358
15,371
717,869
305,546
18,909
18,797
21,706
4,250
116,035
82,546
79,339
35,319
9,596
55,584
800
-
1,260,364
951,136
291,483
26,999
51,788
65,099
37,632
870,797
444,938
35,093
13,233
20,024
2,871
224,112
78,242
90,708
22,109
55,881
57,579
27,986
-
1,849,196
1,330,511
551,638
128,819
57,535
63,671
36,300
849,678
412,383
35,444
12,001
20,007
3,596
216,598
75,152
89,932
21,385
56,222
56,606
28,435
-
1,834,181
1,223,415
540,036
145,491
57,494
63,826
35,454
854,400
398,559
36,116
11,469
20,144
3,826
217,095
72,832
90,568
20,692
57,267
57,854
28,990
-
1,836,306
1,142,104
537,114
146,891
58,154
65,080
35,215
882,791
397,869
35,963
11,141
21,069
3,509
223,466
72,458
93,548
20,675
58,182
56,213
27,965
-
1,883,734
1,154,304
554,822
146,700
Table 1-18 NHTSA Car and Truck Definition Manufacturer Volumes
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
Chrysler/Fiat
Chrysler/Fiat
Daimler
Daimler
Ferrari
Ferrari
Vehicle Type
Both
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
2021
Projected
Volume
16,189,066
10,505,165
5,683,902
1,058
-
359,098
128,724
421,013
348,613
300,378
99,449
7,059
-
2022
Projected
Volume
16,439,772
10,735,777
5,703,996
1,049
-
360,034
128,899
424,173
363,008
304,738
100,935
7,138
-
2023
Projected
Volume
16,655,489
10,968,003
5,687,486
1,041
-
360,561
127,521
423,882
361,064
312,507
105,315
7,227
-
2024
Projected
Volume
16,934,087
11,258,138
5,675,949
1,141
-
388,193
146,525
426,017
344,962
332,337
107,084
7,441
-
2025
Projected
Volume
17,250,459
11,541,560
5,708,899
1,182
-
405,256
145,409
436,479
331,762
340,719
101,067
7,658
-
1-28
-------
The Baseline and Reference Vehicle Fleets
Ford
Ford
Geely/Volvo
Geely/Volvo
GM
GM
HONDA
HONDA
HYUNDAI
HYUNDAI
Kia
Kia
Lotus
Lotus
Mazda
Mazda
Mitsubishi
Mitsubishi
Nissan
Nissan
PORSCHE
PORSCHE
Spyker/Saab
Spyker/Saab
Subaru
Subaru
Suzuki
Suzuki
Tata/JLR
Tata/JLR
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
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
1,401,617
714,181
41,768
92,726
1,530,020
1,564,277
1,198,880
535,916
613,355
156,466
331,319
95,432
278
-
274,740
59,227
65,851
35,309
912,629
408,029
36,475
11,242
21,294
3,560
230,780
72,773
95,725
20,767
58,677
58,153
28,623
-
1,903,706
1,215,539
585,607
148,734
1,415,221
714,266
41,686
92,512
1,507,653
1,578,556
1,237,504
539,235
627,964
157,493
339,102
94,694
290
-
281,150
60,307
67,261
35,227
937,447
411,883
36,607
11,385
21,709
3,461
238,613
72,736
97,599
20,734
59,349
58,590
28,369
-
1,986,077
1,235,052
593,314
146,750
1,474,797
700,005
42,031
96,840
1,496,819
1,606,495
1,265,564
536,898
634,308
161,189
342,746
95,688
299
-
296,910
61,966
67,680
35,469
954,340
417,121
36,993
11,370
22,410
3,435
241,612
73,022
99,263
20,803
60,639
58,865
28,150
-
2,036,992
1,224,980
596,749
153,927
1,503,670
688,854
42,461
99,181
1,493,597
1,636,805
1,307,851
536,994
657,710
166,092
351,882
96,119
308
-
300,614
61,971
70,728
36,001
982,771
422,217
39,504
11,409
22,800
3,426
248,283
74,142
100,447
21,162
63,728
57,981
30,862
-
2,080,528
1,208,013
605,336
156,939
1,540,109
684,476
42,588
101,107
1,524,008
1,673,936
1,340,321
557,697
677,250
168,136
362,783
97,653
316
-
306,804
61,368
73,305
36,387
1,014,775
426,454
40,696
11,219
23,130
3,475
256,970
74,722
103,154
21,374
65,418
56,805
31,974
-
2,108,053
1,210,016
630,163
154,284
Table 1-19 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 2008 and 2025. There will be an increase in the number of cars sold, which will
cause the average footprints for cars and trucks combined to be slightly smaller (about 2%).
1-29
-------
The Baseline and Reference Vehicle Fleets
This is the result of AEO projecting an increased number of cars, and CSM predicting that
most of that increase will be in the subcompact segment. Again, we note that in order to
ensure that our baseline inputs were not influenced by the final regulations, agencies re-ran
AEO to hold standards constant after 2016 (the reader will remember from the text above that
CSM had indicated that its projections were not sensitive to assumptions about new
standards).
Table 1-19 Production Weighted Foot Print Mean
Model
Year
2008
2017
2018
2019
2020
2021
2022
2023
2024
2025
Average Footprint of all
Vehicles
48.8
48.0
47.9
47.8
47.8
47.8
47.7
47.7
47.5
47.5
Average Footprint
Cars
45.2
44.6
44.6
44.6
44.6
44.6
44.6
44.6
44.6
44.6
Average Footprint
Trucks
53.9
53.8
53.7
53.6
53.7
53.6
53.6
53.5
53.3
53.3
Table 1-20 below shows the changes in engine cylinders over the model years. The
current assumptions show that engines will be downsized over the model years to which these
final rules apply. This shift is a projected consequence of the expected changes in class and
segment mix as predicted by AEO and CSM, and does not represent engine downsizing
attributable to the 2012-2016 light-duty CAFE and GHG standards.
Table 1-20 Percentages of 4,6,8 Cylinder Engines by Model Year
Model
Year
2008
2017
2018
2019
2020
2021
2022
Trucks
4
Cylinders
10.3%
10.9%
10.6%
10.4%
10.3%
10.3%
10.3%
6
Cylinders
56.4%
63.7%
64.5%
65.5%
65.6%
66.3%
66.7%
8
Cylinders
33.3%
25.4%
24.8%
24.1%
24.1%
23.4%
23.0%
Cars
4
Cylinders
56.9%
60.6%
60.7%
60.7%
60.3%
60.6%
61.1%
6
Cylinders
37.8%
34.5%
34.4%
34.3%
34.7%
34.4%
34.2%
8
Cylinders
5.3%
5.0%
5.0%
5.0%
5.0%
4.9%
4.8%
1-30
-------
The Baseline and Reference Vehicle Fleets
2023
2024
2025
10.3%
10.5%
10.5%
67.7%
68.1%
68.2%
22.0%
21.4%
21.3%
60.9%
61.0%
61.1%
34.3%
34.1%
34.0%
4.8%
4.8%
4.8%
As discussed above, the agencies also developed a second market forecast using updated data.
The following section describes those efforts and their results.
1.4 The 2010 MY Based Fleet
The 2010 MY based fleet is similar to the 2008 MY based fleet in that it was created
with similar types of information. The 2010 MY based fleet uses interim AEO 2012 total car
and truck volumes, a long range forecast from LMC Automotive (formerly J.D. Powers
Forecasting) used for manufacturer market share and product mix, and 2010 CAFE
certification data for 2010 model volumes and technology. The 2008 MY based fleet, in
contrast, uses interim AEO 2011, a long range forecast from CSM World Wide, and 2008
CAFE certification data. The remainder of section 1.4 describes the 2010 based fleet
projection and how it was created.
1.4.1
On what data is the MY 2010 baseline vehicle fleet based?
Similar to the 2008 baseline, most of the information about the vehicles that make up
the 2010 fleet was gathered from EPA's emission certification and fuel economy database,
most of which is available to the public. These data included, by individual vehicle model
produced in MY 2010, vehicle production volume, fuel economy rating for CAFE
certification, carbon dioxide emissions, 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 to this information about each vehicle model produced in MY
2010, the agencies augmented this description with publicly-available data which includes
more complete technology descriptions from Ward's Automotive Group.111'11 As with the
2008 baseline, the agencies also used Edmunds.com and Motortrend.com°'p'q Like the MY
2008 baseline fleet and the baseline vehicle fleet used in the MYs 2012-2016 rulemaking, the
MY 2010 baseline vehicle fleet is developed using publicly-available data to the largest
extent possible.
0 Motortrend.com and Edmunds.com: Used as a source for footprint and vehicle weight data.
p Motortrend.com and Edmunds.com are free, no-fee internet sites.
q A small amount of footprint data from manufacturers' MY 2008 product plans submitted to the agencies was
used in the development of the baseline.
1-31
-------
The Baseline and Reference Vehicle Fleets
The process for creating the 2010 baseline fleet Excel file was streamlined when
compared with the past rulemaking. EPA and NHTSA worked together to create the baseline
using 2010 CAFE certification data from EPA's Verify database. EPA contracted LMC
Automotive (formerly JD Power Forecasting) to produce an up to date long range forecast of
volumes for the future fleet. Using information sources discussed below, NHTSA identified
technology and footprint information for every vehicle model in the 2010 CAFE certification
data. EPA used the forecast from LMC Automotive to project the future fleet's volume
projections (a detailed discussion of the method used to project the future fleet volumes is in
1.4.2.1 of this chapter.)
Both agencies used the previously mentioned data to populate input files for the
agencies' respective modeling systems. The structure of the market forecast input file used
for DOT's CAFE Compliance and Effects Modeling System (a.k.a. "the CAFE model") is
described in the model documentation.5 To help readers who wish to directly examine the
baseline fleet file for EPA's OMEGA model, and to provide some idea of its contents for
those readers who do not, Table 1-21 shows the columns of the complete fleet file, which
includes the MY 2008 baseline data that was compiled. Each column has its name, definition
(description) and source. Most elements shown in Table 1-21 also appear in the market
forecast input file for DOT's modeling system, which accommodates some additional data
elements discussed in the model documentation.
Table 1-21 Data, Definitions, and Sources
Data Item
Index
Manufacturer
CERT
Manufacturer
Name
Name Plate
Model
Reg Class
Our Class
Definition
Index Used to link EPA and NHTSA baselines
Common name of company that manufactured
vehicle. May include more name plates than Cert
Manufacturer Name.
Certification name of company that manufactured
vehicle
Name of Division
Name of Vehicle
EPA Fuel Economy Class Name
If a car's Footprint<43 then "SubCmpctAuto"
If a car's 43<=Footprint<46 then "CompactAuto"
If a car's 46<=Footprint<53 then "MidSizeAuto"
If a car's Footprint >=53 then "LargeAuto"
If a S.U.V.'s Footprint < 43 then "SmallSuv"
If a S U V 's 43<=Footprint<46 then
"MidSizeSuv"
If a S.U.V's Footprint >=46 then "LargeSuv"
If a Truck's Footprint < 50 then "SmallPickup"
If a Truck's Footprint>=50 then "LargPickup"
If a Van's Structure is Ladder then "CargoVan"
If a Van's Structure is Unibody then "Minivan"
Data Type
Number
Name
(Ex.Chrysler)
Name
(Ex.Chrysler)
Name (Ex. Dodge)
Name (Ex. Viper)
EPA Class Name
(Ex.
SUBCOMPACT
CARS)
Name(Ex
SmallSuv)
Wards
Engine
Acronyms
NA
NA
NA
NA
NA
NA
NA
Where The
Data is From
Created
Certification data
Certification data
Certification data
Certification data
Certification data
Derived From
Certification data
and Footprint
1-32
-------
The Baseline and Reference Vehicle Fleets
NEW
SEGMENT
Vehicle Type
Number
Generic
Vehicle Index
From Sum
Page
Vehicle Index
From Sum
Page
Pre2011
NHTSA
Defined C/T
Our Class C/T
Traditional
Car/Truck
NHTSA
Defined New
NHTSA
Car/Truck
Total
Production
Volume
Fuel Econ.
(mpg)
C02
Fuel (G,D,C)
Fuel Type
Disp (lit.)
Effective Cyl
Actual
Cylinders
Valves Per
Cylinder
Valve Type
Valve
Actuation
WT
WLT
LMC Automotive (formerly J. D. Powers) new
segmentation for the vehicle.
Vehicle Type Number assigned to a vehicle based
on its number of cylinders, valves per cylinder, and
valve actuation technology. See Truck Vehicle
Type Map and Car Vehicle Type Map sheets for
details.
Number to be used as a cross reference with the
Sum Pages.
Number to be used as a cross reference with the
Sum Pages.
C= Car, T=Truck. As defined in the certification
database.
C= Car, T=Truck. As defined in the certification
database. Not used in calculations.
DP=Domestic Passenger Cars, I=Import Passenger
Car, LT= Light duty Truck. As defined in the
certification database. Not used in calculations.
New NHTSA Car Truck value as determined by
NHTSA. Used in calculations.
Total number of vehicles produced for that model.
EPA Unadjusted Fuel Economy
CO2 calculated from MPG. CO2 weighted 1.15
times higher for diesel vehicles.
Gas or Diesel or CNG
Gas or Diesel or CNG
Engine Cylinder Displacement Size in Liters
Number of Cylinder + 2 if the engine has a turbo or
super charger.
Actual Number of Engine Cylinders
Number of Valves Per Actual Cylinder
Type of valve actuation.
Type of valve actuation with values compatible
with the package file.
Type of valve timing with values compatible with
the package file.
Type of valve lift with values compatible with the
package file.
Name (Ex.
Compact Sporty,
Large Pickup, etc.)
Number
Number
Number
Letter(C or T)
Letter(C or T)
IP,DP,LT
Letter(C or T)
number(ex.5500)
number(ex.25)
Number
G,D,C
Gas or Diesel or
CNG
number(ex. 4)
number(ex. 6)
number(ex. 4)
number(ex. 4)
Acronym(Ex.
DOHC, SOHC,
OHV, E, R)
DOHC, SOHC,
OHV
VVTC,VVTD,
VVTI
VVTLC, VVTLD
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
DOHC, SOHC,
OHV, E, R
DOHC, SOHC,
OHV
VVTC,VVTD,
VVTI
VVTLC,
VVTLD
LMC Automotive
Mapped by EPA
staff
NA
NA
Certification data
Created
Certification data
Certification data
Certification data
Certification data
Certification data
Certification data
Certification data
Wards/Certificati
on data
Derived From
Certification data.
Certification data
Certification data
Wards (Note:
Type E is from
Cert Data)
Wards
Wards
Wards
1-33
-------
The Baseline and Reference Vehicle Fleets
Deac
Inline or V
Engine
Fuel injection
system
Boost
Engine Cycle
Horsepower
Torque
Cooled EGR
Trans Type
Tran
Num of Gears
Structure
Drive
Drive with
AWD
Wheelbase
Track Width
(front)
Track Width
(rear)
Footprint
Threshold
Footprint
Curb
Weight
ITW
GVWR
Stop-Start/
Hybrid/ Full
EV
Towing
Capacity
(Maximum)
Engine Oil
Viscosity
Low drag brakes
Cylinder Deactivation with a value that is
compatible with the package file.
Configuration of the Engine
Type of fuel injection.
Type of Boost if any.
As Defined by EPA Cert. Definition
Max. Horsepower of the Engine
Max. Torque of the Engine
Cooled Exhaust Gas Recirculation
A=Auto AMT= Automated Manual M=Manual
CVT= Continuously Variable Transmission
Type Code with number of Gears
Number of Gears
Ladder or Unibody
Fwd, Rwd, 4wd
Fwd, Rwd, Awd, 4wd
Length of Wheelbase
Length of Track Width in inches
Length of Track Width in inches
Car and Large Truck Footprints are normal
(Average Track x Wheelbase). Medium and Small
Truck footprints are the production weighted
average for each vehicle.
Footprint valve that will be set to 41 for values less
than 41, Will be set to 56 for car values > 56, and
will be set to 66 for truck values >66
Curb Weight of the Vehicle
Inertia Test Weight
Gross Vehicle Weight Rating of the Vehicle
Type of Electrification if any. Blank = None
Weight a vehicle is rated to tow.
Ratio between the applied shear stress and the rate
of shear, which measures the resistance of flow of
the engine oil (as per SAE Glossary of Automotive
Terms)
See Volpe Documentation
Deac
lorV
DI, MPI
Super Charged
(Single), Turbo
(Single)
Letter Ex. G for
Gas)
number(ex. 125)
number(ex. 125)
YorN
letter(ex. A)
letters and possible
a number(ex.A5,
ex. CVT)
number(ex. 4)
Unibody or Ladder
(Ex. Ladder)
Acronym(Ex. Rwd)
Acronym(Ex. Awd)
number(ex. 125)
number(ex. 45)
number(ex. 45)
Number
Number
number(ex.4500)
number(ex.4500)
number(ex.4500)
EV75,2-Mode-
IMA,Power-
Split, Stop-Start
Number (in
Pounds)
Text (Ex. OW20;
5W20)
See Volpe
CD
lorV
DI, SFI, EFI,
MPI
TRB,SPR
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Wards
Wards
Wards
Wards
Wards
Wards
Wards
Certification data
Certification data
Certification data
Certification data
Volpe Input File
Certification data
Certification data
From
Edmonds.com or
Motortrend.com,
From
Edmonds.com or
Motortrend.com
From
Edmonds.com or
Motortrend.com
From
Edmonds.com or
Motortrend.com
Derived from
data from
Edmunds.com or
Motortrend.com
Certification data
Certification data
Volpe Input File
Certification data
Volpe Input File
Volpe Input File
Volpe Input File
1-34
-------
The Baseline and Reference Vehicle Fleets
Power steering
Technology
Class
Safety Class
Safety Class
Number
Volume 20 10
Volume 20 11
Volume 20 12
Volume 20 13
Volume 20 14
Volume 20 15
Volume 20 16
See Volpe Documentation
For technology application purposes only and
should not be confused with vehicle classification
for regulatory purposes. Defined by DOT.
See Volpe Documentation
See Volpe Documentation
Projected Production Volume for 2010
Projected Production Volume for 2011
Projected Production Volume for 2012
Projected Production Volume for 2013
Projected Production Volume for 2014
Projected Production Volume for 2015
Projected Production Volume for 2016
Documentation
See Volpe
Documentation
Text (Ex.
Subcompact,
Subcompact
Performance,
Compact, Compact
Performance,
Midsize, Midsize
Performance,
Large, Large
Performance,
Minivan, Small LT,
Midsize LT, Large
LT; (LT =
SUV/Pickup/Van))
See Volpe
Documentation
See Volpe
Documentation
Number
Number
Number
Number
Number
Number
Number
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Volpe Input File
Volpe Input File
Volpe Input File
Volpe Input File
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
1-35
-------
The Baseline and Reference Vehicle Fleets
Volume 20 17
Volume 20 18
Volume 20 19
Volume 2020
Volume 2021
Volume 2022
Volume 2023
Volume 2024
Volume 2025
Projected Production Volume for 2017
Projected Production Volume for 2018
Projected Production Volume for 2019
Projected Production Volume for 2020
Projected Production Volume for 2021
Projected Production Volume for 2022
Projected Production Volume for 2023
Projected Production Volume for 2024
Projected Production Volume for 2025
Number
Number
Number
Number
Number
Number
Number
Number
Number
NA
NA
NA
NA
NA
NA
NA
NA
NA
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Calculated based
onMY2010
volume and AEO
and LMC
adjustment
factors.
Table 1-22 displays the engine technologies present in the MY 2010 baseline fleet.
Again, the engine technologies for the vehicles manufactured by these manufacturers in MY
2010 were largely obtained from data found on Ward's Auto online.
1-36
-------
The Baseline and Reference Vehicle Fleets
Table 1-22 2010 Engine Technology Percentages
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
Chrysler/Fiat
Chrysler/Fiat
Daimler
Daimler
Ferrari
Ferrari
Ford
Ford
Geely
Geely
General Motors
General Motors
Honda
Honda
Hyundai
Hyundai
Kia
Kia
Lotus
Lotus
Mazda
Mazda
Mitsubishi
Mitsubishi
Nissan
Nissan
Porsche
Porsche
Spyker
Spyker
Subaru
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
3%
4%
2%
0%
0%
38%
33%
0%
0%
0%
8%
0%
0%
1%
1%
38%
25%
0%
0%
1%
2%
3%
0%
0%
0%
0%
0%
4%
11%
6%
0%
0%
0%
16%
1%
0%
0%
6%
Super Charged
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
16%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Single Overhead Cam
22%
18%
29%
0%
0%
0%
0%
42%
30%
50%
24%
0%
0%
12%
70%
0%
0%
0%
0%
58%
63%
0%
0%
0%
0%
0%
0%
0%
0%
100%
100%
0%
0%
0%
0%
0%
0%
92%
Dual Over Head Cam
68%
78%
50%
100%
0%
100%
100%
49%
4%
50%
76%
100%
0%
88%
30%
100%
100%
45%
75%
42%
37%
100%
100%
100%
100%
100%
0%
100%
100%
0%
0%
100%
100%
100%
100%
0%
0%
8%
Over Head Cam
10%
4%
21%
0%
0%
0%
0%
9%
66%
0%
0%
0%
0%
0%
0%
0%
0%
55%
25%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Variable Valve Timing
Continuous
41%
11%
17%
0%
0%
28%
0%
0%
0%
52%
24%
0%
0%
2%
50%
0%
0%
0%
0%
58%
63%
0%
0%
0%
0%
0%
0%
0%
0%
96%
74%
0%
0%
0%
0%
0%
0%
0%
Variable Valve Timing
Discrete
26%
26%
27%
0%
0%
70%
82%
41%
4%
46%
69%
100%
0%
0%
0%
100%
100%
42%
73%
42%
37%
0%
0%
0%
0%
100%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
2%
Variable Valve Timing
Intake Only
39%
48%
22%
38%
0%
0%
0%
0%
0%
0%
0%
0%
0%
69%
26%
0%
0%
0%
1%
0%
0%
100%
100%
100%
73%
0%
0%
100%
100%
0%
0%
100%
100%
100%
100%
0%
0%
6%
Variable Valve Lift and
Timing Continuous
6%
6%
5%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
42%
37%
0%
0%
0%
0%
23%
0%
0%
0%
0%
0%
9%
0%
100%
100%
0%
0%
2%
Variable Valve Lift and
Timing Discrete
2%
2%
2%
0%
0%
45%
67%
0%
0%
46%
76%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Cylinder Deactivation
4%
3%
7%
0%
0%
0%
0%
5%
13%
0%
0%
0%
0%
0%
0%
0%
0%
4%
3%
17%
45%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
Direct Injection
9%
9%
9%
0%
0%
38%
33%
0%
0%
0%
8%
90%
0%
1%
1%
0%
0%
37%
31%
0%
0%
0%
0%
0%
0%
0%
0%
2%
0%
0%
0%
0%
0%
83%
100%
0%
0%
0%
1-37
-------
The Baseline and Reference Vehicle Fleets
Subaru
Suzuki
Suzuki
Tata
Tata
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
0%
0%
0%
0%
0%
0%
0%
0%
0%
62%
33%
0%
0%
0%
22%
15%
0%
0%
0%
0%
4%
0%
87%
0%
0%
0%
0%
0%
0%
0%
0%
68%
21%
13%
100%
100%
100%
100%
0%
0%
100%
100%
32%
79%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
47%
21%
13%
2%
50%
67%
67%
0%
0%
16%
62%
32%
67%
0%
98%
50%
33%
33%
0%
0%
84%
38%
0%
0%
13%
0%
0%
45%
64%
0%
0%
0%
0%
1%
52%
0%
0%
0%
7%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
67%
67%
0%
0%
4%
0%
68%
100%
The data in Table 1-22 indicate that manufacturers had already begun implementing a
number of fuel economy/GHG reduction technologies in the baseline (2010) fleet. For
example, as in the 2008 baseline fleet, VW stands out as having a significant number of
turbocharged direct injection engines. Some of the valve and cam technologies are quite
common in the baseline fleet: for example, nearly half the baseline fleet already has dual cam
phasing, while Honda and Chrysler have considerable levels of engines with cylinder
deactivation. Honda also has already implemented continuously variable valve lift on a
majority of their engines. Part of the implication of these technologies already being present
in the baseline is that if manufacturers have already implemented them, they are therefore not
available in the rulemaking analysis for improving fuel economy and reducing CC>2 emissions
further, requiring the agencies to look toward increasing penetration of these and other
technologies and increasingly advanced technologies to project continued improvements in
stringency over time.
The data in Table 1-23 shows the changes between the 2010 engine technology
penetrations and the 2008 engine technology penetrations. Perhaps to increase fuel economy,
manufacturers applied considerable additional technology between 2008 and 2010.
Volkswagen's trucks have direct injection increased to 100 percent (although VW's cars had a
21% decrease). Manufacturers changed variable valve timing, presumably based on engine-
specific design considerations. For example, Honda replaced discrete valve timing with
continuous valve lift or timing, and Kia added variable valve lift and timing to 90% of its cars
and 56% of its trucks.
1-38
-------
The Baseline and Reference Vehicle Fleets
Table 1-23 The difference (2010-2008) in Engine Technology Percentages
Manufacturer
All
All
All
Aston Martin
Aston Martin
BMW
BMW
Chrysler/Fiat
Chrysler/Fiat
Daimler
Daimler
Ferrari
Ferrari
Ford
Ford
Geely/Volvo
Geely/Volvo
GM
GM
Honda
Honda
Hyundai
Hyundai
Kia
Kia
Lotus
Lotus
Mazda
Mazda
Mitsubishi
Mitsubishi
Nissan
Nissan
Porsche
Porsche
Spyker/Saab
Spyker/Saab
Subaru
Subaru
Suzuki
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
Cars
1
o
o
|
H
0%
0%
1%
0%
0%
5%
28%
-1%
0%
-2%
-8%
0%
0%
1%
1%
38%
-24%
0%
-1%
1%
-2%
3%
0%
0%
0%
0%
0%
-7%
-13%
0%
0%
0%
0%
-1%
-11%
NA
NA
-9%
-3%
0%
Super Charged
0%
0%
0%
0%
0%
-1%
0%
0%
0%
0%
0%
0%
0%
-1%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-61%
0%
0%
0%
0%
0%
0%
0%
0%
0%
NA
NA
0%
0%
0%
Single Overhead Cam
2%
1%
5%
0%
0%
-14%
0%
21%
-9%
-5%
-12%
0%
0%
-3%
5%
0%
0%
0%
0%
1%
-1%
0%
0%
0%
0%
0%
0%
0%
-1%
0%
0%
0%
0%
0%
0%
NA
NA
23%
17%
0%
Dual Overhead Cam
5%
5%
2%
0%
0%
14%
0%
-23%
0%
5%
12%
0%
0%
3%
-2%
0%
0%
14%
19%
-1%
1%
0%
0%
0%
0%
0%
0%
1%
1%
0%
0%
0%
0%
0%
0%
NA
NA
-23%
-17%
0%
Overhead Cam
-7%
-5%
-8%
0%
0%
0%
0%
1%
9%
0%
0%
0%
0%
0%
-3%
0%
0%
-14%
-19%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
NA
NA
0%
0%
0%
Variable Valve Timing
Continuous
33%
2%
11%
0%
0%
14%
0%
0%
0%
-20%
-11%
0%
0%
-2%
22%
0%
0%
-5%
-29%
58%
63%
0%
0%
0%
0%
0%
0%
0%
0%
-4%
36%
0%
0%
0%
0%
NA
NA
0%
0%
0%
Variable Valve Timing
Discrete
4%
2%
8%
-100%
0%
-16%
-18%
-1%
0%
42%
52%
0%
0%
0%
-1%
0%
0%
25%
42%
15%
33%
0%
0%
0%
0%
0%
0%
-7%
-13%
0%
0%
-4%
0%
-100%
-100%
NA
NA
2%
-10%
2%
Variable Valve Timing
Intake Only
9%
13%
-1%
38%
0%
0%
0%
0%
0%
-13%
-47%
0%
0%
22%
17%
0%
0%
-14%
0%
-20%
-28%
0%
0%
90%
56%
0%
0%
8%
13%
0%
0%
4%
0%
100%
100%
NA
NA
-25%
-7%
98%
Variable Valve Lift and
Timing Continuous
6%
6%
5%
-24%
0%
0%
0%
0%
0%
0%
0%
-29%
0%
0%
0%
0%
0%
0%
0%
42%
37%
0%
0%
0%
0%
23%
0%
0%
0%
0%
0%
9%
0%
100%
100%
NA
NA
2%
13%
0%
Variable Valve Lift and
Timing Discrete
-10%
-11%
-8%
0%
0%
32%
67%
0%
0%
46%
76%
0%
0%
0%
0%
0%
0%
0%
0%
-100%
-100%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
-100%
NA
NA
-1%
-27%
0%
o
ta
>
o
1
>->
U
-2%
0%
-4%
0%
0%
0%
0%
0%
9%
0%
0%
0%
0%
0%
0%
0%
0%
-36%
-1%
6%
45%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
0%
NA
NA
0%
0%
0%
.0
o
o
'&
Q
4%
2%
6%
0%
0%
5%
27%
0%
0%
-2%
-8%
90%
0%
1%
1%
0%
0%
37%
25%
0%
-4%
0%
0%
0%
0%
0%
0%
-9%
-24%
0%
0%
0%
0%
66%
0%
NA
NA
0%
0%
0%
1-39
-------
The Baseline and Reference Vehicle Fleets
Suzuki
Tata/JLR
Tata/JLR
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
0%
0%
0%
NA
NA
0%
0%
19%
1%
0%
22%
-5%
NA
NA
0%
0%
4%
0%
0%
0%
0%
NA
NA
0%
0%
-17%
0%
0%
0%
0%
NA
NA
0%
0%
17%
100%
0%
0%
0%
NA
NA
0%
0%
0%
0%
0%
0%
0%
NA
NA
0%
0%
47%
0%
50%
-9%
67%
NA
NA
-13%
1%
-16%
99%
50%
9%
-67%
NA
NA
13%
-1%
0%
0%
0%
45%
64%
NA
NA
0%
0%
1%
0%
0%
7%
0%
NA
NA
0%
0%
-1%
79%
0%
0%
0%
NA
NA
0%
0%
0%
0%
0%
67%
67%
NA
NA
-4%
-6%
-21%
100%
The section below provides further detail on the conversion of the MY 2010 baseline
into the MYs 2017-2025 reference fleet. It also describes more of the data contained in the
baseline spreadsheet.
1.4.2
The MY 2010 Based MY 2017-2025 Reference Fleet
The reference fleet aims to reflect the current market conditions and expectations
about conditions of the vehicle fleet during the model years to which the agencies' rules
apply. Fundamentally, constructing this fleet involved projecting the MY 2010 baseline fleet
into the MYs 2017-2025 model years. It also included the assumption that none of the vehicle
models had changes during this period. Projecting this future fleet is a process that is
necessarily uncertain. As with the MY 2008-based MY 2017-2025 reference fleet, NHTSA
and EPA relied on many sources of reputable information to make these projections.
1.4.2.1 On what data is the reference vehicle fleet based (using the MY2010 baseline)?
EPA and NHTSA have based the projection of total car and light truck sales on the
most recent projections available made by the Energy Information Administration (EIA).
EIA's Annual Energy Outlook (AEO) projects future energy production, consumption and
prices.6 EIA issued an "early release" version of AEO 2012 in January 2012. The complete
final version of AEO 2012 was released June 25, 2012, but by that time EPA/NHTSA had
already completed analyses supporting the final 2017-2025 standards using the interim data
release. Similar to the analyses supporting the MYs 2012-2016 rulemaking and for the 2008
based fleet projection, the agencies have used the Energy Information Administration's
(EIA's) National Energy Modeling System (NEMS) to estimate the future relative market
shares of passenger cars and light trucks. However, as explained above, NEMS shifts the
market toward passenger cars in order to ensure compliance with EISA's requirement that
CAFE standards cause the fleet to achieve 35 mpg by 2020. Because we use our market
projection as a baseline relative to which we measure the effects of new standards, and we
attempt to estimate the industry's ability to comply with new standards without changing
product mix (i.e.., we analyze the effects of the final rules assuming manufacturers will not
change fleet composition as a compliance strategy), using the Interim AEO 2012-projected
shift in passenger car market share as provided by EIA would cause the agencies to understate
the cost of achieving compliance through additional technology, alone. Therefore, for the
current analysis, the agencies developed a new projection of passenger car and light truck
sales shares by using NEMS to run scenarios from the Interim AEO 2012 reference case, after
1-40
-------
The Baseline and Reference Vehicle Fleets
first deactivating the above-mentioned sales-volume shifting methodology and holding post-
2017 CAFE standards constant at MY 2016 levels. Incorporating these changes reduced the
projected passenger car share of the light vehicle market by an average of about 5% during
2017-2025. As with the comparable exercise for the 2008 MY baseline fleet, this case is
referred to as the "Unforced Reference Case," and the values are shown below in Table 1-24.
Table 1-24 AEO 2012 Interim Unforced Reference Case Values used in the 2010 Market Fleet Projection
Model Year
2017
2018
2019
2020
2021
2022
2023
2024
2025
Cars
8,713,800
8,631,900
8,688,600
8,774,500
8,898,400
9,033,900
9,179,600
9,368,800
9,525,700
Trucks
7,098,300
6,973,500
6,973,500
6,855,700
6,831,700
6,853,300
6,827,600
6,878,200
6,929,100
Total Vehicles
15,812,100
15,605,400
15,662,100
15,630,200
15,730,100
15,887,200
16,007,200
16,247,000
16,454,800
In 2017, car and light truck sales are projected to be 8.7 and 7.1 million units,
respectively (compared to 8.4 and 7.4 million in the 2010 AEO projection). While the total
level of sales of 15.8 million units is similar to pre-2008 levels, the fraction of car sales in
2017 and beyond is projected to be higher than in the previous AEO projections.
In addition to a shift towards more car sales, sales of segments within both the car and
truck markets have also been changing and are expected to continue to change in the future.
The agencies also wanted to use the most updated information on Chrysler projections, as the
older NPRM projection conducted by CSM showed Chrysler sales to be very low in 2025.
The agencies agree with the Chrysler comments that the NPRM projections are most likely
outdated and too low with respect to Chrysler's market share. In order to reflect these
changes in fleet makeup, EPA and NHTSA used a custom long range forecast purchased from
LMC Automotive (formerly J.D. Powers Forecasting). J.D. Powers is a well-known industry
analyst. NHTSA and EPA decided to use the forecast from LMC Automotive (J.D. Powers
Forecasting) for MY2010-based market forecast for several reasons. First, Like CSM, LMC
Automotive uses a ground up approach (e.g., looking at the number of plants and capacity for
specific engines, transmissions, and vehicles) for their forecast, which the agencies believe is
a robust forecasting approach. Second, LMC Automotive allows us to publish their entire
forecast in the public domain. Third, the LMC Automotive forecast covered all the timeframe
of greatest relevance to this analysis (2017-2025 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
1-41
-------
The Baseline and Reference Vehicle Fleets
guide, such that the agencies could include only the vehicle types covered by the final
standards. And finally, it had a more updated projection of Chrysler sales.
LMC Automotive created a forecast that covered model years 2010-2025. Since the
agencies used this forecast to generate the reference fleet (i.e., the fleet expected to be sold
absent any increases in the stringency regulations after the 2016 model year), it is important
for the forecast to be independent of increases during 2017-2025 in the stringency of CAFE/
GHG standards. LMC Automotive does not use the CAFE or GHG standard as an input to
their model, and specifically had no assumption of increase in stringency in the 2017-2025
time frame.
The agencies combined the LMC Automotive forecast with data from other sources to
create the 2010 baseline reference fleet projections. This process is discussed in sections that
follow.
1.4.2.2 How do the agencies develop the 2010 baseline 2017-2025 reference vehicle
fleet?
The process of producing the MY 2010 baseline 2017-2025 reference fleet 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.
1.4.2.3 How was the 2010 baseline data merged with the LMC Automotive data?
EPA and NHTSA employed a different method from the method used in the NPRM
for mapping certification vehicles to LMC Automotive (LMC) vehicles. Merging the 2010
baseline data with the 2017-2025 LMC data required a thorough mapping of certification
vehicles to LMC vehicles by individual make and model. One challenge that the agencies
faced when determining a reference case fleet was that the sales data projected by LMC had
different market segmentation than the data contained in EPA's internal database. In order to
create a common segmentation between the two databases, the agencies performed a side-by-
side comparison of each vehicle model in both datasets, and created an additional "NEW
SEGMENT" modifier in the spreadsheet to map the two datasets. The reference fleet sales
based on the "NEW SEGMENT" 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.7 The spreadsheet provides
specific details on the sources and definitions for the data. The Excel file contains several
tabs. They are: "Data", "Data Tech Definitions", "SUM", "SUM Tech Definitions", "Truck
Vehicle Type Map", and "Car Vehicle Type Map". "Data" is the tab with the raw data.
"Data Tech Definitions" is the tab where each column is defined and its data source named.
"SUM" is the tab where the raw data is processed to be used in the OMEGA and Volpe
models. The "SUM" tab minus columns A-F and minus the Generic vehicles is the input file
for the models. The "Generic" manufacturer (shown in the "SUM" tab) is the sum of all
1-42
-------
The Baseline and Reference Vehicle Fleets
manufacturers and is calculated as a reference, and for data verification purposes. It is used to
validate the manufacturers' totals. It also gives an overview of the fleet.
Table 1-6 shows some of the unique models chosen from the "SUM" tab. A model is
made up of a unique combination of segment and vehicle type. The number of models is
determined by the number of unique segment and vehicle type combinations. These
combinations of segment and vehicle type (the vehicle type number is the same as the
technology package number) are determined by the technology packages discussed in the
EPA RIA. "SUM Tech Definitions" is the tab where the columns of the "SUM" tab are
defined.
Table 1-25 Models from the SUM Tab Model
Model
Car Like LargeSuv 14 Vehicle Type: 7
Car Like LargeSuv 14, V6 Vehicle Type: 8
Car Like Large Suv V6 Vehicle Type: 9
Car Like MidSizeSuv 14 Vehicle Type: 7
Car Like MidSizeSuv 14, V6 Vehicle Type: 8
Car Like MidSizeSuv V6 Vehicle Type: 9
Car Like SmallSuv V6 Vehicle Type: 10
Large Auto V6 Vehicle Type: 3
Large Auto V6 Vehicle Type: 4
Large Auto >=V6 Vehicle Type: 5
Large Auto >=V8 Vehicle Type: 6
MidSizeAuto 14 Vehicle Type: 2
MidSizeAuto 14, V6 Vehicle Type: 3
MidSize Auto V6 Vehicle Type: 4
MidSize Auto >=V6 Vehicle Type: 5
MidSize Auto V8 Vehicle Type: 6
In the combined EPA certification and LMC data, all 2010 vehicle models were
assumed to continue out to 2025, though their volumes changed in proportion to LMC
projections. Also, any new models expected to be introduced within the 2011-2025
timeframe are not included in the data. These volumes are reassigned to the existing models
to keep the overall fleet volume the same. All MYs 2017-2025 vehicles are mapped to the
existing vehicles by a process of mapping to manufacturer's future segment volumes. The
mappings are discussed in the next section. Further discussion of this limitation is discussed
below in section 1.4.2.4. The statistics of this fleet will be presented below since further
modifications were required to the volumes as the next section describes.
1-43
-------
The Baseline and Reference Vehicle Fleets
1.4.2.4 How were the LMC forecast and the AEO forecast used to project the future
fleet volumes?
As with the comparable step in the MY 2008 baseline 2017-2025 reference fleet
process, the next step in the agencies' generation of the reference fleet is one of the more
complicated steps to explain. First, the 2010 CAFE data was mapped to the LMC segments.
Second, the breakdown of segment volumes by manufacturer was compared between the
LMC and CAFE data sets. Third, a correction was applied for Class 2B vehicles (Large
Pickup Trucks) in the LMC 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 Early Release of the 2012 Annual Energy
Outlook (AEO).
The process started with mapping the LMC segments to the CAFE data. The process
was simple yet time consuming. The mapping required determining the LMC segment by
looking at each of the 1171 vehicles in the LMC quarter forecast, and labeling it in the "New
Segment" column of the new data spreadsheet. The segments were somewhat different from
the ones employed by CSM. LMC has 27 segments and CSM has 18 segments. Table 1-26
has both the LMC Segments and the CSM segments for reference. Table 1-27 shows some of
the Chrysler/Fiatr vehicles in the CAFE data with their "New Segment" identified.
Table 1-26 List of LMC Segments and CSM Segments
LMC Segments
Compact Conventional
Compact CUV
Compact MPV
Compact Premium Conventional
Compact Premium CUV
Compact Premium Sporty
Compact Sporty
Compact Utility
Large Conventional
Large Pickup
Large Premium Conventional
Large Premium Sporty
Large Premium Utility
Large Utility
Large Van
Midsize Conventional
Midsize CUV
Midsize Pickup
Midsize Premium Conventional
Midsize Premium CUV
Midsize Premium Sporty
Midsize Premium Utility
CSM Class
Full-Size Car
Full-Size CUV
Full-Size Pickup
Full-Size SUV
Full-Size Van
Luxury Car
Mid-Size Car
Mid-Size CUV
Mid-Size MAV
Mid-Size Pickup
Mid-Size SUV
Mid-Size Van
Mini Car
Small Car
Small CUV
Small MAV
Small SUV
Specialty Car
r Chrysler/Fiat is being used as an example throughout this section to make the example calculations easier to
follow.
1-44
-------
The Baseline and Reference Vehicle Fleets
Midsize Sporty
Midsize Utility
Midsize Van
Table 1-27 Example of Chrysler/Fiat vehicles being mapped to segments based on the LMC Forecast
Manufacturer
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Name Plate
Chrysler
Chrysler
Chrysler
Chrysler
Dodge
Dodge
Dodge
Dodge
Dodge
Dodge
Model
300 AWD
PT Cruiser
Sebring
Town & Country FWD
Caliber
Challenger
Charger
Dakota Pickup 2wd
Grand Caravan FWD
Journey 2wd
NEW SEGMENT
Large Conventional
Compact MPV
Midsize Conventional
Midsize Van
Compact Conventional
Midsize Sporty
Large Conventional
Midsize Pickup
Midsize Van
Midsize CUV
In this next step, segment volume by manufacturer was compared between the LMC
and CAFE data sets. This is necessary to determine if all of the segments a manufacturer will
produce in the future are currently represented by the 2010 CAFE data. Almost all the future
segments matched the current segments with the exception of some premium vs. standard
class vehicles. In cases where there was not a vehicle model in a premium class (such as
Compact Premium CUV) in the future, but there was a model in the standard class (Compact
CUV), the future premium class volume was added to the standard class volume. The same
thing was done if the opposite was true, i.e. if there was not a vehicle in a standard class (such
as Compact CUV) in the future, but there was one in the premium class (Compact Premium
CUV), the future standard class volume was added to the premium class volume. Table 1-28
shows the New Segments, the LMC 2010 Volumes, and the LMC 2018 Volumes for
Chrysler/Fiat. The Compact Premium Conventional, Compact Premium CUV, and Compact
Premium Sporty were not available from Chrysler/Fiat in 2010, but are available in 2018. As
mentioned, the volumes from all three of those premium segments were added to the standard
segments Compact Conventional, Compact CUV, and Compact Sporty in years were the
premium segments were produced.
Table 1-28 Example Chrysler/Fiat 2010 Volumes by Segment from the LMC Forecast
NEW SEGMENT
Compact Conventional
Compact CUV
Compact MPV
Compact Premium Conventional
Compact Premium CUV
Compact Premium Sporty
Compact Utility
Large Conventional
Large Pickup
Large Van
LMC 2010 Volume
45,082
54,514
9,440
-
-
-
166,492
112,513
199,652
-
LMC 2018 Volume
91,136
78,307
61,461
35,027
12,783
209
210,979
185,553
284,583
1-45
-------
The Baseline and Reference Vehicle Fleets
Midsize Conventional
Midsize CUV
Midsize Pickup
Midsize Premium Conventional
Midsize Premium CUV
Midsize Premium Sporty
Midsize Sporty
Midsize Utility
Midsize Van
Sub-Compact Conventional
89,508
48,577
13,047
-
-
392
36,791
93,352
215,598
-
88,007
91,880
27,141
9,309
12,476
3,014
-
154,401
155,408
97,342
A step that is related to the comparison step is the filtering of Class 2b vehicles from
the LMC forecast. LMC includes Class 2b 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 CAFE and GHG programs. Class 2b large pickup trucks, however, are not
regulated under the light-duty CAFE and GHG programs (rather under the medium- and
heavy-duty fuel efficiency and GHG programs, see 76 FR at 57120), and must therefore be
removed from the forecast. This is accomplished by a creating a multiplier for each
manufacturer's large pickup trucks and applying it to each manufacturer's large pickup truck
volume every model year in the LMC forecast; specifically, by taking a manufacturer's 2010
model year large pickup CAFE volume and dividing its 2010 model year large pickup LMC
volume. Table 1-29 shows the volumes and the resulting multiplier for Chrysler/Fiat, while
Table 1-30 shows the 2025 LMC volume, the multiplier and the result of applying the
multiplier to the original volume for Chrysler/Fiat.
Table 1-29 Example Values Used to Determine the Class 2b Truck Multiplier for Chrysler/Fiat
Manufacturer
Chrysler/Fiat
NEW SEGMENT
Large Pickup
LMC 2010
Volume
199,652
2010 CAFE
Volume
120,645
Truck
Multiplier
0.60
Table 1-30 Example Values Used to Determine Chrysler/Fiat's 2025 Truck Volume
Manufacturer
Chrysler/Fiat
NEW SEGMENT
Large Pickup
Original 2025
Volume
382,492
Truck
Multiplier
0.60
2025
Volume
after
Multiplier
231,131
After correcting for Class 2b vehicles being in the LMC forecast, it was time to create
individual manufacturer segment multipliers to be used with the individual 2010 CAFE
vehicle volumes to create projections for the future fleet. The individual manufacturer
1-46
-------
The Baseline and Reference Vehicle Fleets
segment multipliers are created by dividing each year of the LMC forecast's individual
manufacturer segment volume by the manufacturer's individual segment volume determined
using 2010 CAFE data. Table 1-31 has the 2010 CAFE Volume, the 2025 LMC large pickup
volume after Class 2b vehicles were removed, and the individual manufacturer volume for
large pickup trucks. The multiplier is the result of dividing the 2025 volume by the 2010
volume.
Table 1-31 Example Values Used to Determine Chrysler/Fiat 2025 Individual Large Pickup Multiplier
Manufacturer
Chrysler/Fiat
NEW SEGMENT
Large Pickup
2010 Cafe Volume
120,645
2025 Volume after Multiplier
231,131
Fiat/Chrysler Individual Large Pickup
Multiplierfor2025
192%
Now that the individual manufacturer segment multipliers are calculated, they can be
applied to each vehicle in the 2010 CAFE data. The segment multipliers are applied by
multiplying the 2010 CAFE volume for a vehicle by the multiplier for its manufacturer and
segment. Table 1-32 shows the 2010 CAFE volumes, the individual manufacturer segment
multipliers, and the result of multiplying the multiplier and the volume for 2025 project
volumes for many of Chrysler/Fiat's large pickup trucks.
Table 1-32 Example Applying the Individual Large Pickup Multiplier for Chrysler/Fiat
Manufacturer
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Model
Ram 1500 Pickup 2wd
Ram 1500 Pickup 2wd
Ram 1500 Pickup 2wd
Ram 1500 Pickup 2wd
Ram 1500 Pickup 2wd
Ram 1500 Pickup 4wd
Ram 1500 Pickup 4wd
Ram 1500 Pickup 4wd
Ram 1500 Pickup 4wd
NEW SEGMENT
Large Pickup
Large Pickup
Large Pickup
Large Pickup
Large Pickup
Large Pickup
Large Pickup
Large Pickup
Large Pickup
2010 CAFE Volume
23,686
938
3,029
16,505
7,698
1,162
51,417
15,498
712
Fiat/Chrysler
Individual Large
Pickup Multiplier
for 2025
192%
192%
192%
192%
192%
192%
192%
192%
192%
2025 Project Volume
Before AEO
Normalization
45,377
1,797
5,803
31,620
14,748
2,226
98,504
29,691
1,364
Normalizing to AEO 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 LMC 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 AEO are
used to determine a normalizing multiplier. Table 1-33 has the 2025 car and truck totals
1-47
-------
The Baseline and Reference Vehicle Fleets
before normalization, the 2025 AEO car and truck total, and the multipliers which are the
result of dividing the AEO totals by totals before normalization.
Table 1-33 Example 2025 AEO Truck and Car Multipliers
Vehicle Type
Trucks
Cars
2025 Total before Normalization
8,242,936
8,954,382
2025 AEO Total
6,929,100
9,525,700
AEO 2025
Normalizing
Multiplier
84%
106%
The final step in creating the reference volumes is applying the AEO multipliers. The
AEO multipliers are applied by vehicle type. Table 1-34 shows the normalized volume, the
AEO 2025 truck multiplier, and the final resulting volume for a number of Chrysler/Fiat
pickups.
Table 1-34 Example Applying the AEO Truck Multiplier to Chrysler/Fiat Pickups
Manufacturer
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Chrysler/Fiat
Model
Ram 1500 Pickup 2wd
Ram 1500 Pickup 2wd
Ram 1500 Pickup 2wd
Ram 1500 Pickup 2wd
Ram 1500 Pickup 2wd
Ram 1500 Pickup 4wd
Ram 1500 Pickup 4wd
Ram 1500 Pickup 4wd
Ram 1500 Pickup 4wd
Vehicle Type
Truck
Truck
Truck
Truck
Truck
Truck
Truck
Truck
Truck
2025 Project
Volume Before
AEO
Normalization
45,377
1,797
5,803
31,620
14,748
2,226
98,504
29,691
1,364
AEO 2025 Truck
Multiplier
84%
84%
84%
84%
84%
84%
84%
84%
84%
2025 Project
Volume with
AEO
Normalization
38,145
1,511
4,878
26,580
12,397
1,871
82,804
24,959
1,147
1.4.3 What are the sales volumes and characteristics of the MY 2010 based
reference fleet?
Table 1-35 and Table 1-37 below contain the sales volumes that result from the
process above for MY 2010 and 2017-2020.
Table 1-36 and Table 1-38 below contain the sales volumes that result from the process above
for MY 2021-2025.
Table 1-35 Vehicle Segment Volumes"
Reference Class
Segment
Large Auto
Actual and Projected Sales
Volume
2010
393,049
2017
567,514
2018
579,808
2019
598,784
2020
617,135
1-48
-------
The Baseline and Reference Vehicle Fleets
Mid-Size Auto
Compact Auto
Sub-Compact Auto
Large Pickup
Small Pickup
Large SUV
Mid-Size SUV
Small SUV
Mini Van
Cargo Van
2,189,552
1,894,017
1,615,536
1,201,518
74,780
2,066,629
1,058,340
113,716
565,527
17,516
3,446,643
2,561,669
2,258,243
1,747,062
39,095
3,259,969
1,068,111
148,142
686,492
29,160
3,413,476
2,525,760
2,231,633
1,723,045
39,793
3,208,284
1,036,455
143,413
674,803
28,929
3,523,692
2,524,658
2,161,935
1,773,581
49,185
3,157,778
1,058,492
142,957
641,731
29,308
3,577,767
2,537,591
2,169,551
1,757,204
55,481
3,086,726
1,037,464
142,894
618,567
29,821
a Volumes in this table are based on the pre-2011 NHTSA definition of Cars and Trucks.
Table 1-36 Vehicle Segment Volumes"
Reference Class
Segment
Large Auto
Mid-Size Auto
Compact Auto
Sub-Compact Auto
Large Pickup
Small Pickup
Large SUV
Mid-Size SUV
Small SUV
Mini Van
Cargo Van
Projected Sales Volume
2021
627,571
3,644,746
2,571,913
2,188,554
1,759,426
58,848
3,067,335
1,026,207
143,576
612,054
29,868
2022
641,252
3,684,993
2,613,050
2,236,339
1,761,341
62,556
3,064,546
1,040,034
145,165
607,502
30,422
2023
657,367
3,763,193
2,649,239
2,256,403
1,763,299
66,735
3,043,294
1,031,240
146,476
599,255
30,699
2024
665,152
3,819,396
2,709,562
2,334,855
1,770,423
71,587
3,049,618
1,047,527
148,201
600,002
30,678
2025
678,652
3,902,811
2,750,233
2,359,545
1,787,445
75,596
3,064,625
1,052,812
152,103
599,779
31,198
' Volumes in this table are based on the pre-2011 NHTSA definition of Cars and Trucks.
Table 1-37 2011+ NHTSA Car and Truck Definition Based Volumes
Vehicle Type
Cars
Trucks
Cars and Trucks
Actual and Projected Sales Volume
2010
7,176,330
4,013,850
11,190,180
2017
10,213,312
5,598,788
15,812,100
2018
10,088,966
5,516,434
15,605,400
2019
10,139,761
5,522,339
15,662,100
2020
10,194,353
5,435,847
15,630,200
1-49
-------
The Baseline and Reference Vehicle Fleets
Table 1-38 2011+ NHTSA Car and Truck Definition Based Volumes
Vehicle Type
Cars
Trucks
Cars and Trucks
Projected Sales Volume
2021
10,310,594
5,419,506
15,730,100
2022
10,455,061
5,432,139
15,887,200
2023
10,593,727
5,413,473
16,007,200
2024
10,811,530
5,435,470
16,247,000
2025
10,981,082
5,473,718
16,454,800
Table 1-40 and Table 1-40 below contain the sales volumes by manufacturer and
vehicle type for MY 2010 and 2017-2025. Tesla did not report any vehicle sales in 2010 so
their projected volume is zero. Spyker/Saab sold no vehicles under the Spyker brand in 2010
so their volume is also zero.
Table 1-39 NHTSA Car and Truck Definition Manufacturer Volumes
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
Chrysler/Fiat
Chrysler/Fiat
Daimler
Daimler
Ferrari
Ferrari
Ford
Ford
Geely
Geely
General Motors
General Motors
Honda
Honda
Hyundai
Hyundai
Kia
Vehicl
e 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
2010
Baseline
Sales
11,190,180
7,176,330
4,013,850
601
-
143,638
26,788
496,998
665,806
157,453
72,393
1,780
-
940,241
858,798
28,223
29,719
1,010,524
735,367
845,318
390,028
375,656
35,360
226,157
2017
Projected
Volume
11,190,180
10,213,312
5,598,788
634
-
320,634
106,150
728,817
774,065
252,820
99,125
1,878
-
1,348,543
1,035,400
60,422
35,087
1,652,946
1,213,192
1,122,558
536,998
865,069
131,912
345,314
2018
Projected
Volume
15,605,400
10,088,966
5,516,434
617
-
318,821
104,625
736,022
743,375
240,222
108,510
1,828
-
1,347,544
1,023,955
57,655
32,438
1,616,449
1,201,479
1,139,856
525,327
849,727
127,289
339,180
2019
Projected
Volume
15,662,100
10,139,761
5,522,339
620
-
327,091
105,104
769,256
749,206
245,807
108,294
1,836
-
1,341,628
1,016,328
60,338
33,299
1,611,415
1,217,167
1,147,055
527,814
857,497
122,193
328,872
2020
Projected
Volume
15,630,200
10,194,353
5,435,847
620
-
329,304
101,805
786,344
740,640
245,888
108,598
1,837
-
1,347,596
995,702
60,040
32,149
1,612,666
1,211,435
1,167,627
517,268
861,062
118,265
327,694
1-50
-------
The Baseline and Reference Vehicle Fleets
Kia
Lotus
Lotus
Mazda
Mazda
Mitsubishi
Mitsubishi
Nissan
Nissan
Porsche
Porsche
Spyker/Saab
Spyker/Saab
Subaru
Subaru
Suzuki
Suzuki
Tata/JLR
Tata/JLR
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
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
21,721
354
-
249,489
61,451
54,263
9,146
619,918
255,566
11,937
3,978
-
-
184,587
73,665
25,002
3,938
11,279
37,475
-
-
1,508,866
696,324
284,046
36,327
43,374
374
-
254,270
59,862
61,058
13,701
889,039
305,943
18,430
20,105
-
-
209,137
96,938
43,253
3,399
28,012
54,033
-
-
1,528,208
966,417
481,894
103,088
43,209
364
-
249,048
59,114
58,152
13,840
867,771
306,537
18,138
19,647
-
-
205,550
94,441
42,515
3,347
27,188
53,423
-
-
1,501,492
955,281
470,826
100,596
41,648
365
-
247,203
55,108
60,387
14,276
873,076
309,179
17,255
19,573
-
-
205,868
92,177
43,399
3,690
28,194
52,682
-
-
1,509,270
951,691
463,329
102,910
40,270
365
-
248,350
53,334
60,619
14,262
874,098
304,196
17,065
18,851
-
-
205,749
90,751
44,081
3,676
28,430
51,461
-
-
1,515,051
932,267
459,868
100,916
Table 1-40 NHTSA Car and Truck Definition Manufacturer Volumes
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
Chrysler/Fiat
Chrysler/Fiat
Daimler
Daimler
Ferrari
Ferrari
Ford
Ford
Vehicle Type
Both
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
2021
Projected
Volume
15,730,100
10,310,594
5,419,506
623
-
335,753
101,238
805,113
733,257
249,219
110,235
1,845
-
1,359,990
990,243
2022
Projected
Volume
15,887,200
10,455,061
5,432,139
626
-
341,613
100,345
828,656
735,937
251,461
112,133
1,853
-
1,377,947
990,827
2023
Projected
Volume
16,007,200
10,593,727
5,413,473
630
-
346,903
99,084
850,402
731,269
253,688
113,550
1,865
-
1,394,907
985,782
2024
Projected
Volume
16,247,000
10,811,530
5,435,470
634
-
357,948
101,174
877,751
722,213
258,742
116,867
1,878
-
1,418,568
991,767
2025
Projected
Volume
16,454,800
10,981,082
5,473,718
639
-
363,380
101,013
899,843
726,403
261 ,242
119,090
1,894
-
1,441,350
997,694
1-51
-------
The Baseline and Reference Vehicle Fleets
Geely
Geely
General Motors
General Motors
Honda
Honda
Hyundai
Hyundai
Kia
Kia
Lotus
Lotus
Mazda
Mazda
Mitsubishi
Mitsubishi
Nissan
Nissan
Porsche
Porsche
Spyker
Spyker
Subaru
Subaru
Suzuki
Suzuki
Tata/JLR
Tata/JLR
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
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
61,433
31,977
1,624,561
1,218,265
1,187,756
512,800
873,625
117,565
330,416
39,205
367
-
249,288
52,946
61,785
14,307
879,450
303,616
17,289
18,863
-
-
206,863
91,673
44,765
3,760
28,977
50,984
-
-
1,530,699
927,227
460,777
101,344
62,399
31,598
1,638,066
1,226,184
1,212,900
515,656
887,004
116,208
335,846
38,857
368
-
252,522
52,752
63,390
14,778
884,816
304,381
17,216
18,598
-
-
209,828
91,940
45,769
3,879
29,416
50,767
-
-
1,548,354
925,277
465,011
102,022
63,076
31,007
1,652,324
1,232,502
1,238,278
509,628
899,936
115,339
338,791
38,203
371
-
254,751
52,158
63,937
14,824
893,622
304,703
17,292
18,562
-
-
211,621
92,337
46,590
3,939
29,898
50,280
-
-
1,567,676
918,749
467,170
101,558
65,157
31,796
1,676,558
1,244,178
1,267,745
505,534
918,938
116,430
346,828
38,034
374
-
259,488
52,998
67,026
15,229
907,823
308,510
17,517
18,861
-
-
215,567
94,300
47,824
4,085
30,546
50,340
-
-
1,598,715
918,479
475,903
104,673
65,883
31,528
1,696,474
1,261,546
1,295,234
504,020
935,619
117,662
350,765
37,957
377
-
262,732
53,183
67,925
15,464
919,920
312,005
17,609
19,091
-
-
218,870
96,326
48,710
4,173
30,949
50,369
-
-
1 ,622,242
921,183
479,423
105,009
Table 1-41 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 2010 and 2025. The footprints are somewhat larger than in the 2008 based fleet
projection (Table 1-19).
Table 1-41 Production Weighted Foot Print Mean
Model
Average Footprint of all
Average Footprint
Average Footprint
1-52
-------
The Baseline and Reference Vehicle Fleets
Year
2010
2017
2018
2019
2020
2021
2022
2023
2024
2025
Vehicles
48.6
48.7
48.8
48.8
48.8
48.8
48.7
48.7
48.6
48.6
Cars
45.2
45.4
45.4
45.5
45.5
45.5
45.5
45.5
45.5
45.5
Trucks
54.5
54.9
54.9
55.0
55.0
55.0
55.0
55.0
54.9
55.0
Table 1-42 below shows the changes in engine cylinders over the model years. The
current assumptions show that engines will increase in size between 2010 and 2017 and then
remain relatively constant over the model years to which these final rules apply.
Table 1-42 Percentages of 4,6,8 Cylinder Engines by Model Year
Model
Year
2010
2017
2018
2019
2020
2021
2022
2023
2024
2025
Trucks
4
Cylinders
15.7%
13.9%
13.7%
13.6%
13.5%
13.4%
13.5%
13.5%
13.7%
13.6%
6
Cylinders
52.5%
50.2%
50.3%
50.0%
49.9%
49.8%
49.7%
49.6%
49.6%
49.5%
8
Cylinders
31.8%
35.9%
36.0%
36.4%
36.7%
36.8%
36.8%
36.9%
36.8%
36.8%
Cars
4
Cylinders
69.2%
66.3%
66.2%
65.7%
65.7%
65.7%
65.8%
65.7%
65.9%
65.9%
6
Cylinders
26.6%
29.0%
29.1%
29.6%
29.6%
29.6%
29.5%
29.5%
29.4%
29.4%
8
Cylinders
4.1%
4.7%
4.7%
4.7%
4.7%
4.7%
4.8%
4.8%
4.7%
4.8%
1-53
-------
The Baseline and Reference Vehicle Fleets
1.5 What are the differences in the sales volumes and characteristics of the MY 2008
based and the MY 2010 based reference fleets?
This section compares some of the differences between the fleet based on MY 2008
CAFE and the fleet based on MY 2010 CAFE data. As stated before, the 2008 fleet
projection is based on MY 2008 CAFE data, a long range forecast provided by CSM, and
interim AEO 2011. The 2010 fleet projection is based on MY 2010 CAFE, a long range
forecast provided by LMC Automotive, and interim AEO 2012.
Table 1-43, Table 1-44, Table 1-45 and Table 1-46 below contain the sales volume
differences between the two fleets, from subtracting the 2008 MY based fleet projection from
the 2010 MY based fleet projection.
The sales in MY 2010 are significantly lower (by 2,661,581 vehicles) than in MY
2008 (reflecting the continued economic recession, as noted earlier). The sales in MY 2010
are depressed but sales are expected to recover to their MY 2008 levels before 2017.
There is an increase in the number of large trucks, midsize autos, and large autos by
2025. There is also decreased volume in the remaining segment in 2025. These differences
are due to the LMC forecast and the newer AEO projection.
Table 1-43 Vehicle Segment Volumes Differences"
Reference Class
Segment
LargeAuto
MidSizeAuto
CompactAuto
SubCmpctAuto
LargePickup
SmallPickup
LargeSUV
MidSizeSUV
SmallSUV
MiniVan
CargoVan
Actual
Sales
Volume
2010-2008
-169,191
-909,375
-85,444
249,703
-380,708
-102,717
-717,320
-205,020
-171,639
-76,528
-93,342
Projected Sales Volume
2017
191,407
135,375
213,689
-199,979
232,443
-117,132
65,480
-290,644
-109
-68,070
-156,681
2018
223,040
123,068
200,367
-222,479
279,279
-118,139
58,183
-272,757
-6,520
-64,748
-170,305
2019
245,175
220,071
155,357
-327,273
390,391
-111,567
-20,090
-208,902
-11,718
-75,334
-172,666
2020
222,271
195,982
89,570
-383,799
371,009
-90,548
-116,518
-248,358
-19,783
-95,756
-189,807
a Volumes in this table are based on the pre-2011 NHTSA definition of Cars and Trucks.
1-54
-------
The Baseline and Reference Vehicle Fleets
Table 1-44 Vehicle Segment Volumes Differences"
Reference Class
Segment
LargeAuto
MidSizeAuto
CompactAuto
SubCmpctAuto
LargePickup
SmallPickup
LargeSUV
MidSizeSUV
SmallSUV
MiniVan
CargoVan
Projected Sales Volume
2021
247,379
202,630
50,936
-437,810
391,125
-91,275
-245,579
-255,033
-23,647
-117,024
-180,671
2022
282,957
136,730
20,851
-450,828
411,920
-84,582
-298,062
-243,210
-24,478
-131,480
-172,390
2023
294,695
70,660
16,313
-464,699
462,006
-84,580
-369,459
-237,048
-23,763
-141,530
-170,886
2024
308,979
67,900
-35,072
-461,206
498,672
-83,040
-426,255
-245,135
-24,990
-120,718
-166,222
2025
309,809
87,870
-92,836
-518,743
527,056
-79,242
-456,367
-252,550
-23,610
-126,477
-170,570
a Volumes in this table are based on the pre-2011 NHTSA definition of Cars and Trucks.
Table 1-45 2011+ NHTSA Car and Truck Definition Based Volumes Differences
Vehicle Type
Cars
Trucks
Cars and Trucks
Actual Sales
Volume
2010-2008
-1,054,238
-1,607,343
-2,661,581
Projected Sales Volume
2017
225,645
-219,867
5,778
2018
183,602
-154,612
28,990
2019
144,065
-60,623
83,442
2020
-97,209
-168,530
-265,739
Table 1-46 2011+ NHTSA Car and Truck Definition Based Volumes Differences
Vehicle Type
Cars
Trucks
Cars and Trucks
Projected Sales Volume
2021
-194,571
-264,396
-458,966
2022
-280,716
-271,857
-552,572
2023
-374,276
-274,013
-648,289
2024
-446,608
-240,479
-687,087
2025
-560,478
-235,181
-795,659
1-55
-------
The Baseline and Reference Vehicle Fleets
Table 1-47 and Table 1-48 below contain the differences in sales volumes by
manufacturer and vehicle type between the 2008 MY based fleet and the 2010 MY based
fleet. Table 1-48 shows that Chrysler/Fiat cars and trucks, Ford trucks, Hyundai cars, and
Porsche trucks are projected to have significant increases in volume in MY 2025, though
Table 1-48 also shows the market down overall in MY 2025 by 795,659 vehicles.
Table 1-47 NHTSA Car and Truck Definition Manufacturer Volumes Differences
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
Chrysler/Fiat
Chrysler/Fiat
Daimler
Daimler
Ferrari
Ferrari
Ford
Ford
Geely/Volvo
Geely/Volvo
GM
GM
HONDA
HONDA
HYUNDAI
HYUNDAI
Kia
Kia
Lotus
Lotus
Mazda
Mazda
Mitsubishi
Mitsubishi
Vehicle 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
2010-2008
Difference
in Sales
-2,661,581
-1,054,238
-1,607,343
-769
NA
-148,158
-34,536
-206,160
-290,986
-50,742
-6,742
330
NA
-16,458
44,604
-4,525
-35,930
-497,273
-852,024
-161,321
-115,112
37,787
-17,798
4,177
-37,751
102
NA
2,828
5,566
-31,095
-6,225
2017
Difference
in Volume
-4,616,142
225,645
-219,867
-401
NA
7,612
-31,903
310,054
364,363
-32,027
12,212
-4,798
NA
48,644
271,851
18,535
-53,147
290,185
-249,012
-32,042
-59,483
273,042
-20,973
23,270
-55,328
134
NA
730
8,074
-4,041
-23,931
2018
Difference
in Volume
28,990
183,602
-154,612
-434
NA
-4,118
-27,317
338,484
355,517
-36,187
24,859
-4,872
NA
36,077
275,126
15,468
-56,956
178,094
-272,597
1,769
-19,292
271,354
-24,172
26,810
-55,071
121
NA
-13,464
1,579
-5,519
-22,460
2019
Difference
in Volume
83,442
144,065
-60,623
-452
NA
-18,984
-26,269
377,567
382,759
-35,618
20,106
-4,958
NA
9,589
298,555
17,213
-58,276
106,390
-276,344
2,416
279
274,526
-33,449
13,993
-59,031
115
NA
-19,748
-2,386
-3,439
-21,178
2020
Difference
in Volume
-265,739
-97,209
-168,530
-414
NA
-28,638
-26,534
371,025
379,963
-45,101
15,679
-5,079
NA
-31,193
278,665
17,425
-60,854
81,911
-333,548
3,961
-7,821
262,779
-35,908
4,018
-56,265
99
NA
-21,728
-4,820
-4,461
-20,953
1-56
-------
The Baseline and Reference Vehicle Fleets
Nissan
Nissan
PORSCHE
PORSCHE
Spyker/Saab
Spyker/Saab
Subaru
Subaru
Suzuki
Suzuki
Tata/JLR
Tata/JLR
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
-97,951
-49,980
-6,972
-14,819
NA
NA
68,552
-8,881
-54,337
-31,381
1,683
-18,109
NA
NA
248,502
-254,812
-7,437
9,328
18,242
-138,995
-16,663
6,872
NA
NA
-14,975
18,696
-47,455
-18,710
-27,869
-3,546
NA
NA
-320,988
-364,094
-69,744
-25,731
18,093
-105,846
-17,306
7,646
NA
NA
-11,048
19,289
-47,417
-18,038
-29,034
-3,183
NA
NA
-332,689
-268,134
-69,210
-44,895
18,676
-89,380
-18,861
8,104
NA
NA
-11,227
19,345
-47,169
-17,002
-29,073
-5,172
NA
NA
-327,036
-190,413
-73,785
-43,981
-8,693
-93,673
-18,898
7,710
NA
NA
-17,717
18,293
-49,467
-16,999
-29,752
-4,752
NA
NA
-368,683
-222,037
-94,954
-45,784
Table 1-48 NHTSA Car and Truck Definition Manufacturer Volumes Differences
Manufacturers
All
All
All
Aston Martin
Aston Martin
BMW
BMW
Chrysler/Fiat
Chrysler/Fiat
Daimler
Daimler
Ferrari
Ferrari
Ford
Ford
Geely/JLR
Geely/JLR
GM
GM
Vehicle Type
Both
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
Cars
Trucks
2021
Difference
in Volume
-458,966
-194,571
-264,396
-435
NA
-23,345
-27,486
384,100
384,644
-51,159
10,786
-5,214
NA
-41,627
276,062
19,665
-60,749
94,541
-346,012
2022
Difference
in Volume
-552,572
-280,716
-271,857
-423
NA
-18,421
-28,554
404,483
372,929
-53,277
11,198
-5,285
NA
-37,274
276,561
20,713
-60,914
130,413
-352,372
2023
Difference
in Volume
-648,289
-374,276
-274,013
-411
NA
-13,658
-28,437
426,520
370,205
-58,819
8,235
-5,362
NA
-79,890
285,777
21,045
-65,833
155,505
-373,993
2024
Difference
in Volume
-687,087
-446,608
-240,479
-507
NA
-30,245
-45,351
451,734
377,251
-73,595
9,783
-5,563
NA
-85,102
302,913
22,696
-67,385
182,961
-392,627
2025
Difference
in Volume
-795,659
-560,478
-235,181
-543
NA
-41,876
-44,396
463,364
394,641
-79,477
18,023
-5,764
NA
-98,759
313,218
23,295
-69,579
172,466
-412,390
1-57
-------
The Baseline and Reference Vehicle Fleets
HONDA
HONDA
HYUNDAI
HYUNDAI
Kia
Kia
Lotus
Lotus
Mazda
Mazda
Mitsubishi
Mitsubishi
Nissan
Nissan
PORSCHE
PORSCHE
Spyker/Saab
Spyker/Saab
Subaru
Subaru
Suzuki
Suzuki
Tata/JLR
Tata/JLR
Tesla
Tesla
Toyota
Toyota
Volkswagen
Volkswagen
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
-11,124
-23,116
260,270
-38,901
-903
-56,227
89
NA
-25,452
-6,281
-4,066
-21,002
-33,179
-104,413
-19,186
7,621
NA
NA
-23,917
18,900
-50,960
-17,007
-29,700
-7,169
NA
NA
-373,007
-288,312
-124,830
-47,390
-24,604
-23,579
259,040
-41,285
-3,256
-55,837
78
NA
-28,628
-7,555
-3,871
-20,449
-52,631
-107,502
-19,391
7,213
NA
NA
-28,785
19,204
-51,830
-16,855
-29,933
-7,823
NA
NA
-437,723
-309,775
-128,303
-44,728
-27,286
-27,270
265,628
-45,850
-3,955
-57,485
72
NA
-42,159
-9,808
-3,743
-20,645
-60,718
-112,418
-19,701
7,192
NA
NA
-29,991
19,315
-52,673
-16,864
-30,741
-8,585
NA
NA
-469,316
-306,231
-129,579
-52,369
-40,106
-31,460
261,228
-49,662
-5,054
-58,085
66
NA
-41,126
-8,973
-3,702
-20,772
-74,948
-113,707
-21,987
7,452
NA
NA
-32,716
20,158
-52,623
-17,077
-33,182
-7,641
NA
NA
-481,813
-289,534
-129,433
-52,266
-45,087
-53,677
258,369
-50,474
-12,018
-59,696
61
NA
-44,072
-8,185
-5,380
-20,923
-94,855
-114,449
-23,087
7,872
NA
NA
-38,100
21,604
-54,444
-17,201
-34,469
-6,436
NA
NA
-485,811
-288,833
-150,740
-49,275
Table 1-49 shows the difference in footprint distributions between the 2010 based fleet
projection and the 2008 based fleet projection. The differences between MYs 2010 and 2008
are small and are just the result of the manufacturers' product mix in those model years. MY
2025 shows an increase in both the average truck and average car footprints. This is due to
the increased number of large cars and large trucks forecast in the 2010 based fleet projection.
Also, in several MYs, the change in the average footprint of all vehicles is outside the range
between the changes in the corresponding car and truck fleets. This is due to production
weighting. 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 effect when taking the difference of the averages.
1-58
-------
The Baseline and Reference Vehicle Fleets
Table 1-49 Production Weighted Foot Print Mean Difference*
Model
Year
2010-2008
2017
2018
2019
2020
2021
2022
2023
2024
2025
Average Footprint of all
Vehicles
48.6 -48. 8 = -0.2
48.7-48.0 = 0.7
48.8-47.9 = 0.9
48.8-47.8 = 1.0
48.8-47.8 = 1.0
48.8-47.7 = 1.0
48.7-47.7 = 1.0
48.7-47.7 = 1.0
48.6-47.5 = 1.1
48.6-47.5 = 1.1
Average Footprint
Cars
45.2-45.2 = 0.0
45.4-44.6 = 0.8
45.4.44.6 = 0.8
45.5-44.6 = 0.9
45.5.44.6 = 0.9
45.5-44.6 = 0.9
45.5-44.6 = 0.9
45.5-44.6 = 0.9
45.5-44.6 = 0.9
45.5-44.6 = 0.9
Average Footprint
Trucks
54.5-53.9 = 0.6
54.9-53.8 = 1.1
54.9-53.7 = 1.2
55.0-53.6 = 1.4
55.0-53.7 = 1.3
55.0-53.6 = 1.4
55.0-53.6 = 1.4
55.0-53.5 = 1.5
54.9-53.3 = 1.6
55.0-53.3 = 1.7
*Note: This table is the difference calculated from Table 1-19 and Table 1-41.
Table 1-50 shows the difference in engine cylinders distribution between the 2010 MY
based fleet and the 2008 MY based fleet. MY 2010 has fewer vehicles with 6 cylinder
engines. Fewer 6 cylinders in the baseline fleet along with vehicle mix changes results in
more 4 and 8 cylinder engines in trucks and more 4 cylinder cars by 2025.
Table 1-50 Differences in Percentages of 4,6,8 Cylinder Engines by Model Year
Model
Year
2010-2008
2017
2018
2019
2020
2021
2022
2023
2024
2025
Trucks
4
Cylinders
5.40%
3.00%
3.10%
3.20%
3.20%
3.10%
3.20%
3.20%
3.20%
3.10%
6
Cylinders
-3.90%
-13.50%
-14.20%
-15.50%
-15.70%
-16.50%
-17.00%
-18.10%
-18.50%
-18.70%
8
Cylinders
-1.50%
10.50%
11.20%
12.30%
12.60%
13.40%
13.80%
14.90%
15.40%
15.50%
Cars
4
Cylinders
12.30%
5.70%
5.50%
5.00%
5.40%
5.10%
4.70%
4.80%
4.90%
4.80%
6
Cylinders
-11.20%
-5.50%
-5.30%
-4.70%
-5.10%
-4.80%
-4.70%
-4.80%
-4.70%
-4.60%
8
Cylinders
-1.20%
-0.30%
-0.30%
-0.30%
-0.30%
-0.20%
0.00%
0.00%
-0.10%
0.00%
1-59
-------
The 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; DOT/NHTSA's CAFE Compliance and Effects
Modeling System (commonly known as the "Volpe Model") and input and output sheets are
available at http://www.nhtsa.gov/fuel-economy. It is also available in the docket (Docket
EPA-HQ-O AR-2010-0799)
9 _
http://www.nhtsa.gov/Laws+&+Regulations/CAFE+-
+Fuel+Economy/CAFE+Compliance+and+Effects+Modeling+System:+The+Volpe+Model
3 Department of Energy, Energy Information Administration, Annual Energy Outlook (AEO)
2011, Early Release. Available at http://www.eia.gov/forecasts/aeo/.
4 The baseline Excel file ("2008-2025 Production Summary Data _Definitions Docket
08_27_2009") is available in the docket (Docket EPA-HQ-O AR-2010-0799).
5 http://www.nhtsa.gov/Laws+&+Regulations/CAFE+-
+Fuel+Economy/CAFE+Compliance+and+Effects+Modeling+System:+The+Volpe+Model
6 Department of Energy, Energy Information Administration, Annual Energy Outlook (AEO)
20112012, Early Release. Available at http://www.eia.gov/forecasts/aeo/ (last accessed Aug.
15, 2011 April 9, 2012).
7 The baseline Excel file ("2010-2025 Production Summary Data_Defmitions Docket
05.01.2012") is available in the docket (Docket EPA-HQ-O AR-2010-0799).
1-60
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Chapter 2: What are the Attribute-Based Curves the Agencies
are Adopting, and How Were They Developed?
2.1 Why are standards attribute-based and defined by a mathematical function?
As in the MYs 2012-2016 CAFE/GHG rules, and as NHTSA did in the MY 2011
CAFE rule, NHTSA and EPA are promulgating attribute-based CAFE and CC>2 standards that
are defined by a mathematical function. EPCA, as amended by EISA, expressly requires that
CAFE standards for passenger cars and light trucks be based on one or more vehicle attributes
related to fuel economy, and be expressed in the form of a mathematical function.1 The CAA
has no such requirement, although such an approach is permissible under section 202 (a) and
EPA has used the attribute-based approach in issuing standards under both section 202 (a) and
under analogous provisions of the CAA (e.g., criteria pollutant standards for non-road diesel
engines using engine size as the attribute,2 in the recent GHG standards for heavy duty
pickups and vans using a work factor attribute,3 and in the MYs 2012-2016 GHG rule which
used vehicle footprint as the attribute). Public comments on the MYs 2012-2016 rulemaking
widely supported attribute-based standards for both agencies' standards. Comments received
on the MY 2017 and later proposal also generally supported an attribute-based standard, as
further discussed in section 2.2.
Under an attribute-based standard, every vehicle model has a performance target (fuel
economy and CC>2 emissions for CAFE and CC>2 emissions standards, respectively), the level
of which depends on the vehicle's attribute (for this rule, footprint, as discussed below). The
manufacturers' fleet average performance is determined by the production-weighted51 average
(for CAFE, harmonic average) of those targets.
The agencies believe that an attribute-based standard is preferable to a single-industry-
wide average standard in the context of CAFE and CC>2 standards for several reasons. First, if
the shape is chosen properly, every manufacturer is more likely to be required to continue
adding more fuel efficient technology each year across their fleet, because the stringency of
the compliance obligation will depend on the particular product mix of each manufacturer.
Therefore a maximum feasible attribute-based standard will tend to require greater fuel
savings and CO2 emissions reductions overall than would a maximum feasible flat standard
(that is, a single mpg or CC>2 level applicable to every manufacturer).
Second, depending on the attribute, attribute-based standards reduce the incentive for
manufacturers to respond to CAFE and CC>2 standards in ways harmful to safety. Because
a Production for sale in the United States.
b The 2002 NAS Report described at length and quantified the potential safety problem with average fuel
economy standards that specify a single numerical requirement for the entire industry. See 2002 NAS Report at
5, finding 12. Ensuing analyses, including by NHTSA, support the fundamental conclusion that standards
structured to minimize incentives to downsize all but the largest vehicles will tend to produce better safety
outcomes than flat standards.
2-1
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
each vehicle model has its own target (based on the attribute chosen), properly fitted attribute-
based standards provide little, if any, incentive to build smaller vehicles simply to meet a
fleet-wide average, because the smaller vehicles will be subject to more stringent compliance
targets.0
Third, attribute-based standards provide a more equitable regulatory framework for
different vehicle manufacturers.d A single industry-wide average standard imposes
disproportionate cost burdens and compliance difficulties on the manufacturers that need to
change their product plans to meet the standards, and puts no obligation on those
manufacturers that have no need to change their plans. As discussed above, attribute-based
standards help to spread the regulatory cost burden for fuel economy more broadly across all
of the vehicle manufacturers within the industry.
Fourth, attribute-based standards better respect economic conditions and consumer
choice, as compared to single-value standards. A flat, or single value, standard encourages a
certain vehicle size fleet mix by creating incentives for manufacturers to use vehicle
downsizing as a compliance strategy. Under a footprint-based standard, manufacturers have
greater incentive (compared to under a flat standard) to invest in technologies that improve
the fuel economy of the vehicles they sell rather than shifting product mix, because reducing
the size of the vehicle is generally a less viable compliance strategy given that smaller
vehicles have more stringent regulatory targets.
2.2 What attribute are the agencies adopting, and why?
As in the MYs 2012-2016 CAFE/GHG rules, and as NHTSA did in the MY 2011
CAFE rule, NHTSA and EPA are promulgating CAFE and CC>2 standards that are based on
vehicle footprint, which has an observable correlation to fuel economy and emissions. There
are several policy and technical reasons why NHTSA and EPA believe that footprint is the
most appropriate attribute on which to base the standards for the vehicles covered by this
rulemaking, even though some other light-duty vehicle attributes (notably curb weight) are
better correlated to fuel economy and emissions.
First, in the agencies' judgment, from the standpoint of vehicle safety, it is important
that the CAFE and CC>2 standards be set in a way that does not encourage manufacturers to
respond by selling vehicles that are less safe. NHTSA's research of historical crash data has
found that reductions in vehicle size and reductions in the mass of lighter vehicles tend to
compromise overall highway safety, while reductions in the mass of heavier vehicles tend to
improve overall highway safety. If footprint-based standards are defined in a way that creates
relatively uniform burden for compliance for vehicles of all sizes, then footprint-based
standards will not incentivize manufacturers to downsize their fleets as a strategy for
0 Assuming that the attribute is related to vehicle size.
d Mat 4-5, finding 10.
2-2
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
compliance which could compromise societal safety, or to upsize their fleets which might
reduce the program's fuel savings and GHG emission reduction benefits. Footprint-based
standards also enable manufacturers to apply weight-efficient materials and designs to their
vehicles while maintaining footprint, as an effective means to improve fuel economy and
reduce GHG emissions. On the other hand, depending on their design, weight-based
standards can create disincentives for manufacturers to apply weight-efficient materials and
designs. This is because weight-based standards would become more stringent as vehicle
mass is reduced. The agencies discuss mass reduction and its relation to safety in more detail
in Preamble section II.G.
Further, although we recognize that weight is better correlated with fuel economy and
CC>2 emissions than is footprint, we continue to believe that there is less risk of "gaming"
(changing the attribute(s) to achieve a more favorable target) by increasing footprint under
footprint-based standards than by increasing vehicle mass under weight-based standards—it is
relatively easy for a manufacturer to add enough weight to a vehicle to decrease its applicable
fuel economy target a significant amount, as compared to increasing vehicle footprint. We
also continue to agree with concerns raised in 2008 by some commenters on the MY 2011
CAFE rulemaking that there would be greater potential for gaming under multi-attribute
standards, such as those that also depend on weight, torque, power, towing capability, and/or
off-road capability. The agencies agree with the assessment first presented in NHTSA's MY
2011 CAFE final rule4 that the possibility of gaming an attribute-based standard is lowest
with footprint-based standards, as opposed to weight-based or multi-attribute-based standards.
Specifically, standards that incorporate weight, torque, power, towing capability, and/or off-
road capability in addition to footprint would not only be more complex, but by providing
degrees of freedom with respect to more easily-adjusted attributes, they could make it less
certain that the future fleet would actually achieve the average fuel economy and CO2
reduction levels projected by the agencies.6 This is not to say that a footprint-based system
will eliminate gaming, or that a footprint-based system will eliminate the possibility that
manufacturers will change vehicles in ways that compromise occupant protection. In the
agencies' judgment, footprint-based standards achieved the best balance among affected
considerations.
The agencies recognize that based on economic and consumer demand factors that are
external to this rule, the distribution of footprints in the future may be different (either smaller
or larger) than what is projected in this rule. However, the agencies continue to believe that
there will not be significant shifts in this distribution as a direct consequence of this rule. We
note that comments by CBD, ACEEE, and NACAA referenced a 2011 study by Whitefoot
and Skerlos, "Design incentives to increase vehicle size created from the U.S. footprint-based
fuel economy standards." This study concluded that the proposed MY 2014 standards
eHowever, for heavy-duty pickups and vans not covered by today's standards, the agencies determined that use
of footprint and work factor as attributes for heavy duty pickup and van GHG and fuel consumption standards
could reasonably avoid excessive risk of gaming. See 76 FR 57106, 57161-62 (Sept. 15, 2011)
f Available at Docket ID: EPA-HQ-OAR-2010-0799.
2-3
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
"create an incentive to increase vehicle size except when consumer preference for vehicle size
is near its lower bound and preference for acceleration is near its upper bound."g The
commenters who cited this study generally did so as part of arguments in favor of flatter
standards (i.e., curves that are flatter across the range of footprints) for MYs 2017-2025.
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 rules regarding the
baseline fleet, the cost and efficacy of potential future technologies, and the relationship
between vehicle footprint and fuel economy. Were the agencies to use the Whitefoot and
Skerlos methodology (e.g., methods to simulate manufacturers' potential decisions to increase
vehicle footprint) with the actual inputs to the MYs 2012-2016 rules, the agencies would
likely obtain different findings. Underlining the potential uncertainty, considering a range of
scenarios, the authors obtained a wide range of results in their analyses. The agencies discuss
this study more fully in the Section II of the preamble, the NHTSA RIA, and the EPA
response to comments document.
The agencies also recognize that some international attribute-based standards use
attributes other than footprint and that there could be benefits for a number of manufacturers
if there was greater international harmonization of fuel economy and GHG standards for light-
duty vehicles, but this is largely a question of how stringent standards are and how they are
tested and enforced. It is entirely possible that footprint-based and weight-based systems can
coexist internationally and not present an undue burden for manufacturers if they are carefully
crafted. Different countries or regions may find different attributes appropriate for basing
standards, depending on the particular challenges they face—from fuel prices, to family size
and land use, to safety concerns, to fleet composition and consumer preference, to other
environmental challenges besides climate change. The agencies anticipate working more
closely with other countries and regions in the future to consider how to address these issues
in a way that least burdens manufacturers while respecting each country's need to meet its
own particular challenges.
In the proposal, the agencies found that footprint was the most appropriate attribute
upon which to base the proposed standards. Recognizing strong public interest in this issue,
the agencies sought comment on whether a different attribute or combination of attributes
should be considered in setting standards for the final rule. The agencies specifically
requested that the commenters address the concerns raised in the proposal regarding the use of
other attributes, and explain how standards should be developed using the other attribute(s) in
a way that contributes more to fuel savings and CO2 reductions than the footprint-based
standards, without compromising safety.
The agencies received several comments regarding the attribute(s) upon which new
CAFE and GHG standards should be based. NADAh and the Consumer Federation of
8Ibid, page 410
h NAD A, Docket No. NHTSA-2010-0131-0261, at 11.
2-4
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
America (CFA)1 expressed support for attribute-based standards, generally, indicating that
such standards accommodate consumer preferences, level the playing field between
manufacturers, and remove the incentive to push consumers into smaller vehicles. Many
commenters, including automobile manufacturers, NGOs, trade associations and parts
suppliers (e.g.. General Motors,J Ford, American Chemistry Council, Alliance of
Automobile Manufacturers,™ International Council on Clean Transportation,11 Insurance
Institute for Highway Safety,0 Society of the Plastics Industry,15 Aluminum Association,11
Motor and Equipment Manufacturers Association/ and others) expressed support for the
continued use of vehicle footprint as the attribute upon which to base CAFE and CO2
standards, citing advantages similar to those mentioned by NADA and CFA. Conversely, the
Institute for Policy Integrity (IPI) at the New York University School of Law questioned
whether non-attribute-based (flat) or an alternative attribute basis would be preferable to
footprint-based standards as a means to increase benefits, improve safety, reduce "gaming,"
and/or equitably distribute compliance obligations.8 IPI argued that, even under flat standards,
credit trading provisions would serve to level the playing field between manufacturers. IPI
acknowledged that NHTSA, unlike EPA, is required to promulgate attribute-based standards,
and agreed that a footprint-based system could have much less risk of gaming than a weight-
based system. IPI suggested that the agencies consider a range of options, including a fuel-
based system, and select the approach that maximizes net benefits. Ferrari and BMW
suggested that the agencies consider weight-based standards, citing the closer correlation
between fuel economy and footprint, and BMW further suggested that weight-based standards
might facilitate international harmonization (i.e., between U.S. standards and related standards
in other countries).1 Porsche commented that the footprint attribute is not well suited for
manufacturers of high performance vehicles with a small footprint."
Regarding the comments from IPI, as IPI appears to acknowledge, EPCA/EISA
expressly requires that CAFE standards be attribute-based and defined in terms of
mathematical functions. Also, NHTSA has, in fact, considered and reconsidered options
other than footprint, over the course of multiple CAFE rulemakings conducted throughout the
past decade. When first contemplating attribute-based systems, NHTSA considered attributes
such as weight, "shadow" (overall area), footprint, power, torque, and towing capacity.
NHTSA also considered approaches that would combine two or potentially more than two
1 CFA, Docket No. EPA-HQ-OAR-2010-0799-9419at 8, 44.
J GM, Docket No. NHTSA-2010-0131-0236, at 2.
kFord, Docket No. NHTSA-2010-0131-0235, at 8.
1ACC, Docket No. EPA-HQ-OAR-2010-0799-9517at 2.
m Alliance, Docket No. NHTSA-2010-0131-0262, at 85.
"ICCT, Docket No. NHTSA-2010-0131-0258, at 48.
0IIHS, Docket No. NHTSA-2010-0131-0222, at 1.
p SPI, Docket No. EPA-HQ-OAR-2010-0799-9492, at 4.
q Aluminum Association, Docket No. NHTSA-2010-0131-0226, at 1.
r MEMA, Docket No. EPA-HQ-OAR-2010-0799-9478], at 1.
s IPI, Docket No. EPA-HQ-OAR-2010-0799-11485at 13-15.
1 BMW, Docket No. NHTSA-2010-0131-0250, at 3.
u Porsche, EPA-HQ-OAR-2010-0799-9264
2-5
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
such attributes. To date, every time NHTSA (more recently, with EPA) has reconsidered
options, the agency has concluded that a properly designed footprint-based approach provides
the best means of achieving the basic policy goals (i.e., better balancing compliance burdens
among full-line and limited-line manufacturers and reducing incentives for manufacturers to
respond to standards by reducing vehicle size in ways that could compromise overall highway
safety) involved in applying an attribute-based standards, and at the same time structuring
footprint-based standards in a way that furthers the energy and environmental policy goals of
EPCA and the CAA by controlling incentives to increase vehicle size in ways that could
increase fuel consumption and GHG emissions/ In response to IPI's suggestion to use fuel-
based standards as a type of attribute, although neither NHTSA nor EPA have presented
quantitative analysis of standards that differentiate between fuel type for light-duty vehicles,
such standards would effectively use fuel type to identify different subclasses of vehicles, thus
requiring mathematical functions—not addressed by IPI's comments—to recombine these
fuel types into regulated classes.w Insofar as EPCA/EISA already specifies how different fuel
types are to be treated for purposes of calculating fuel economy and CAFE levels, and
moreover, insofar as the EISA revisions to EPCA removed NHTSA's previously-clear
authority to set separate CAFE standards for different classes of light trucks, using fuel type
to further differentiate subclasses of vehicles could conflict with the intent, and possibly the
letter, of NHTSA's governing statute. Finally, in the agencies' judgment, while regarding
IPI's suggestion that the agencies select the attribute-based approach that maximizes net
benefits may have merit, net benefits are but one of many considerations which lead to the
setting of the standard. Also, such an undertaking would be impracticable at this time,
considering that the mathematical forms applied under each attribute-based approach would
also need to be specified, and that the agencies lack methods to reliably quantify the relative
potential for induced changes in vehicle attributes.
Regarding Ferrari's and BMW's comments, as stated previously, in the agencies'
judgment, footprint-based standards (a) discourage vehicle downsizing that might
compromise occupant protection, (b) encourage the application of technology, including
weight-efficient materials (e.g., high-strength steel, aluminum, magnesium, composites, etc.),
and (c) are less susceptible than standards based on other attributes to "gaming" that could
lead to less-than-projected energy and environmental benefits. It is also important to note that
there are many differences between both the standards and the on-road light-duty vehicle
v See 71 FR 17566, at!7595-17596 (April 6, 2006); 74 FR 14196, at!4359 (March 30, 2009); 75 FR 25324 at
25333 (May 7, 2010).
w The agencies did adopt separate standards for gasoline and diesel heavy-duty pickups and vans based on
technological differences between gasoline and diesel engines. See 76 FR at 57163-65. However, the agencies
stated that "standards that do not distinguish between fuel types are generally preferable where technological and
market-based reasons do not strongly argue otherwise. These technological differences exist presently between
gasoline and diesel engines for GHGs ... The agencies emphasize, however, that they are not committed to
perpetuating separate GHG standards for gasoline and diesel heavy-duty vehicles and engines, and expect to
reexamine the need for separate gasoline/diesel standards in the next rulemaking." 76 FR at 57165. IPI did not
suggest that there were any such technological distinctions justifying separate fuel-based attributes for light duty
vehicles, and the agencies note that EPCA/EISA already specifies how different fuels are to be treated for
purposes of CAFE
2-6
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
fleets in Europe and the United States. The stringency of standards, independent of the
attribute used, is another factor that influences harmonization. While the agencies agree that
international harmonization of test procedures, calculation methods, and/or standards could be
a laudable goal, again, harmonization is not simply a function of the attribute upon which the
standards are based. Given the differences in the on-road fleet (including vehicle
classification and use), in fuel composition and availability, in regional consumer preferences
for different vehicle characteristics, in other vehicle regulations besides for fuel economy/CO2
emissions, it would not necessarily be expected that the CAFE and GHG emission standards
would align with standards of other countries. Thus, the agencies continue to judge vehicle
footprint to be a preferable attribute for the same reasons enumerated in the proposal and
reiterated above.
Finally, as explained in section III.B.6 and documented in section III.D.6 below, EPA
agrees with Porsche that the MY 2017 GHG standards, and the GHG standards for the
immediately succeeding model years, pose special challenges of feasibility and (especially)
lead time for intermediate volume manufacturers, in particular for limited-line manufacturers
of smaller footprint, high performance passenger cars. It is for this reason that EPA has
provided additional lead time to these manufacturers. NHTSA, however, is providing no such
additional lead time. Under EISA/EPCA, manufacturers continue—as since the 1970s—to
have the option of paying civil penalties in lieu of achieving compliance with the standards.
2.3 What mathematical functions have the agencies previously used, and why?
2.3.1 NHTSA in MY 2008 and MY 2011 CAFE (constrained logistic)
For the MY 2011 CAFE rule, NHTSA estimated fuel economy levels after
normalization for differences in technology, but did not make adjustments to reflect other
vehicle attributes (e.g., power-to-weight ratios).x Starting with the technology adjusted
passenger car and light truck fleets, NHTSA used minimum absolute deviation (MAD)
regression without sales weighting to fit a logistic form as a starting point to develop
mathematical functions defining the standards. NHTSA then identified footprints at which to
apply minimum and maximum values (rather than letting the standards extend without limit)
and transposed these functions vertically (i.e., on a gpm basis, uniformly downward) to
produce the promulgated standards. In the preceding rule, for MYs 2008-2011 light truck
standards, NHTSA examined a range of potential functional forms, and concluded that,
compared to other considered forms, the constrained logistic form provided the expected and
appropriate trend (decreasing fuel economy as footprint increases), but avoided creating
"kinks" the agency was concerned would provide distortionary incentives for vehicles with
neighboring footprints/
xSee 74 FR 14196, 14363-14370 (Mar. 30, 2009) for NHTSA discussion of curve fitting in the MY 2011 CAFE
final rule.
y See 71 FR 17556, 17609-17613 (Apr. 6, 2006) for NHTSA discussion of "kinks" in the MYs 2008-2011 light
truck CAFE final rule (there described as "edge effects"). A "kink," as used here, is a portion of the curve where
a small change in footprint results in a disproportionally large change in stringency.
2-7
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
2.3.2 MYs 2012-2016 Light Duty GHG/CAFE (constrained/piecewise linear)
For the MYs 2012-2016 rules, NHTSA and EPA re-evaluated potential methods for
specifying mathematical functions to define fuel economy and GHG standards. The agencies
concluded that the constrained logistic form, if applied to post-MY 2011 standards, would
likely contain a steep mid-section that would provide undue incentive to increase the footprint
of midsize passenger cars.5 The agencies judged that a range of methods to fit the curves
would be reasonable, and used a minimum absolute deviation (MAD) regression without sales
weighting on a technology-adjusted car and light truck fleet to fit a linear equation. This
equation was used as a starting point to develop mathematical functions defining the standards
as discussed above. The agencies then identified footprints at which to apply minimum and
maximum values (rather than letting the standards extend without limit) and transposed these
constrained/piecewise linear functions vertically (i.e., on a gpm or CC>2 basis, uniformly
downward) to produce the fleetwide fuel economy and CO2 emission levels for cars and light
trucks described in the final rule.6
2.3.3 How have the agencies defined the mathematical functions for the MYs 2017-
2025 standards, and why?
By requiring NHTSA to set CAFE standards that are attribute-based and defined by a
mathematical function, NHTSA interprets Congress as intending that the post-EISA standards
to be data-driven - a mathematical function defining the standards, in order to be "attribute-
based," should reflect the observed relationship in the data between the attribute chosen and
fuel economy.2 EPA is also setting attribute-based CO2 standards defined by similar
mathematical functions, for the reasonable technical and policy grounds discussed below and
in section II of the preamble to the rule, and to harmonize with the CAFE standards.
The relationship between fuel economy (and GHG emissions) and footprint, though
directionally clear (i.e., fuel economy tends to decrease and CO2 emissions tend to increase
with increasing footprint), is theoretically vague and quantitatively uncertain; in other words,
not so precise as to a priori yield only a single possible curve.aa There is thus a range of
legitimate options open to the agencies in developing curve shapes. The agencies may of
course consider statutory objectives in choosing among the many reasonable alternatives since
the statutes do not dictate a particular mathematical function for curve shape. For example,
curve shapes that might have some theoretical basis could lead to perverse outcomes contrary
z A mathematical function can be defined, of course, that has nothing to do with the relationship between fuel
economy and the chosen attribute - the most basic example is an industry-wide standard defined as the
mathematical function average required fuel economy =X, where X is the single mpg level set by the agency.
Yet a standard that is simply defined as a mathematical function that is not tied to the attribute(s) would not meet
the requirement of EISA.
aa In fact, numerous manufacturers have confidentially shared with the agencies what they describe as "physics
based" curves, with each OEM showing significantly different shapes, and footprint relationships. The sheer
variety of curves shown to the agencies further confirm the lack of an underlying principle of "fundamental
physics" driving the relationship between CO2 emission or fuel consumption and footprint, and the lack of an
underlying principle to dictate any outcome of the agencies' establishment of footprint-based standards.
2-8
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
to the intent of the statutes to conserve energy and reduce GHG emissions. .bb Thus, the
decision of how to set the target curves cannot always be just about most "clearly" using a
mathematical function to define the relationship between fuel economy and the attribute; it
often has to have reflect legitimate policy judgments, where the agencies adjust the function
that would define the relationship in order to achieve environmental goals, reduce petroleum
consumption, encourage application of fuel-saving technologies, not adversely affect highway
safety, reduce disparities of manufacturers' compliance burdens (thereby increasing the
likelihood of improved fuel economy and reduced GHG emissions across the entire spectrum
of footprint targets), preserve consumer choice, etc. This is true both for the decisions that
guide the mathematical function defining the sloped portion of the target curves, and for the
separate decisions that guide the agencies' choice of "cutpoints" (if any) that define the fuel
economy/CO2 levels and footprints at each end of the curves where the curves become flat.
Data informs these decisions, but how the agencies define and interpret the relevant data, and
then the choice of methodology for fitting a curve to the data, must include a consideration of
both technical data and policy goals. Supporting the consideration and selection of
mathematical functions upon which to base new CAFE and GHG standards, the agencies
conducted a broad-ranging analysis spanning different techniques for adjusting data and
fitting linear functions. The next sections examine the policy concerns that the agencies
considered in developing the target curves that define the MYs 2017-2025 CAFE and CO2
standards, technical work (expanding on similar analyses performed by NHTSA when the
agency proposed MY 2011-2015 standards, and by both agencies during consideration of
options for MY 2012-2016 CAFE and GHG standards) that was completed in the process of
reexamining potential mathematical functions for this rulemaking, how the agencies have
defined the data, and how the agencies explored statistical curve-fitting methodologies in
order to arrive at proposed and final curves. Because the agencies are finalizing the target
curves for MYs 2017-2025 as proposed, the following discussion largely mirrors the
discussion in the version of the TSD that accompanied the proposal; it is repeated here for the
reader's convenience.
2.4 What did the agencies propose for the MYs 2017-2025 curves?
The mathematical functions for the proposed MYs 2017-2025 standards were
somewhat changed from the functions for the MYs 2012-2016 standards, in response to
comments received from stakeholders both pre-proposal and during the public comment
period and in order to address technical concerns and policy goals that the agencies judged
more significant in this nine-model year rulemaking than in the prior one, which only
included five model years.cc This section (2.4) discusses the methodology the agencies
bb For example, if the agencies set weight-based standards defined by a steep function, the standards might
encourage manufacturers to keep adding weight to their vehicles to obtain less stringent targets.
cc We note that although, due to statutory constraints, NHTSA is finalizing standards for only MYs 2017-2021
and presenting augural standards for MYs 2022-2025, the joint analysis was conducted by NHTSA and EPA
with respect to shapes of target curves for all nine model years - both because EPA is indeed finalizing all nine
years of standard curves, and because NHTSA's augural standards for MYs 2022-2025 represent the agency's
2-9
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
selected as best addressing those technical concerns and policy goals for this rulemaking,
given the various technical inputs to the agencies' current analyses. Section 2.5 discusses
how the agencies determined the cutpoints and the flat portions of the MYs 2017-2025 target
curves. We note that both of these sections address only how the target curves were fit to fuel
consumption and CC>2 emission values determined using the city and highway test procedures,
and that in determining respective regulatory alternatives, the agencies made further
adjustments to the resultant curves in order to account for adjustments for improvements to
mobile air conditioners.
Thus, recognizing that there are many reasonable statistical methods for fitting curves
to data points that define vehicles in terms of footprint and fuel economy, the agencies chose
for the proposed rule to fit curves using an ordinary least-squares formulation, on sales-
weighted data, using a fleet that has had technology applied, and after adjusting the data for
the effects of weight-to-footprint, as described below. This represents a departure from the
statistical approach for fitting the curves in the MYs 2012-2016 rules, as explained in the next
section (2.4.1). The agencies considered a wide variety of reasonable statistical methods in
order to better understand the range of uncertainty regarding the relationship between fuel
consumption (the inverse of fuel economy), CC>2 emission rates, and footprint, thereby
providing a range within which decisions about standards would be potentially supportable.
2.4.1 What concerns were the agencies looking to address that led them to change
from the approach used for the MYs 2012-2016 curves?
Before the MY 2017 and later proposal was issued, NHTSA and EPA received a
number of comments from stakeholders on how curves should be fitted to the passenger car
and light truck fleets.dd Some limited-line manufacturers argued that curves should generally
be flatter in order to avoid discouraging production of small vehicles, because steeper curves
tend to result in more stringent targets for smaller vehicles. Most full-line manufacturers
argued that a passenger car curve similar in slope to the MY 2016 passenger car curve would
be appropriate for future model years, but that the light truck curve should be revised to be
less stringent for manufacturers selling the largest full-size pickup trucks. These
manufacturers argued that the MY 2016 light truck curve was not "physics-based," and that in
order for future tightening of standards to be feasible for full-line manufacturers, the truck
curve for later model years should be steeper and extended further (i.e., made less stringent)
into the larger footprints. As stated in the TSD accompanying the proposal, the agencies do
not agree that the MY 2016 light truck curve was somehow deficient in lacking a "physics
basis," or that it was somehow overly stringent for manufacturers selling large pickups—
manufacturers making these arguments presented no "physics-based" model to explain how
fuel economy should depend on footprint.66 The same manufacturers indicated that they
best estimate, based on the information currently before it, of the standards that the agency would finalize had it
the authority to do so. NHTSA will fully revisit all aspects of the MYs 2022-2025 standards as part of the later
rulemaking concurrent with the mid-term evaluation.
dd See 75 FR at 76341 for a general summary.
ee See footnote aa.
2-10
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
believed that the light truck standard should be somewhat steeper after MY 2016, primarily
because, after more than ten years of progressive increases in the stringency of applicable
CAFE standards, large pickups would be less capable of achieving further improvements
without compromising load carrying and towing capacity.
In developing the curve shapes for the proposed rule, the agencies were aware of the
current and prior technical concerns raised by OEMs concerning the effects of the stringency
on individual manufacturers and their ability to meet the standards with available
technologies, while producing vehicles at a cost that allowed them to recover the additional
costs of the technologies being applied. Although we continue to believe that the
methodology for fitting curves for the MYs 2012-2016 standards was technically sound, we
recognize manufacturers' technical concerns regarding their abilities to comply with a
similarly shallow curve after MY 2016 given the anticipated mix of light trucks in MYs 2017-
2025. As in the MYs 2012-2016 rules, the agencies considered these concerns in the analysis
of potential curve shapes for the MYs 2017-2025 proposal. The agencies also considered
safety concerns which could be raised by curve shapes creating an incentive for vehicle
downsizing, as well as the potential loss to consumer welfare should vehicle upsizing be
unduly disincentivized. In addition, the agencies sought to improve the balance of
compliance burdens among manufacturers, and thereby increase the likelihood of improved
fuel economy and reduced GHG emissions across the entire spectrum of footprint targets.
Among the technical concerns and resultant policy trade-offs the agencies considered were
the following:
Flatter standards (i.e., curves) increase the risk that both the weight and size of
vehicles will be reduced, potentially compromising highway safety.
Flatter standards potentially impact the utility of vehicles by providing an incentive for
vehicle downsizing.
Steeper footprint-based standards may create incentives to upsize vehicles, thus
increasing the possibility that fuel economy and greenhouse gas reduction benefits will
be less than expected.
Given the same industry-wide average required fuel economy or CC>2 standard, flatter
standards tend to place greater compliance burdens on full-line manufacturers
Given the same industry-wide average required fuel economy or CC>2 standard, steeper
standards tend to place greater compliance burdens on limited-line manufacturers
(depending of course, on which vehicles are being produced).
If cutpoints are adopted, given the same industry-wide average required fuel economy,
moving small-vehicle cutpoints to the left (i.e., up in terms of fuel economy, down in
terms of CC>2 emissions) discourages the introduction of small vehicles, and reduces
the incentive to downsize small vehicles in ways that could compromise overall
highway safety.
If cutpoints are adopted, given the same industry-wide average required fuel economy,
moving large-vehicle cutpoints to the right (i.e., down in terms of fuel economy, up in
terms of CC>2 emissions) better accommodates the design requirements of larger
vehicles—especially large pickups—and extends the size range over which
downsizing is discouraged.
2-11
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
All of these were policy goals that required weighing and consideration. Ultimately,
the agencies rejected the argument that the MY 2017 target curves for the proposal, on a
relative basis, should be made significantly flatter than the MY 2016 curve,ff as we believed
that this would undo some of the safety-related incentives and balancing of compliance
burdens among manufacturers—effects that attribute-based standards are intended to provide.
Nonetheless, the agencies recognized full-line OEM concerns and tentatively
concluded that further increases in the stringency of the light truck standards would be more
feasible if the light truck curve is made steeper than the MY 2016 truck curve and the right
(large footprint) cut-point is extended over time to larger footprints. This conclusion was
supported by the agencies' technical analyses of regulatory alternatives defined using the
curves developed in the manner described below.
2.4.2 What methodologies and data did the agencies consider in developing the
2017-2025 curves presented in the proposal?
In considering how to address the various policy concerns discussed in the previous
sections, the agencies revisited the data and performed a number of analyses using different
combinations of the various statistical methods, weighting schemes, adjustments to the data
and the addition of technologies to make the fleets less technologically heterogeneous. As
discussed in 2.3.3, in the agencies' judgment, there is no single "correct" way to estimate the
relationship between CO2 or fuel consumption and footprint - rather, each statistical result is
based on the underlying assumptions about the particular functional form, weightings and
error structures embodied in the representational approach. These assumptions are the subject
of the following discussion. This process of performing many analyses using combinations of
statistical methods generated many possible outcomes, each embodying different potentially
reasonable combinations of assumptions and each thus reflective of the data as viewed
through a particular lens. The choice of a standard developed by a given combination of these
statistical methods was consequently a decision based upon the agencies' determination of
how, given the policy objectives for this rulemaking and the agencies' MY 2008-based
forecast of the market through MY 2025, to appropriately reflect the current understanding of
the evolution of automotive technology and costs, the future prospects for the vehicle market,
and thereby establish curves (i.e., standards) for cars and light trucks.
2.4.2.1 For the MYs 2017-2025 standards, what information did the agencies use to
estimate a relationship between fuel economy, CO2 and footprint?
For each fleet, the agencies began with the MY 2008-based market forecast developed
to support the proposal (i.e., the baseline fleet), with vehicles' fuel economy levels and
technological characteristics at MY 2008 levels.gg The development, scope, and content of
ff While "significantly" flatter is subjective qualitative description, the year over year change in curve shapes is
discussed in greater detail in Section 2.5.3.1.
88 While the agencies jointly conducted this analysis, the coefficients ultimately used in the slope setting analysis
are from the CAFE model.
2-12
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
this market forecast is discussed in detail in Chapter 1 of the joint Technical Support
Document supporting the proposed rulemaking.
Figure 2-1 shows the MY 2008 CC^by car and truck class as it existed in the EPA
OMEGA and NHTSA CAFE NPRM model data files (for a gasoline-only fleet, fuel
consumption—the inverse of fuel economy—is directly proportional to CC^). This fleet was
the starting point for all analysis in the proposal.
700
f
2008 CO2 v. Footprint
500 -
O
U
200 -
3>$$'.:•• -ff :*.. - •
«| |U'" : •
&-k-
0 50000
O 100000
O 150000
O 200000
O 250000
40 50 BO 70 40 50 60 70
Footprint
Figure 2-1 2008 CO2 vs. Footprint by Car and Truck
Although the agencies are finalizing the target curves as proposed, the agencies have
also revisited and updated their analyses for this final rule, and found that the proposed curves
are well within the ranges spanned by the final rule analyses. See section 2.6 below. As
discussed in Chapter 1 of this TSD, the agencies have used two different market forecasts to
conduct additional analyses supporting this final rule. The first, referred to here as the "MY
2008-Based Fleet Projection," is largely identical to that used for analysis supporting the
NPRM, but includes some corrections to the footprint of some vehicle models discussed in
Chapter 1, as well as other minor changes. The second, referred to here as the "MY 2010-
Based Fleet Projection," is a post-proposal market forecast based on the MY 2010 fleet of
vehicles. Using both of these projected fleets, the agencies repeated the analyses described
below, and obtained broadly similar results, details of which are presented in a memorandum
2-13
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
available in NHTSA's docket.1* Because the agencies are promulgating target curve standards
identical to those proposed in the NPRM, the remainder of this chapter reviews results
supporting the development of those proposed standards. This chapter concludes with a
summary of results of the agencies' updated analysis, and discussion of the consideration that
analysis was given in selecting mathematical functions upon which to base the standards in
the final rules.
2.4.2.2 What adjustments did the agencies evaluate?
As indicated in the TSD supporting the NPRM, one possible approach is to fit curves to
the minimally adjusted data shown above (the approach still includes sales mix adjustments,
which influence results of sales-weighted regressions), much as DOT did when it first began
evaluating potential attribute-based standards in 2003.7 However, the agencies found, as in
prior rulemakings, that the data are so widely spread (i.e., when graphed, they fall in a loose
"cloud" rather than tightly around an obvious line) that they indicate a relationship between
footprint and CC>2 and fuel consumption that is real but not particularly strong (Figure 2-1).
Therefore, as discussed below, the agencies also explored possible adjustments that could
help to explain and/or reduce the ambiguity of this relationship, or could help to produce
policy outcomes the agencies judged to be more desirable.
2.4.2.3 Adjustment to reflect differences in technology
As in prior rulemakings, the agencies considered technology differences between vehicle
models to be a significant factor producing uncertainty regarding the relationship between
CC>2/fuel consumption and footprint. Noting that attribute-based standards are intended to
encourage the application of additional technology to improve fuel efficiency and reduce CC>2
emissions, the agencies, in addition to considering approaches based on the unadjusted
engineering characteristics of MY 2008 vehicle models, therefore also considered approaches
in which, as for previous rulemakings, technology is added to vehicles for purposes of the
curve fitting analysis in order to produce fleets that are less varied in technology content.
This approach helps to reduce "noise" (i.e.., dispersion) in the plot of vehicle footprints and
fuel consumption levels and to identify a more technology-neutral relationship between
footprint and fuel economy / CC>2 emissions.
For the analysis supporting the NPRM, the agencies adjusted the NPRM baseline fleet for
technology by adding all technologies considered, except for, diesel engines, integrated starter
generators, strong FtEVs, PFtEVs, EVs, FCVs, and the most advanced high-BMEP (brake
14 Docket No. NHTSA-2010-0131. As with the NPRM analysis, EPA and NHTSA jointly analyzed the fleet
projections used in this final rulemaking. While the proposal and final rulemaking analyses shown in this
chapter are from the NHTSA CAFE model, the EPA OMEGA results are generally similar, and support the same
conclusions. A memo containing the OMEGA results for the FRM can be found in EPA docket EPA-HQ-OAR-
2010-0799.
2-14
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
mean effective pressure) gasoline engines." The agencies included 15 percent mass
reduction on all vehicles. Figure 2-2 shows the same fleet, with technology adjustment and
2021 sales applied, and the baseline diesel fueled vehicles, HEV and EVs removed from the
fleet. Of note, the fleet is now more closely clustered" (and lower in emissions), but the same
basic pattern emerges; in both figures, the CO2 emission rate (which, as mentioned above, is
directly proportional to fuel consumption for a gasoline-only fleet) increases with increasing
footprint, although the relationship is less pronounced for larger light trucks.
Max ICE Tech - CO2 v. Footprint
8
o
|300-
(0
E
0
>*
°2QQ -
0
1
0 -
1
•1
*
i
J
Vj
•
a°
0 "*
°£
•"t^J1
:S
1
C
*
a a
?••
F
0*
.
• •
•
• •?
•1
^*
W
T
.
iT *°
f'
a
1%*
*&
|
.i-
2021 Projected Sales
n o
[• I 50000
[•^ 100000
• 150000
• 200000
• 250000
0 300000
40 50 60 70 80 40 50 60 70 80
Footprint
Figure 2-2 2008 CO2 vs. Footprint by Car and Truck, after Adjustment Reflecting Technology
Differences, and removing diesel fueled vehicles, HEVs and EVs
Updating this analysis using the current MY2008- and MY2010-based fleet projection
yielded results generally similar to those shown above. Detailed results of the analyses with
11 As described in the preceding paragraph, applying technology in this manner serves to reduce the effect of
technology differences across the vehicle fleet. The particular technologies used for the normalization were
chosen as a reasonable selection of technologies which could potentially be used by manufacturer over this time
period.
JJ For cars, the standard deviation of the CO2 data is reduced from 81 to 54 through the technology normalization.
For trucks, the standard deviation is reduced from 62 to 36.
2-15
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
the final rulemaking fleet projections are presented in a memorandum available in NHTSA's
docket.1*
2.4.2.4 Adjustments reflecting differences in performance and "density"
As discussed in Section 2.4.1, during stakeholder meetings the agencies held while
developing the NPRM, some manufacturers indicated that they believed that the light truck
standard should be somewhat steeper after MY 2016. As a means to produce a steeper light
truck curve, the agencies considered adjustments for other differences between vehicle
models (i.e., inflating or deflating the fuel economy of each vehicle model based on the extent
to which one of the vehicle's attributes, such as power, is higher or lower than average).
Previously, NHTSA had rejected such adjustments because they imply that a multi-attribute
standard may be necessary, and as explained above, the agencies judged most multi-attribute
standards to be more subject to gaming than a footprint-only standard.mm'8 Having considered
this issue again for purposes of this rulemaking, NHTSA and EPA concluded the need to
accommodate in the target curves the challenges faced by manufacturers of large pickups
currently outweighs these prior concerns (comments on this topic are discussed in Section 0
and 2.4.2.11 and in Section II. C of the preamble). Therefore, the agencies also evaluated
curve fitting approaches through which fuel consumption and CC>2 levels were adjusted with
respect to weight-to-footprint alone, and in combination with power-to-weight. While the
agencies examined these adjustments for purposes of fitting curves, the agencies did not
propose a multi-attribute standard; the proposed fuel economy and CC>2 targets for each
vehicle were still functions of footprint alone. The agencies are not promulgating a multi-
attribute standard, and no adjustment will be used in the compliance process.
The agencies also examined some differences between the technology-adjusted car
and truck fleets in order to better understand the relationship between footprint and CC>2/fuel
consumption in the agencies' MY 2008 based forecast. More direct measures (such as
coefficients of drag and rolling resistance), while useful for vehicle simulation, were not
practical or readily available at the fleet level. Given this issue, and based on analysis
published in the MYs 2012-2016 rule,9 the agencies investigated a sales-weighted (i.e.,
treating every vehicle unit sold as a separate observation) regression equation involving
power to weight ratio and vehicle weight (Equation 2-1).nn This equation provides for a
^ Docket No. NHTSA-2010-0131.
11 See Preamble I.A.2 for a discussion of the stakeholder meetings before the NPRM.
mm For example, in comments on NHTSA's 2008 NPRM regarding MY 2011-2015 CAFE standards, Porsche
recommended that standards be defined in terms of a "Summed Weighted Attribute", wherein the fuel economy
target would calculated as follows: target =f(SWA), where target is the fuel economy target applicable to a
given vehicle model and SWA =footprint + torque111 s + weight112 \ (NHTSA-2008-0089-0174). While the
standards the agencies proposed for MY 2017-2025 are not multi-attribute standards, that is the target is only a
function of footprint, we proposed curve shapes that were developed considering more than one attribute.
1111 These parameters directly relate to the amount of energy required to move the vehicle. As compared to a
lighter vehicle, more energy is required to move a heavier vehicle the same distance. Similarly, a more powerful
engine, when technology adjusted, is less efficient than a less powerful engine.
2-16
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
strong correlation between HP/WT, weight and CC>2 emissions (R2=0.78, Table 2-1) after
accounting for technology adjustments.00
Equation 2-1 - Relationship between vehicle attributes and emissions or fuel consumption
Where:
HP/Weight= the rated horsepower of the vehicle divided by the curb weight
Weight = the curb weight of the vehicle in pounds
C = a constant.
Table 2-1 - Physical Regression Coefficients against Technology Adjusted CO2
R2
F-test p
C
Cars
0.78
<0.01
1.09*10'
3.29*10'2
-3.29
Light Trucks
0.78
<0.01
1.13*10'
3.45*10'2
2.73
*In this gasoline only fleet, these coefficients can be divided by 8887 (the amount of
CO2 produced by the combustion of a gallon of the fuel used to certify the fuel
economy and emissions of gasoline vehicles) to yield the corresponding fuel
consumption coefficients.
Updating this analysis using the MY 2008- and MY 2010-based fleet projections
yielded results generally similar to those shown above. Detailed results of the analyses with
the final rulemaking fleet projections are presented in a memorandum available in NHTSA's
docket.pp
The coefficients above show, for the agencies' MY 2008-based market forecast as
developed for the NPRM, strong correlation between these vehicle attributes and the fuel
consumption and emissions of the vehicle, as well as strong similarity between car and truck
coefficients. (As explained in section 2.6 below, our analysis using the corrected version of
the MY 2008 based market forecast used for the final rule, as well as the alternative 2010
based market forecast, is consistent with these results.) Given these very similar parameters,
00 As R does not equal 1, there are remaining unaccounted for differences beyond technology, power and
weight. These may include gear ratios, axle ratios, aerodynamics, and other vehicle features not captured in this
equation.
pp Docket No. NHTSA-2010-0131.
2-17
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
similar distributions of power and weight would be expected to produce similarly arrayed
plots of CC>2(or equivalently, fuel consumption) by footprint, regardless of car or truck class.
Based on the differences seen in the technology-adjusted plot (Figure 2-2), the agencies
further investigated these particular attributes and their relationship to footprint in the
agencies' MY 2008-based market forecast developed for the NPRM, to examine the
differences across the footprint distribution.
Figure 2-3 shows vehicle curb weight charted against footprint, with sales weighted
ordinary least squares sales fit (blue) and sales-weighted LOESS fit (red) imposed. For cars,
the LOESS fit, which weights nearby points more heavily, qq is nearly identical to the linear fit
in the data filled region between about 40 and 56 sq ft (with the gray bar showing standard
error on the Loess fit). For this market forecast, average car curb weight is linearly
proportional to car footprint between 40 and 56 sq ft, or in other words, cars progress in
weight in a regular fashion as they get larger (Figure 2-3).
Figure 2-3 By contrast, a linear fit does not overlap with the LOESS fit on the truck
side, which indicates that for this market forecast, truck curb weight does not linearly increase
with footprint, at least not across the entire truck fleet. The LOESS fit shows that larger
trucks (those on the right side of the data bend in Figure 2-2) have a different trend than
smaller trucks, and after about 55 sq ft, no longer proportionally increases in weight. The
same pattern is seen in Figure 2-1 and Figure 2-2 above.
qq: In a LOESS regression, "fitting is done locally. That is, for the fit at point x, the fit is made using points in a
neighborhood of x, weighted by their distance from x (with differences in 'parametric' variables being ignored
when computing the distance). The size of the neighborhood is controlled by a For a < 1, the neighborhood
includes proportion a of the points, and these have tricubic weighting (proportional to (1 - (dist/maxdist)^3)^3.
For a > 1, all points are used, with the 'maximum distance' assumed to be a^l/p times the actual maximum
distance forp explanatory variables."
A span of 1 was used in these images, http://cran.r-project.org/doc/manuals/fullrefman.pdf. p. 1406.
2-18
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
WT v. FP - Weighted OLS and Loess Fit
5000 -
3000 -
• . I
40 50 60 70 40 50 60 70
Footprint
Figure 2-3 Relationship between Weight and Footprint in Agencies' MY2008-Based Market
Forecast
Updating this analysis using the revised MY 2008- and the MY 2010-based fleet
projection yielded results generally similar to those shown above. Detailed results of the
analyses with the final rulemaking fleet projections are presented in a memorandum available in
NHTSA's docket."
To further pursue this topic, weight divided by footprint (WT/FP) can be thought of as
a "density" of a vehicle (although dimensionally it has units of pressure). As seen in Figure
2-4, the trend in WT/FP in the agencies' MY2008-based market forecast is different in trucks
than in cars. The linear trend on cars is an increase in WT/FP as footprint increases (Figure
2-4). In contrast, light trucks do not consistently increase in WT/FP ratio as the vehicles grow
larger, but WT/FP actually decreases (Figure 2-4).
Docket No. NHTSA-2010-0131.
2-19
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
WT/FP v, FP - Weighted OLS
70 40
Footprint
Figure 2-4 Relationship between Weight/FP and Footprint in Agencies' MY2008-Based Market Forecast
Updating this analysis using the current MY 2008- and MY 2010-based fleet
projection yielded results generally similar to those shown above. Detailed results of the
analyses with the final rulemaking fleet projections are presented in a memorandum available
inNHTSA'sdocket.88
The heterogeneity of the truck fleet explains part of the WT/FP trend, where the
pickup truck fleet is largest in footprint, but is also relatively light for its size due to the flat
bed (Figure 2-5). Note that the two light truck classes with the smallest WT/FP ratios are
small and large pickups. Further, as the only vehicle class with a sales-weighted average
footprint above 60 square feet, the large pickup trucks have a strong influence on the slope of
the truck curve. As the correlation between weight and CC>2 is strong (Table 2-1), having
proportionally lighter vehicles at one extreme of the footprint distribution can bias a curve fit
to these vehicles. If no adjustment is made to the curve fitted to the truck fleet, and no other
compensating flexibilities or adjustments are made available, manufacturers selling
significant numbers of vehicles at the large end of the truck distribution will face compliance
burdens that are comparatively more challenging that those faced by manufacturers not
serving this part of the light truck market. As noted further below, this consideration
1 Docket No. NHTSA-2010-0131.
2-20
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
provided the basis for the agencies' proposal to change the cutpoint for larger light trucks
from 66 feet to 74 feet, and to steepen the slope of the light truck curve for larger light trucks.
WT/FP by Vehicle Class
Truckfleetavg.: 84.3
Carfleetavg.; 72.7
Footprint
Figure 2-5 Class and the WT/FP distribution
Updating this analysis using the revised MY 2008- and the MY 2010-based market
forecasts yielded results generally similar to those shown above. Detailed results of the
analyses with the final rulemaking fleet projections are presented in a memorandum available
inNHTSA'sdocket*
The agencies also investigated the relationship between HP/WT and footprint in the
agencies' MY 2008-based market forecast developed for the NPRM (Figure 2-6). On a sales
weighted basis, cars tend to become proportionally more powerful as they get larger. In
contrast, there is a minimally positive relationship between HP/WT and footprint for light
trucks, indicating that light trucks become only slightly more powerful as they get larger, but
that the trend is not especially pronounced.
'DocketNo. NHTSA-2010-0131.
2-21
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
HP/WT v. FP - Weighted OLS
020 •
015 -
010 -
005
V • Tt *
40 50 60 70 40 SO 60 70
Footprint
2021 Salts
« 0
^ 50000
• 100000
I* 150000
• 200000
• 250000
• 300000
Figure 2-6 HP/WT v. FP
Updating this analysis using the revised MY 2008- and the MY 2010-based fleet
projection yielded results generally similar to those shown above. Detailed results of the
analyses with the final rulemaking fleet projections are presented in a memorandum available
inNHTSA'sdocket.1111
One factor influencing results of this analysis is the non-homogenous nature of the
truck fleet; some vehicles at the smaller end of the footprint curve are different in design and
utility from others at the larger end (leading to the observed bend in the LOESS fit, Figure
2-6). There are many high volume four-wheel drive vehicles with smaller footprint in the
truck fleet (such as the Chevrolet Equinox, Dodge Nitro, Ford Escape, Honda CR-V, Hyundai
Santa Fe, Jeep Liberty, Nissan Rogue, Toyota RAV4, and others) exhibit only select truck
characteristics.™ By contrast, the largest pickup trucks in the light truck fleet have unique
aerodynamic and power characteristics that tend to increase CO2 emissions and fuel
consumption. These disparities contribute to the slopes of lines fitted to the light truck fleet.
uu Docket No. NHTSA-2010-0131.
vv In most cases, these vehicles have four-wheel drive, but no significant towing capability, and no open-bed.
Many of these vehicles are also offered without four-wheel drive, and these two-wheel drive versions are
classified as passenger cars, not light trucks.
2-22
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Several comments, such as those by CBD and ACEEE, were submitted with regard to
the non-homogenous nature of the truck fleet, and the "unique" attributes of pickup trucks.
Ford Motor Company described the attributes of these vehicles, noting that "towing capability
generally requires increased aerodynamic drag caused by a modified frontal area, increased
rolling resistance, and a heavier frame and suspension to support this additional capability."^
Ford further noted that these vehicles further require auxiliary transmission oil coolers,
upgraded radiators, trailer hitch connectors and wiring harness equipment, different steering
ratios, upgraded rear bumpers and different springs for heavier tongue load (for upgraded
towing packages), body-on-frame (vs. unibody) construction (also known as ladder frame
construction) to support this capability and an aggressive duty cycle, and lower axle ratios for
better pulling power/capability. In the agencies' judgment, the curves and cutpoints defining
the light truck standards appropriately account for engineering differences between different
types of vehicles. For example, the agencies' estimates of the applicability, cost, and efficacy
of different fuel-saving technologies differentiate between small, medium, and large light
trucks. Further discussion on this topic is contained in Section II.C.
The agencies' technical analyses of regulatory alternatives developed using curves
fitted as described below supported OEM comments that there would be significant
compliance challenges for the manufacturers of large pickup trucks, and led toward the
agencies' policy goal of a steeper slope for the light truck curve relative to MY 2016. Three
primary drivers were as follows: (a) the largest trucks have unique equipment and design, as
described in the Ford comment referenced above; (b) the agencies agree with those large truck
manufacturers who indicated in discussions prior to the proposal that they believed that the
light truck standard should be somewhat steeper after MY 2016, primarily because, after more
than ten recent years of progressive increases in the stringency of applicable CAFE standards
(after nearly ten years during which Congress did not allow NHTSA to increase light truck
CAFE standards), manufacturers of large pickups would have limited options to comply with
more stringent standards without resorting to compromising large truck load carrying and
towing capacity; and (c) given the relatively few platforms which comprise the majority of the
sales at the largest truck footprints, the agencies were concerned about requiring levels of
average light truck performance that might lead to overly aggressive advanced technology
penetration rates in this important segment of the work fleet. Specifically, the agencies were
concerned at proposal, and remain concerned, about issues of lead time and cost with regard
to manufacturers of these work vehicles. As noted later in this chapter, while the largest
trucks are a small segment of the overall truck fleet, and an even smaller segment of the
overall fleet, ** these changes to the truck slope have been made in order to provide a clearer
path toward compliance for manufacturers of these vehicles, and reduce the potential that new
ww Ford comments, Docket No. [fill in], at [page number].
xx The agencies' market forecast used at proposal includes about 24 vehicle configurations above 74 square feet
with a total volume of about 50,000 vehicles or less during any MY in the 2017-2025 time frame, In the
MY2010 based market forecast, there are 14 vehicle configurations with a total volume of 130,000 vehicles or
less during any MY in the 2017-2025 time frame. This is a similarly small portion of the overall number of
vehicle models or vehicle sales.
2-23
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
standards would lead these manufacturers to choose to downpower, modify the structure, or
otherwise reduce the utility of these work vehicles.
Some commenters disagreed with these policy goals concerning the largest light trucks
and argued that higher fuel economy for the largest light trucks is fully compatible with
maintaining towing and hauling capacity. These comments, which largely deal with
stringency, are addressed in each agency's respective preamble section (HID and IV.F), as
well as in Section II. C, which addresses the shapes of the target curves. Consequently, the
agencies considered options including fitting curves developed using results of the analysis
described above. Specifically, the agencies note that the WT/FP ratio of the light duty fleet
potentially has a large impact on a sales-weighted regression.37 The increasing trend in
WT/FP versus footprint for cars in the 2008 MY baseline would steepen the slope of the car
curve, while the decreasing trend in WT/FP would flatten the truck slope, as compared to a
WT/FP adjusted fleet. This result was reflected in the MYs 2012-2016 final rulemaking,10
where the agencies noted the steep car curves resulting from a weighted least-squares
analysis.
Based on the above analysis, the agencies also considered adjustments for other
differences between vehicle models. Therefore, utilizing the coefficients derived in Equation
2-1, the agencies also evaluated curve fitting approaches through which fuel consumption and
CC>2 levels were adjusted with respect to weight-to-footprint alone, and in combination with
power-to-weight. This adjustment procedure inflates or deflates the fuel economy or CO2
emissions of each vehicle model based on the extent to which one of the vehicle's attributes,
such as power, is higher or lower than average. As mentioned above, while the agencies
considered this technique for purposes of fitting curves, the agencies did not propose a multi-
attribute standard, as the proposed fuel economy and CO2 targets for each vehicle were still
functions of footprint alone. The agencies are not promulgating a multi-attribute standard,
and no adjustment would be used in the compliance process.
The basis for the gallon-per-mile (GPM) adjustments is the sales-weighted linear
regression discussed in 2.4 (Equation 2-1, Table 2-1). The coefficients to this equation give
the impact of the various car attributes on CC>2 emissions and fuel consumption in the
agencies' MY 2008-based market forecast used in the NPRM. For example, gives the
impact of weight while holding the ratio horsepower to weight constant. Importantly, this
means that as weight changes, horsepower must change as well to keep the power/weight ratio
constant. Similarly, gives the CC>2 impact of changing the performance of the vehicle
while keeping the weight constant. These coefficients were used to perform an adjustment of
the gallons per mile measure for each vehicle to the respective car or truck—i.e., in the case
of a HP/WT adjustment, to deflate or inflate the fuel consumption of each vehicle model
based on the extent to which the vehicle's power-to-weight ratio is above or below the
regression-based value at that footprint.
yy As mentioned above, the agencies also performed the same analysis without sales-weighting, and found that
the WT/FP ratio also had a directionallv similar effect on the fitted car and truck curves.
2-24
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
The agencies performed this normalization to adjust for differences in vehicle weight
per square foot observations in the data discussed in Section 2.4. This adjustment process
requires two pieces of information: the weight coefficient from Equation 2-1 and the average
weight per footprint (i.e., pounds per square foot) for that vehicle's group. Two groups,
passenger cars and light trucks, were used. For each group, the average weight per footprint
was calculated as a weighted average with the weight being the same as in the above
regression (projected sales by vehicle in 2021). The equation below indicates how this
adjustment was carried out.
2-25
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Equation 2-2 WT/FP adjustment
The term in parentheses represents the vehicle's deviation from an "expected weight."
That is, multiplying the average weight per footprint for a group of vehicles (cars or trucks)
by a specific vehicle's footprint gives an estimate of the weight of that specific vehicle if its
density were "average," based on the analyzed fleet. Put another way, this factor represents
what the weight is "expected" to be, given the vehicle's footprint, and based on the analyzed
fleet. This "expected weight" is then subtracted from the vehicle's actual weight. Vehicles
that are heavier than their "expected weight" will receive a positive value (i.e., a deflated fuel
economy value) here, while vehicles that are lighter than their "expected weight" will receive
a negative number (i.e., an inflated fuel economy value).
This deviation from "expected weight" is then converted to a gallon value by the
regression coefficient. The units on this coefficient are gallons per mile per pound, as can be
deduced from equation 1. This value is then subtracted from the vehicle's actual gallons per
mile measure. Note that the adjusted truck data no longer exhibits the bend seen in Figure 2-1
and Figure 2-2.
in
o
£ 8
CM
O
40 50 60 70 80 40 50 60 70 80
Figure 2-7 WT/FP Adjusted Fuel Consumption vs. Footprint
2-26
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Updating this analysis using the revised MY 2008- and the MY 2010-based fleet
projection yielded results generally similar to those shown above. Detailed results of the
analyses with the final rulemaking fleet projections are presented in a memorandum available
inNHTSA'sdocket.22
This adjustment serves to reduce the variation in gallons per mile measures caused by
variation in weight in the agencies' MY 2008-based market forecast used in the NPRM.
Importantly, this adjustment serves to reduce the fuel consumption (i.e., inflate fuel economy)
for those vehicles which are heavier than their footprint would suggest while increasing the
gallons per mile measure (i.e., deflating fuel economy) for those vehicles which are lighter.
For trucks, a linear trend is more evident in the data cloud.aaa The following table shows the
degree of adjustment for several vehicle models:
zz Docket No. NHTSA-2010-0131.
aaa Using EPA's dataset, R2 for the sales weighted ordinary least squared linear fit between footprint and CO2
improved from 0.38 (technology adjusted CO2) to 0.64 (technology and weight / footprint adjusted CO2)
2-27
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Table 2-2 - Sample Adjustments for Weight to Footprint, Cars
Manufacturer
HONDA
TOYOTA
FORD
GENERAL
MOTORS
HONDA
NISSAN
GENERAL
MOTORS
FORD
TOYOTA
VOLKSWAGEN
FORD
HONDA
HYUNDAI
HONDA
Model
HONDA FIT
TOYOTA
COROLLA
FORD FOCUS
CHEVROLET
MALIBU
HONDA
ACCORD
INFINITIG37
CHEVROLET
CORVETTE
FORD
MUSTANG
TOYOTA
CAMRY
VOLKSWAGEN
JETTA
FORD FUSION
HONDA
ACCORD
HYUNDAI
SONATA
HONDA CIVIC
Name Plate
FIT
COROLLA
FOCUS FWD
MALIBU
ACCORD 4DR
SEDAN
G37 COUPE
CORVETTE
MUSTANG
CAMRY
SOLARA
CONVERTIBLE
JETTA
FUSION FWD
ACCORD 2DR
COUPE
SONATA
CIVIC
Weight /
Footprint
64.4
61.3
62.9
73.5
69.6
76.7
69.3
74.7
75.6
78.0
72.2
71.6
70.7
59.9
Footprint
39.5
42.5
41.7
46.9
46.6
47.6
46.3
46.7
46.9
42.4
46.1
46.6
46.0
43.2
GPM
0.01
0.01
0.02
0.02
0.02
0.02
0.02
0.03
0.02
0.02
0.02
0.02
0.02
0.02
MPG
69.40
69.94
61.94
53.70
57.57
47.83
40.84
31.32
50.87
46.77
59.96
56.92
61.72
64.25
Adjusted
GPM
0.0157
0.0164
0.0177
0.0185
0.0179
0.0200
0.0251
0.0316
0.0191
0.0211
0.0168
0.0178
0.0166
0.0177
Adjusted
MPG
63.73
60.80
56.34
54.08
55.73
50.08
39.83
31.67
52.27
47.47
59.61
56.26
60.34
56.38
GPM%
Adjustment
8.9%
15.0%
9.9%
-0.7%
3.3%
-4.5%
2.5%
-1.1%
-2.7%
-1.5%
0.6%
1.2%
2.3%
14.0%
2-28
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Table 2-3 - Sample Adjustments for Weight to Footprint, Trucks
Manufacturer
FORD
GENERAL
MOTORS
FIAT
HONDA
TOYOTA
FORD
FIAT
TOYOTA
TATA
GENERAL
MOTORS
GENERAL
MOTORS
GENERAL
MOTORS
TOYOTA
Model
FORD ESCAPE
CHEVROLET
CIS
JEEP GRAND
CHEROKEE
HONDA PILOT
TOYOTA
HIGHLANDER
FORD F150
DODGE RAM
TUNDRA
LAND ROVER
RANGE
ROVER SPORT
CHEVROLET
UPLANDER
HUMMER H3
PONTIAC
TORRENT
TACOMA
Name Plate
ESCAPE FWD
CIS
SILVERADO
2WD 119WB
GRAND
CHEROKEE
4WD
PILOT 4WD
HIGHLANDER
4WD
F150 FFV
4WD 145 WB
RAM 1500
PICKUP 4WD
140 WB
TOYOTA
TUNDRA
4WD 145 WB
RANGE
ROVER
SPORT
UPLANDER
FWD
H34WD
TORRENT
FWD
TOYOTA
TACOMA
4WD
Weight /
Footprint
80.1
85.9
103.7
85.2
79.6
73.8
78.1
79.3
118.6
114.4
99.9
84.2
74.8
Footprint
65.2
55.9
47.1
51.3
49.0
67.4
66.3
68.7
47.5
49.2
50.7
48.2
53.4
GPM
0.02
0.03
0.02
0.02
0.02
0.03
0.03
0.03
0.03
0.02
0.03
0.02
0.02
MPG
51.00
39.76
41.45
40.95
45.90
32.70
33.75
32.07
33.17
45.46
36.71
46.64
43.01
Adjusted
GPM
0.0181
0.0248
0.0222
0.0243
0.0227
0.0334
0.0316
0.0325
0.0239
0.0163
0.0242
0.0215
0.0252
Adjusted
MPG
55.11
40.29
44.98
41.22
44.05
29.97
31.65
30.73
41.92
61.34
41.30
46.56
39.63
GPM%
Adjustment
-7.5%
-1.3%
-7.9%
-0.6%
4.2%
9.1%
6.6%
4.3%
-20.9%
-25.9%
-11.1%
0.2%
8.5%
Updating this analysis using the revised MY 2008- and the MY 2010-based fleet
projection yielded results generally similar to those shown above. Detailed results of the
analyses with the final rulemaking fleet projections are presented in a memorandum available
inNHTSA'sdocket'"
bbb
Based on Equation 2-1, the agencies also evaluated an adjustment of GPM and CO2
based on HP/WT.
Equation 2-3 -Adjustment based on HP/WT
bbb
Docket No. NHTSA-2010-0131.
2-29
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Figure 2-8 shows the adjusted data and the estimated relationship between the adjusted
GPM values and footprint.
CO
o
CL
O
CM
O
50
60
70 80 40
Footprint
50
60
70
80
Figure 2-8 HP/WT Adjusted Fuel Consumption v. Footprint
Table 2-4 shows the degree of adjustment for several vehicle models. Those vehicles
which have more power than average for their actual curb weight are adjusted downward (i.e.,
fuel economy ratings are inflated), while those that have less power than average are adjusted
upward (i.e., fuel economy ratings are deflated).
Table 2-4 - Sample Adjustments for Horsepower to Weight, Cars
Manufacturer
HONDA
TOYOTA
FORD
GENERAL
MOTORS
Model
HONDA FIT
TOYOTA
COROLLA
FORD FOCUS
CHEVROLET
MALIBU
Name Plate
FIT
COROLLA
FOCUS FWD
MALIBU
Horsepower
109
126
140
169
Footprint
39.5
42.5
41.7
46.9
GPM
0.01
0.01
0.02
0.02
MPG
69.40
69.94
61.94
53.70
Adjusted
GPM
0.0157
0.0164
0.0177
0.0185
Adjusted
MPG
63.73
60.80
56.34
54.08
GPM%
Adjustment
8.9%
15.0%
9.9%
-0.7%
2-30
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
HONDA
NISSAN
GENERAL
MOTORS
FORD
TOYOTA
VOLKSWAGEN
FORD
HONDA
HYUNDAI
HONDA
HONDA
ACCORD
INFINITIG37
CHEVROLET
CORVETTE
FORD
MUSTANG
TOYOTA
CAMRY
VOLKSWAGEN
JETTA
FORD FUSION
HONDA
ACCORD
HYUNDAI
SONATA
HONDA CIVIC
ACCORD 4DR
SEDAN
G37 COUPE
CORVETTE
MUSTANG
CAMRYSOLARA
CONVERTIBLE
JETTA
FUSION FWD
ACCORD 2DR
COUPE
SONATA
CIVIC
190
330
400
500
225
170
160
190
162
140
46.6
47.6
46.3
46.7
46.9
42.4
46.1
46.6
46.0
43.2
0.02
0.02
0.02
0.03
0.02
0.02
0.02
0.02
0.02
0.02
57.57
47.83
40.84
31.32
50.87
46.77
59.96
56.92
61.72
64.25
0.0179
0.0200
0.0251
0.0316
0.0191
0.0211
0.0168
0.0178
0.0166
0.0177
55.73
50.08
39.83
31.67
52.27
47.47
59.61
56.26
60.34
56.38
3.3%
-4.5%
2.5%
-1.1%
-2.7%
-1.5%
0.6%
1.2%
2.3%
14.0%
Table 2-5 - Sample Adjustments for Horsepower to Weight, Trucks
Manufacturer
FORD
GENERAL
MOTORS
FIAT
HONDA
TOYOTA
FORD
FIAT
TOYOTA
TATA
GENERAL
MOTORS
GENERAL
MOTORS
GENERAL
MOTORS
TOYOTA
Model
FORD ESCAPE
CHEVROLET
CIS
JEEP GRAND
CHEROKEE
HONDA PILOT
TOYOTA
HIGHLANDER
FORD F150
DODGE RAM
TUNDRA
LAN DROVER
RANGE ROVER
SPORT
CHEVROLET
UPLANDER
HUMMER H3
PONTIAC
TORRENT
TACOMA
Name Plate
ESCAPE FWD
CIS SILVERADO
2WD 119WB
GRAND CHEROKEE
4WD
PILOT 4WD
HIGHLANDER
4WD
F150 FFV 4WD
145 WB
RAM 1500 PICKUP
4WD 140 WB
TOYOTA TUNDRA
4WD 145 WB
RANGE ROVER
SPORT
UPLANDER FWD
H34WD
TORRENT FWD
TOYOTA TACOMA
4WD
Horsepower
153
195
210
244
270
300
345
381
300
240
242
185
236
Footprint
65.2
55.9
47.1
51.3
49.0
67.4
66.3
68.7
47.5
49.2
50.7
48.2
53.4
GPM
0.02
0.03
0.02
0.02
0.02
0.03
0.03
0.03
0.03
0.02
0.03
0.02
0.02
MPG
51.00
39.76
41.45
40.95
45.90
32.70
33.75
32.07
33.17
45.46
36.71
46.64
43.01
Adjusted
GPM
0.0181
0.0248
0.0222
0.0243
0.0227
0.0334
0.0316
0.0325
0.0239
0.0163
0.0242
0.0215
0.0252
Adjusted
MPG
55.11
40.29
44.98
41.22
44.05
29.97
31.65
30.73
41.92
61.34
41.30
46.56
39.63
GPM %
Adjustment
-7.5%
-1.3%
-7.9%
-0.6%
4.2%
9.1%
6.6%
4.3%
-20.9%
-25.9%
-11.1%
0.2%
8.5%
Updating this analysis using the revised MY 2008- and the MY 2010-based fleet
projection yielded results generally similar to those shown above. Detailed results of the
2-31
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
analyses are with the final rulemaking fleet projections presented in a memorandum available
inNHTSA'sdocket.ccc
The above approaches resulted in three data sets each for (a) vehicles without added
technology and (b) vehicles with technology added to reduce technology differences, any of
which may provide a reasonable basis for fitting mathematical functions upon which to base
the slope of the standard curves: (1) vehicles without any further adjustments; (2) vehicles
with adjustments reflecting differences in "density" (weight/footprint); and (3) vehicles with
adjustments reflecting differences in "density," and adjustments reflecting differences in
performance (power/weight). Further, these sets were developed for both the revised MY
2008-based fleet projection and the post-proposal MY 2010-based fleet projection. Detailed
results using these market forecasts are presented in a memorandum available in NHTSA's
docket.ddd
2.4.2.5 What statistical methods did the agencies evaluate?
Using these data sets, the agencies tested a range of regression methodologies, each
judged to be possibly reasonable for application to at least some of these data sets.
2.4.2.6 Regression Approach
In the MYs 2012-2016 final rules, the agencies employed a robust regression approach
(minimum absolute deviation, or MAD), rather than an ordinary least squares (OLS)
regression.11 MAD is generally applied to mitigate the effect of outliers in a dataset, and thus
was employed in that rulemaking as part of our interest in attempting to best represent the
underlying technology. NHTSA had used OLS in early development of attribute-based
CAFE standards, but NHTSA (and then NHTSA and EPA) subsequently chose MAD instead
of OLS for both the MY 2011 and the MYs 2012-2016 rulemakings. These decisions on
regression technique were made both because OLS gives additional emphasis to outliers12 and
because the MAD approach helped achieve the agencies' policy goals with regard to curve
slope in those rulemakings.13 In the interest of taking a fresh look at appropriate regression
methodologies as promised in the 2012-2016 light duty rulemaking, in developing this
proposal, the agencies gave full consideration to both OLS and MAD. The OLS
representation, as described, uses squared errors, while MAD employs absolute errors and
thus weights outliers less.
As noted, one of the reasons stated for choosing MAD over least square regression in
the MYs 2012-2016 rulemaking was that MAD reduced the weight placed on outliers in the
data. As seen in Figure 2-1, there clearly are some outliers in the data, mostly to the high CO2
and fuel consumption side. However, the agencies have further considered whether it is
appropriate to classify these vehicles as outliers. Unlike in traditional datasets, these vehicles'
performance is not mischaracterized due to errors in their measurement, a common reason for
ccc Docket No. NHTSA-2010-0131.
ddd Docket No. NHTSA-2010-0131.
2-32
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
outlier classification. Being certification data, the chances of large measurement errors
should be near zero, particularly towards high CC>2 or fuel consumption. Thus, they can only
be outliers in the sense that the vehicle designs are unlike those of other vehicles. These
outlier vehicles may include performance vehicles, vehicles with high ground clearance,
4WD, or boxy designs. Given that these are equally legitimate on-road vehicle designs, the
agencies concluded that it would appropriate to reconsider the treatment of these vehicles in
the regression techniques.
Based on these considerations as well as on the adjustments discussed above, the
agencies concluded it was not meaningful to run MAD regressions on gpm data that had
already been adjusted in the manner described above. Normalizing already reduced the
variation in the data, and brought outliers towards average values. This was the intended
effect, so the agencies deemed it unnecessary to apply an additional remedy to resolve an
issue that had already been addressed, but we sought comment on the use of robust regression
techniques under such circumstances. One commenter, ACEEE, addressed this question in
this rulemaking, indicating (consistent with the agencies' views) that MAD and OLS are both
technically sound methods for fitting functions.
2.4.2.7 Sales Weighting
Likewise, in the proposal, the agencies reconsidered the application of sales-weighting
to represent the data. As explained below, the decision to sales weight or not is ultimately
based upon a choice about how to represent the data, and not by an underlying statistical
concern. Sales weighting is used if the decision is made to treat each (mass produced) unit
sold as a unique physical observation. Doing so thereby changes the extent to which different
vehicle model types are emphasized as compared to a non-sales weighted regression. For
example, while total General Motors Silverado (332,000) and Ford F-150 (322,000) sales
differed by less than 10,000 in MY 2021 market forecast used for the NPRM, 62 F-150s
models and 38 Silverado models were reported in the agencies baselines. Without sales-
weighting, the F-150 models, because there were more of them, were given 63 percent more
weight in the regression despite comprising a similar portion of the marketplace and a
relatively homogenous set of vehicle technologies.
The agencies did not use sales weighting in the MYs 2012-2016 rulemaking analysis
of the curve shapes. A decision to not perform sales weighting reflects judgment that each
vehicle model provides an equal amount of information concerning the underlying
relationship between footprint and fuel economy. Sales-weighted regression gives the highest
sales vehicle model types vastly more emphasis than the lowest-sales vehicle model types
thus driving the regression toward the sales-weighted fleet norm. For unweighted regression,
vehicle sales do not matter. The agencies note that the light truck market forecast shows MY
2025 sales of 218,000 units for Toyota's 2WD Sienna, and shows 66 model configurations
with MY 2025 sales of fewer than 100 units. Similarly, the agencies' market forecast shows
MY 2025 sales of 267,000 for the Toyota Prius, and shows 40 model configurations with
MY2025 sales of fewer than 100 units. Sales-weighted analysis would give the Toyota
Sienna and Prius more than a thousand times the consideration of many vehicle model
configurations. Sales-weighted analysis would, therefore, cause a large number of vehicle
model configurations to be virtually ignored in the regressions.14
2-33
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
However, the agencies did note in the MYs 2012-2016 final rules that, "sales weighted
regression would allow the difference between other vehicle attributes to be reflected in the
analysis, and also would reflect consumer demand." 15 In reexamining the sales-weighting for
this analysis, the agencies note that there are low-volume model types that account for many
of the passenger car model types (50 percent of passenger car model types account for 3.3
percent of sales), and it is unclear whether the engineering characteristics of these model types
should equally determine the standard for the remainder of the market.
In the interest of taking a fresh look at appropriate methodologies as promised in the
last final rule, in developing proposed and final standards for MYs 2017-2025, the agencies
gave full consideration to both sales-weighted and unweighted regressions.
2.4.2.8 Analyses Performed
We performed regressions describing the relationship between a vehicle's CCVfuel
consumption and its footprint, in terms of various combinations of factors: initial (raw) fleets
with no technology, versus after technology is applied; sales-weighted versus non-sales
weighted; and with and without two sets of normalizing factors applied to the observations.
The agencies excluded diesels and dedicated AFVs because the agencies anticipate that
advanced gasoline-fueled vehicles are likely to be dominant through MY2025.
Results supporting development of the proposed and finalized standards are depicted
graphically in Figures 2-9 through 2-16, below.
Thus, the basic OLS regression on the initial data (with no technology applied) and no
sales-weighting represents one perspective on the relation between footprint and fuel
economy. Adding sales weighting changes the interpretation to include the influence of sales
volumes, and thus steps away from representing vehicle technology alone. Likewise, MAD is
an attempt to reduce the impact of outliers, but reducing the impact of outliers might perhaps
be less representative of technical relationships between the variables, although that
relationship may change over time in reality. Each combination of methods and data reflects
a perspective, and the regression results reflect that perspective in a simple quantifiable
manner, expressed as the coefficients determining the line through the average (for OLS) or
the median (for MAD) of the data. It is left to policy makers to determine an appropriate
perspective and to interpret the consequences of the various alternatives.
The agencies sought comment on the application of the weights as described above,
and the implications for interpreting the relationship between fuel efficiency and footprint.
ACEEE questioned adjustment of the light truck data based on differences in weight/footprint,
indicating that, in their view, the adjustment produces too steep a slope and potentially
implies overstatement of the efficacy of some technologies as applied to pickup trucks.
ACEEE also suggested that adjustment based on differences in power/weight would yield
flatter curves and be more consistent with how the EU constructed related CO2 targets. The
Alliance, in contrast, supported the weightings applied by the agencies, and the resultant
relationships between fuel efficiency and footprint. Both ACEEE and the Alliance
commented that the agencies should revisit the application of weights—and broader aspects
of analysis to develop mathematical functions—in the future. Moreover, although ACEEE
2-34
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
expressed concern regarding the outcomes of the application of the weight/footprint
adjustment, the agencies maintain that the adjustments (including no adjustments) considered
in the NPRM are all potentially reasonable to apply for purposes of developing fuel economy
and GHG target curves. This issue is discussed in greater detail in Section II.C of the
preamble, and related issues-the slope and stringency of the light truck standards—are
addressed further in Sections III and IV of the preamble.
2.4.2.9 What results did the agencies obtain?
Both agencies employed the same statistical approaches. For regressions against data
including technology normalization, NHTSA used the CAFE modeling system, and EPA used
EPA's OMEGA model. The agencies obtained similar regression results, and based the joint
proposal on those obtained by NHTSA.
For illustrative purposes, the set of figures below show the range of curves determined
by the possible combinations of regression techniques, with and without sales weighting, with
and without the application of technology, and with various adjustments to the gpm variable
prior to running a regression. Again, from a statistical perspective, each of these regressions
simply represents the assumptions employed. Since they are all univariate linear regressions,
they describe the line that will result from minimizing the sum of the residuals (for MAD) or
sum of squared residuals (for OLS). Figures show the results for passenger cars, then light
trucks, for ordinary least squares (OLS) then similar results for MAD regressions for cars and
light trucks, respectively. The various equations are represented by the string of attributes
used to define the regression. See the table, Regression Descriptors, below, for the legend.
Thus, for example, the line representing "ols_LT_wt_ft_adj_init_w" should be read as
follows: an OLS regression, for light trucks, using data adjusted according to weight to
footprint, no technology added, and weighted by sales.
Table 2-6 Regression Descriptors
Notation
ols or mad
PC or LT
hp wt adj
wt ft adj
wt ft hp wt adj
init or final
u or w
Description
Ordinary least squares or mean absolute deviation
Passenger car or light truck
Adjustment for horsepower to weight
Adjustment for weight to footprint
Adjustment for both horsepower to weight and weight to
Vehicles with no technology (initial) or with technology
footprint
added (final)
Unweighted or weighted by sales
2-35
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Thus, the next figures, for example, represent a family of curves (lines) fit using ordinary least
squares on data for passenger cars, not modified for technology, and which therefore permits
comparisons of results in terms of the factors that change in each regression. These factors
are whether the data are sales-weighted (denoted "w") or unweighted (denoted "u"), as well as
the adjustments described above. Each of these adjustments has an influence on the
regressions results, depicted in the figures below.
Updating this analysis using the revised MY 2008- and the MY 2010-based fleet
projection yielded results generally similar to those shown above. See section 2.6 below.
Detailed results of the analysis with the final rulemaking fleet projections are presented in a
memorandum available in NHTSA's docket.666
' Docket No. NHTSA-2010-0131.
2-36
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
0.01
•55
45
75
85
55 65
Footprint
olsPC init_w olsPC init_u
— olsPC_hp_wt_adj init_w ^^— olsPC_hp_wt_adj init_u
olsPC_wt_ft_hp_wt_adj init_w olsPC_wtJt_hp_wt_adj_init_u
o Is P C_w t_f t_a clj i n i t_w o I s P C_ w t_f t_a dj i n it_u
o No Technology Fleet n Technology Fleet
Figure 2-9 Best Fit Results for Various Regressions: Cars, No Added Technology, OLS
Figure 2-10, below, shows comparable results, this time with data representing the
additional technology that has been added to reduce technological heterogeneity. Note that
the data now pass through the relevant data "cloud" for the fleet with the technology
2-37
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
adjustment applied. The slopes of the lines are somewhat more clustered (less divergent) in
the chart depicting added technology (as discussed in footnote ii)
o.oi
25 35
•olsPC final_w
•olsPC_hp_wt_adj final_w
55 65 75
R««*PtiR,t»C_final_u
olsPC_hp_wt_aclj final_u
85
^^— o I s P C_w t_f t_hp_wt_a clj f i na l_w o I s P C_ w t_f t_hp_ wt_a clj f i na l_u
olsPC_wt_ft_aclj final_w - olsPC_wt_ft_adj final_u
o No Technology Fleet o Technology Fleet
Figure 2-10 Best Fit Results for Various Regressions: Cars, with Added Technology, OLS
Similar to the figures displaying the results for passenger cars, the figures below
display regression lines for trucks, first with no technology added, then subsequently, for the
case where technology has been added. Slopes appear more similar to each other here than of
passenger cars.
2-38
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
0.01
25 35
^^— o I s LT i n i t_w
olsLT_hp_wt_adj init_w —— olsLT_hp_wt_aclj init_u
ol$LT_wt_ft_hp_wt_adj init_w olsl_T_wt_ft_hp_wt_adj init_u
olsLT_wt_ft_adj_init_w olsLT_wt_ft_adj init_u
O No Technology Fleet D Technology Fleet
Figure 2-11 Best Fit Results for Various Regressions: Trucks, No Added Technology, OLS
2-39
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
0.01
75
85
25 35 45 55 65
ol$LT final_w olsLT final_u
«^^ olsLT_hp_wt_atlj f inal_w ^^ olsl_T_hp_wt_adj final_u
ol$LT_wt_ft_hp_wt_adj final_w olsl_T_wt_ft_hp_wt_adj final_u
ol$LT_wt_ft_adj finaljw olsLT_vvt_ft_adj final_u
o No Technology Fleet n Technology Fleet
Figure 2-12 Best Fit Results for Various Regressions: Trucks, With Added Technology, OLS
Figure 2-13, below, displays regression results for the passenger car MAD fitted
curves. The technology adjustment does not have, however, the same degree of impact in
2-40
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
reducing the difference in the attained slopes (between those with and without the addition of
technology) evidenced in the OLS regressions.
25 35
- maclPC_init_w
^— maclPC_hp_wt_adj_init_w
— m a d P C_w t_f t_hp_v»' t_a clj_i n i t_ w
madPC_wt_ft_adj_init_w
o No Technology Fleet
madPC_hp_wt_adj_init_u
m a d P C_ w t_f t_hp_w t_a dj_i n i t_u
iiiadPC_wt_ft_adj_Wt_u
n Technology Fleet
Figure 2-13 Best Fit Results for Various Regressions: Cars, No Added Technology, MAD
2-41
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
75
85
maclPC_final_w
45 55 65
Footprint
madPC_final_u
v ^—madPC_hp_wt_adj_final_u
maclPC_wt_ft_hp_wt_adj_final_w madPC_wt_ft_hp_wt_adj_final_u
madPC_wt_ft_adj_final_w madPC_wt_ft_aclj_final_u
o No Technology Fleet P Technology Fleet
Figure 2-14 Best Fit Results for Various Regressions: Cars, Added Technology, MAD
The MAD regression results below in Figure 2-15 show a grouping of the fitted lines
similar to that displayed in the OLS fits for trucks. As expected, an additional reduction in
2-42
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
divergence is seen in the case where technology has been added, in Figure 2-15, which can be
ascribed to the reduction in heterogeneity of the fleet brought about by the addition of the
technology.
25 35 45
madLT_init_w
madLT_hp_Vv't_aclj_init_w
55
65
75
85
m a d LT_ w t_f t_a clj_i n i t_w
O No Technology Fleet
maclLT_h|>_wt_aclj_init_u
madLT_wt_ft_hp_wt_adj_init_u
mad LT_ w t_f t_a cl j_i n i t_u
D Technology Fleet
Figure 2-15 Best Fit Results for Various Regressions: Trucks, No Added Technology, MAD
2-43
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
0.01
25 35
•maclLT_final_w
•mad LT_h p_ w t_a dj_f i n a l_w
madLT_wt_ft_adj_final_w
O No Technology Fleet
Footprint
madLT_flnal_u
madLT_hp_wt_adj_final_u
w-^—madLT_wt_ft_hp_wt_adj_final_u
niadLT_wt_ft_adj_final_n
D Technology Fleet
Figure 2-16 Best Fit Results for Various Regressions: Trucks, with Added Technology, MAD
Updating this analysis using the revised MY 2008- and the MY 2010-based fleet
projections yielded results generally similar to those shown above. Detailed results of the
2-44
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
analyses with the final rulemaking fleet projections are presented in a memorandum available
inNHTSA'sdocket.^
2.4.2.10 Which methodology did the agencies choose for the proposal, and why was it
reasonable?
For the proposal, the choice among the alternatives presented above was to use the
OLS formulation, on sales-weighted data, using a fleet that has had technology applied, and
after adjusting the data for the effect of weight-to-footprint, as described above. The agencies
believe that this represented a technically reasonable approach for purposes of developing
target curves to define the proposed standards, and that it represents a reasonable trade-off
among various considerations balancing statistical, technical, and policy matters, which
include the statistical representativeness of the curves considered and the steepness of the
curve chosen. The agencies judged the application of technology prior to curve fitting to
provide a reasonable means—one consistent with the rule's objective of encouraging
manufacturers to add technology in order to increase fuel economy and reduce GHG
emissions—of reducing variation in the data and thereby helping to estimate a relationship
between fuel consumption/CO2 and footprint.
Similarly, for the agencies' NPRM MY 2008-based market-forecast and the agencies'
estimates of future technology effectiveness, the inclusion of the weight-to-footprint data
adjustment prior to running the regression also helped to improve the fit of the curves by
reducing the variation in the data, and the agencies believed that the benefits of this
adjustment for the proposed rule likely outweighed the potential that resultant curves might
somehow encourage reduced load carrying capability or vehicle performance (note that we
were not suggesting that we believed these adjustments would reduce load carrying capability
or vehicle performance). In addition to reducing the variability, the truck curve was also
steepened, and the car curve flattened compared to curves fitted to sales weighted data that do
not include these normalizations. The agencies agreed with manufacturers of full-size pick-up
trucks that in order to maintain towing and hauling utility, the engines on pick-up trucks must
be more powerful, than their low "density" nature statistically suggested based on the
agencies' NPRM MY 2008-based market forecast and the agencies' estimates of the
effectiveness of different fuel-saving technologies. Therefore, the agencies judged that it may
be more appropriate (i.e., in terms of relative compliance challenges faced by different light
truck manufacturers) to adjust the slope of the curves defining fuel economy and CC>2 targets.
The results of the normalized regressions are displayed in Table, below.ggg
Table 2-7 Regression Results
fff Docket No. NHTSA-2010-0131.
888 As presented in the draft TSD supporting the NPRM, this table erroneously reported coefficients from the
regression using normalization based on differences in horsepower to weight rather than differences in weight
per footprint. The differences in this Table as presented in this final TSD reflect this correction.
2-45
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Vehicle
Passenger cars
Light trucks
Slope
(gallons/mile)
0.00037782
0.00038891
Constant
(gallons/mile)
0.00181033
0.00401336
Updating this analysis using the corrected MY 2008- and the MY 2010-based fleet
projection yielded results generally similar to those shown above. Detailed results of the
analyses with the final rulemaking fleet projections are presented in a memorandum available
inNHTSA'sdocket.'"
hhh
As described above, however, other approaches are also technically reasonable, and
also represent a way of expressing the underlying relationships. The agencies revisited the
analysis for the final rule, after correcting the underlying MY 2008 based market forecast,
developing a MY 2010 based market forecast, updating estimates of technology effectiveness
and cost, and after considering relevant public comments. As presented below in section 2.6,
results of these updated analyses were generally similar to those supporting the NPRM
analysis results, and the agencies' balancing of considerations led the agencies to select final
curves unchanged from the NPRM curves.
As shown in the figures below, the line represents the sales-weighted OLS regression
fit of gallons per mile regressed on footprint, with the proposal data first adjusted by weight to
footprint, as described above. This introduces weight as an additional consideration into the
slope of the footprint curve, although in a manner that adjusts the data as described above, and
thus maintains a simple graphical interpretation of the curve in a two dimensional space
(gallons per mile and footprint).
hhh
Docket No. NHTSA-2010-0131.
2-46
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
GPM vs. Footprint - Cars
Figure 2-17 Gallons per Mile versus Footprint, Cars
(Data adjusted by weight to footprint).
2-47
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
GPM vs. Footprint - Trucks
Figure 2-18 Gallons per Mile versus Footprint, Trucks
(data adjusted by weight to footprint).
Updating this analysis using the revised MY 2008- and the MY 2010-based fleet
projection yielded results generally similar to those shown above. Detailed results of the
analyses with the final rulemaking fleet projections are presented in a memorandum available
inNHTSA'sdocket.1"
In the preceding two figures, passenger car and light truck data is represented for the
specification chosen, with the size of the observation scaled to sales. The agencies note with
regard to light trucks that for the MYs 2012-2016 analysis NPRM and final rule analyses,
some models of pickups are aggregated , when, for example, the same pickup had been
available in different cab configurations with different wheelbases.16 For the analysis
presented above, these models have been disaggregated and are represented individually,
which leads to a slightly different outcome in the regression results than had they remained
aggregated.
111 Docket No. NHTSA-2010-0131.
2-48
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
2.4.2.11 Implications of the adopted slopes compared to the slopes in MYs 2012-2016
Rules
The slope first proposed, and now adopted by the agencies has several implications
relative to the MY 2016 curves, with the majority of changes affecting the truck curve. The
selected car curve has a slope similar to that finalized in the MYs 2012-2016 rulemaking (4.7
g/mile in MY 2016, vs. 4.5 g/mile proposed in MY 2017). By contrast, the truck curve is
steeper in MY 2017 than in MY 2016 (4.0 g/mile in MY 2016 vs. 4.9 g/mile in MY 2017).
As discussed previously, a steeper slope relaxes the stringency of targets for larger vehicles
relative to those for smaller vehicles, thereby shifting relative compliance burdens among
manufacturers based on their respective product mix. Comments regarding the slope of the
agencies' proposed curves are discussed in Section II.C of the preamble to today's final rule.
2.5 Once the agencies determined the appropriate slope for the sloped part, how did the
agencies determine the rest of the mathematical function?
The agencies continue to believe that without a limit at the smallest footprints, the
function—whether logistic or linear—can reach values that would be unfairly burdensome for
a manufacturer that elects to focus on the market for small vehicles; depending on the
underlying data, an unconstrained form could result in stringency levels that are
technologically infeasible and/or economically impracticable for those manufacturers that
may elect to focus on the smallest vehicles. On the other side of the function, without a limit
at the largest footprints, the function may provide no floor on required fuel economy. Also,
the safety considerations that support the provision of a disincentive for downsizing as a
compliance strategy apply weakly, if at all, to the very largest vehicles. Limiting the
function's value for the largest vehicles thus leads to a function with an inherent absolute
minimum level of performance, while remaining consistent with safety considerations.
Just as for slope, in determining the appropriate footprint and fuel economy values for
the "cutpoints," the places along the curve where the sloped portion becomes flat, the
agencies took a fresh look for purposes of this rulemaking, taking into account the updated
market forecasts and new assumptions about the availability of technologies. The next two
sections discuss the agencies' approach to cutpoints for the passenger car and light truck
curves separately, as the policy considerations for each vary somewhat.
2.5.1 Cutpoints for Passenger Car curve
The passenger car fleet upon which the agencies based the proposed target curves for
MYs 2017-2025 was derived from MY 2008 data, as discussed above. In MY 2008,
passenger car footprints ranged from 36.7 square feet, the Lotus Exige 5, to 69.3 square feet,
the Daimler Maybach 62. In that fleet, several manufacturers offer small, sporty coupes
below 41 square feet, such as the BMW Z4 and Mini, Honda S2000, Mazda MX-5 Miata,
Porsche Carrera and 911, and Volkswagen New Beetle. Because such vehicles represent a
2-49
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
small portion (less than 10 percent) of the passenger car market, yet often have performance,
utility, and/or structural characteristics that could make it technologically infeasible and/or
economically impracticable for manufacturers focusing on such vehicles to achieve the very
challenging average requirements that could apply in the absence of a constraint, EPA and
NHTSA again proposed to cut off the sloped portion of the passenger car function at 41
square feet, consistent with the MYs 2012-2016 rulemaking. The agencies recognized that for
manufacturers who make small vehicles in this size range, putting the cutpoint at 41 square
feet creates some incentive to downsize (i.e., further reduce the size, and/or increase the
production of models currently smaller than 41 square feet) to make it easier to meet the
target. Putting the cutpoint here may also create the incentive for manufacturers who do not
currently offer such models to do so in the future. However, at the same time, the agencies
believe that there is a limit to the market for cars smaller than 41 square feet — most
consumers likely have some minimum expectation about interior volume, among other things.
The agencies thus believe that the number of consumers who will want vehicles smaller than
41 square feet (regardless of how they are priced) is small, and that the incentive to downsize
to less than 41 square feet in response to this proposal, if present, will be at best minimal. On
the other hand, the agencies note that some manufacturers are introducing mini cars not
reflected in the agencies MY 2008-based market forecast, such as the Fiat 500, to the U.S.
market, and that the footprint at which the curve is limited may affect the incentive for
manufacturers to do so.
Above 56 square feet, the only passenger car models present in the MY 2008 fleet
were four luxury vehicles with extremely low sales volumes—the Bentley Arnage and three
versions of the Rolls Royce Phantom. As in the MYs 2012-2016 rulemaking, NHTSA and
EPA therefore proposed again to cut off the sloped portion of the passenger car function at 56
square feet.JJJ
While meeting with manufacturers prior to issuing the proposal, the agencies received
comments from some manufacturers that, combined with slope and overall stringency, using
41 square feet as the footprint at which to cap the target for small cars would result in unduly
challenging targets for small cars. The agencies do not agree. No specific vehicle need meet
its target (because standards apply to fleet average performance), and maintaining a sloped
function toward the smaller end of the passenger car market is important to discourage unsafe
downsizing, the agencies thus proposed to again "cut off the passenger car curve at 41 square
feet, notwithstanding these comments.
. The agencies discuss the comments that were received for the cutpoints on both
passenger car and light truck curves in the next section.
JJJ The MY 2010 based market forecast has a similarly small number of cars above a footprint of 56 sq ft. These
nine vehicle models include 5 Rolls Royce models, a Maybach 57-S and three BMW vehicles, with fewer than
20,000 total projected sales in any model year during this timeframe.
2-50
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
2.5.2 Cutpoints for Light Truck curve
The light truck fleet upon which the agencies based the proposed target curves for
MYs 2017-2025, like the passenger car fleet, was derived from MY 2008 data, as discussed in
Section 2.4 above. In MY 2008, light truck footprints ranged from 41.0 square feet, the Jeep
Wrangler, to 77.5 square feet, the Toyota Tundra. For consistency with the curve for
passenger cars, the agencies proposed to cut off the sloped portion of the light truck function
at the same footprint, 41 square feet, although we recognized that no light trucks are currently
offered below 41 square feet. With regard to the upper cutpoint, the agencies heard from a
number of manufacturers during the discussions leading up to the proposal of the MYs 2017-
2025 standards that the location of the cutpoint in the MYs 2012-2016 rules, 66 square feet,
resulted in very challenging targets for the largest light trucks in the later years of that
rulemaking (although, because CAFE and GHG standards are based on average performance,
manufacturers to not need to ensure that every vehicle model meets its fuel economy and
GHG targets). See 76 FR at 74864-65. Those manufacturers requested that the agencies
extend the cutpoint to a larger footprint, to reduce targets for the largest light trucks which
represent a significant percentage of those manufacturers' light truck sales. At the same time,
in re-examining the light truck fleet data, the agencies concluded that aggregating pickup
truck models in the MYs 2012-2016 rule had led the agencies to underestimate the impact of
the different pickup truck model configurations above 66 square feet on manufacturers' fleet
average fuel economy and CO2 levels (as discussed immediately below). In disaggregating
the pickup truck model data, the impact of setting the cutpoint at 66 square feet after model
year 2016 became clearer to the agencies.
In the agencies' view, these comments have a legitimate basis. The agencies' market
forecast used at proposal includes about 24 vehicle configurations above 74 square feet with a
total volume of about 50,000 vehicles or less during any MY in the 2017-2025 time frame.
While a relatively small portion of the overall truck fleet, for some manufacturers, these
vehicles are a non-trivial portion of their sales. As noted above, the very largest light trucks
have significant load-carrying and towing capabilities that make it particularly challenging for
manufacturers to add fuel economy-improving/C(^-reducing technologies in a way that
maintains the full functionality of those capabilities.111 Considering manufacturer CBI and our
estimates of the impact of the 66 square foot cutpoint for future model years, the agencies
determined to adopt curves that transition to a different cut point. While noting that no
specific vehicle need meet its target (because standards apply to fleet average performance),
we believe that the information provided to us by manufacturers (i.e.., information provided
regarding the accumulated impacts, especially on manufacturers' credit balances, of CAFE
standards since MY2005 and GHG standards since MY2012) and our own analysis supported
the gradual extension of the cutpoint for large light trucks in the proposal from 66 square feet
kkk In the MY2010 based market forecast, there are 14 vehicle configurations with a total volume of 130,000
vehicles or less during any MY in the 2017-2025 time frame. This is a similarly small portion of the overall
number of vehicle models or vehicle sales.
Comments on this issue are discussed in section 0.
2-51
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
in MY 2016 out to a larger footprint square feet before MY 2025. The agencies' analyses
with regard to this topic, and how it relates to the stringency of the standards, are presented in
preamble sections HID and IV.F and summarized in preamble section II.C.
Footprint Distribution by Car and Truck
2000000 -
1500000 -
1000000 -
a 500000 -
2000000 -
1500000 -
1000000 -
500000 -
0 -
40
50 60
Footprint
70
Figure 2-19 Footprint Distribution by Car and Truck*
*Proposed truck outpoints for MY 2025 shown in red, car outpoints shown in green
Updating this analysis using the revised MY 2008- and the MY 2010-based market
forecasts yielded results generally similar to those shown above. Detailed results of the
analyses with the final rulemaking fleet projections are presented in a memorandum available
inNHTSA'sdocketmmm
The agencies proposed to phase in the higher cutpoint for the truck curve in order to
avoid any backsliding from the MY 2016 standard. A target that is feasible in one model year
should never become less feasible in a subsequent model year since manufacturers should
have no reason to remove fuel economy-improving/CO2-reducing technology from a vehicle
1 Docket No. NHTSA-2010-0131.
2-52
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
once it has been applied. Put another way, the agencies proposed to not allow "curve
crossing" from one model year to the next. In proposing MYs 2011-2015 CAFE standards
and promulgating MY 2011 standards, NHTSA proposed and requested comment on avoiding
curve crossing, as an "anti-backsliding measure."17 The MY 2016 2-cycle test curves are
therefore a floor for the MYs 2017-2025 curves. For passenger cars, which have minimal
change in slope from the MY 2012-2016 rulemakings and no change in cut points, there were
no curve crossing issues in the proposed (or final) standards.
The minimum stringency determination was done using the two-cycle curves.
Stringency adjustments for air conditioning and other credits were calculated after curves that
did not cross were determined in two-cycle space. The year over year increase in these
adjustments cause neither the GHG nor CAFE curves (with A/C) to contact the 2016 curves
when charted.
The agencies received some comments on the selection of these cutpoints. ACEEE
commented that the extension of the light truck cutpoint upward from 66 s.f to 74 s.f. would
reduce stringency for large trucks even though there is no safety-related reason to discourage
downsizing of these trucks. Sierra Club and Volkswagen commented that moving this
cutpoint could encourage trucks to get larger and may be detrimental to societal fatalities.
Global Automakers commented that the cutpoint for the smallest light trucks should be set at
approximately ten percent of sales (as for passenger cars) rather than at 41 square feet.
Conversely, IIHS commented that, for both passenger cars and light trucks, the 41 s.f.
cutpoint should be moved further to the left (i.e., to even smaller footprints), to reduce the
incentive for manufacturers to downsize the lightest vehicles.
The agencies have considered these comments regarding the cutpoint applied to the
high footprint end of the target function for light trucks, and we judge there to be minimal risk
that manufacturers would respond to this upward extension of the cutpoint by deliberately
increasing the size of light trucks that are already at the upper end of marketable vehicle sizes,
particularly as gasoline prices may continue to increase in the future. Such vehicles have
distinct size, maneuverability, fuel consumption, storage, and other characteristics which
differ from vehicles between 43 and 48 square feet, and are likely not be suited for all
consumers in all usage scenarios. Further, larger vehicles typically also have additional
production costs that make it unlikely that the sales of these vehicles will increase in response
to changes in the cutpoint. Therefore, we remain concerned that not to extend this cutpoint to
74 s.f. would fail to take into adequate consideration the challenges to improving fuel
economy and CC>2 emissions to the levels required by this final rule for vehicles with
footprints larger than 66 s.f, given their increased utility, As noted above, while
manufacturers are not required to ensure that every vehicle model meets its target, the
agencies are concerned that standards with more stringent targets for large trucks would
unduly burden full-line manufacturers active in the market for full-size pickups and other
large light trucks, as discussed earlier, and evidenced by the agencies' estimates of differences
between compliance burdens faced by OEMs active and not active in the market for full-size
2-53
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
pickups. While some manufacturers have recently indicated1111" that buyers are currently
willing to pay a premium for fuel economy improvements, the agencies are concerned that
disparities in long-term regulatory requirements could lead to future market distortions
undermining the economic practicability of the standards. Absent an upward extension of the
cutpoint, such disparities would be even greater. For these reasons, the agencies do not
expect that gradually extending the cutpoint to 74 s.f will incentivize the upsizing of large
trucks and, thus, believe there will be no adverse effects on societal safety. Therefore, we are
promulgating standards that, as proposed, gradually extend the truck curve cutpoint to 74 s.f.
We have also considered the above comments by Global Automakers and IIHS on the
cutpoints for the smallest passenger cars and light trucks. In our judgment, placing these
cutpoints at 41 square feet continues to strike an appropriate balance between (a) not
discouraging manufacturers from introducing new small vehicle models in the U.S. and (b)
not encouraging manufacturers to downsize small vehicles.
2.5.3 Once the agencies determined the complete mathematical function shape, how
did the agencies adjust the curves to develop the proposed standards and
regulatory alternatives?
The curves discussed above all reflect the addition of technology to individual vehicle
models to reduce technology differences between vehicle models before fitting curves. This
application of technology was conducted not to directly determine the proposed standards, but
rather for purposes of technology adjustments, and set aside considerations regarding
potential rates of application (i.e., phase-in caps), and considerations regarding economic
implications of applying specific technologies to specific vehicle models. The following
sections describe further adjustments to the curves discussed above, that affect both the shape
of the curve (section 2.5.3.1), and the location of the curve (2.5.3.2), that helped the agencies
determine curves that defined the proposed standards.
2.5.3.1 Adjusting for Year over Year Stringency
As in the MYs 2012-2016 rules, the agencies developed curves defining regulatory
alternatives for consideration by "shifting" these curves. For the MYs 2012-2016 rules, the
agencies did so on an absolute basis, offsetting the fitted curve by the same value (in gpm or
g/mi) at all footprints. In developing the proposal for MYs 2017-2025, the agencies
reconsidered the use of this approach, and concluded that after MY 2016, curves should be
offset on a relative basis—that is, by adjusting the entire gpm-based curve (and, equivalently,
the CC>2 curve) by the same percentage rather than the same absolute value. The agencies'
estimates of the effectiveness of these technologies are all expressed in relative terms—that is,
each technology (with the exception of A/C) is estimated to reduce fuel consumption (the
inverse of fuel economy) and CC>2 emissions by a specific percentage of fuel consumption
without the technology. It is, therefore, more consistent with the agencies' estimates of
111111 For example, in its June 11, 2012 edition, Automotive News quoted a Ford sales official saying that "fuel
efficiency continues to be a top purchaser driver." ("More MPG - ASAP", Automotive News, Jun 11, 2012.)
2-54
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
technology effectiveness to develop standards and regulatory alternatives by applying a
proportional offset to curves expressing fuel consumption or emissions as a function of
footprint. In addition, extended indefinitely (and without other compensating adjustments),
an absolute offset would eventually (i.e., at very high average stringencies) produce negative
(gpm or g/mi) targets. Relative offsets avoid this potential outcome. Relative offsets do
cause curves to become, on a fuel consumption and CC>2 basis, flatter at greater average
stringencies; however, as discussed above, this outcome remains consistent with the agencies'
estimates of technology effectiveness. In other words, given a relative decrease in average
required fuel consumption or CC>2 emissions, a curve that is flatter by the same relative
amount should be equally challenging in terms of the potential to achieve compliance through
the addition of fuel-saving technology.
On this basis, and considering that the "flattening" occurs gradually for the regulatory
alternatives the agencies have evaluated, the agencies conclude that this approach to offsetting
the curves to develop year-by-year regulatory alternatives neither re-creates a situation in
which manufacturers are likely to respond to standards in ways that compromise highway
safety, nor undoes the attribute-based standard's more equitable balancing of compliance
burdens among disparate manufacturers. The agencies sought comment on these conclusions,
and on any other means that might avoid the potential negative outcomes discussed above.
As indicated earlier, ACEEE and the Alliance both expressed support for the application of
relative adjustments in order to develop year-over-year increases in the stringency of fuel
consumption and CC>2 targets, although the Alliance also commented that this approach
should be revisited as part of the mid-term evaluation.
2.5.3.2 Adjusting for anticipated improvements to mobile air conditioning systems
The fuel economy values in the agencies' market forecasts are based on the 2-cycle
(i.e., city and highway) fuel economy test and calculation procedures that do not reflect
potential improvements in air conditioning system efficiency, refrigerant leakage, or
refrigerant Global Warming Potential (GWP). Recognizing that there are significant and cost
effective potential air conditioning system improvements available in the rulemaking
timeframe (discussed in detail below in Chapter 5), the agencies are increasing the stringency
of the target curves based on the agencies' assessment of the capability of manufacturers to
implement these changes. For the proposed CAFE standards and alternatives, an offset was
included based on air conditioning system efficiency improvements, as these improvements
are the only improvements that effect vehicle fuel economy. For the proposed GHG standards
and alternatives, a stringency increase was included based on air conditioning system
efficiency, leakage and refrigerant improvements. As discussed in Chapter 5 of the joint
TSD, the air conditioning system improvements affect a vehicle's fuel efficiency or CC>2
emissions performance as an additive stringency increase, as compared to other fuel
efficiency improving technologies which are multiplicative. Therefore, in adjusting target
curves for improvements in the air conditioning system performance, the agencies adjusted
the target curves by additive stringency increases (or vertical shifts) in the curves.
For the GHG target curves, the offset for air conditioning system performance is being
handled in the same manner as for the MYs 2012-2016 rules. For the CAFE target curves,
NHTSA for the first time is accounting for potential improvements in air conditioning system
2-55
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
performance. Using this methodology, the agencies first use a multiplicative stringency
adjustment for the sloped portion of the curves to reflect the effectiveness on technologies
other that air conditioning system technologies, creating a series of curve shapes that are
"fanned" based on two-cycle performance. Then the curves are offset vertically by the air
conditioning improvement by an equal amount at every point.
2.6 What does the agencies' updated analysis indicate?
As discussed above in Chapter 1, the agencies have used two different market
forecasts to conduct analyses supporting today's final rule. The first, referred to here as the
"MY 2008-Based Fleet Projection," is largely identical to that used for analysis supporting the
NPRM, but includes some corrections (in particular, to the footprint of some vehicle models)
discussed in Chapter 1 of this TSD. The second, referred to here as the "MY 2010-Based
Fleet Projection," is a post-proposal market forecast based on the MY 2010 fleet of vehicles;
the development of this 2010 based fleet projection is discussed in Chapter 1.
Having made these changes, the agencies repeated the normalization and statistical
analyses describe above, following the same approaches as used in the analysis supporting the
NPRM. The tables and charts that follow compare the results of NHTSA's updated analysis
to those of NHTSA's prior analysis, and compare the resultant fitted lines to the lines (one
each for passenger cars and light trucks) selected for purposes of developing the proposed
attribute-based standards. The charts below present details of the results in graphical form.
2-56
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Tialized for Technology Differences
i_
o
2
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
Yes
Yes
Table 2-8 Fitted Coefficients (Slope in gpm/sf, Intercept in gpm), Passenger Cars
4-1
*-• E
:§ & tt tt
QJ Q rtJ TO
^ £ is is oo
QJ tuO QJ QJ u_ u_
5 ';r *— *— 4-< 4-<
n -2! O O QJ QJ
0 g u. u. ^ _*
,j_j ,j_j •"— •"—
in in i- i-
QJ QJ TO TO to ~o ~O
" " ^ ^ 'w QJ QJ
£- £- >s to to
£ £ QJ .^ T3 T3 -_ TO TO
OJOJ S " >"
^^•^(--2 0 0 ^ ^ ^
"a"a"&CQ: > > , , ,
MM'QJO^ ^ ^ •£ ^ ^
— — >~^ ^ ^ CL CL CL
TO TO > to ' ' ' QJ QJ QJ
EFw^QJ QJ QJ H f H
i_
O
^
No
No
No
No
Yes
Yes
Yes
Yes
No
No
No
No
Yes
Yes
Yes
Yes
No
No
i_
o
^
No
No
No
No
No
No
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
QJ
TO
0)
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
bfl
QJ
C£.
OLS
OLS
MAD
MAD
OLS
OLS
OLS
OLS
MAD
MAD
OLS
OLS
OLS
OLS
OLS
OLS
OLS
OLS
CL
O
0)
0.000648
0.000513
0.000725
0.000359
0.000431
0.000399
0.000161
0.000264
0.001486
0.000942
0.001345
0.001109
0.000984
0.000920
0.000481
0.000669
0.000378
0.000378
CL
O
0)
0.000510
0.000464
0.000560
0.000334
0.000293
0.000351
0.000131
0.000250
0.001220
0.000959
0.001175
0.001085
0.000800
0.000890
0.000452
0.000673
0.000348
0.000362
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
CL
O
0)
000472
000502
000427
000445
000248
000398
000093
000268
001058
000995
001096
001099
000737
000933
000403
000654
000316
000371
QJ
—
-0.01027
0.00009
-0.01408
0.00610
-0.00052
0.00336
0.01155
0.00844
-0.03401
-0.00507
-0.02766
-0.01122
-0.01144
-0.00579
0.01103
0.00367
0.00181
0.00517
QJ
4-1
f—
-0.00450
0.00184
-0.00699
0.00650
0.00520
0.00508
0.01238
0.00873
-0.02131
-0.00572
-0.01974
-0.00983
-0.00299
-0.00425
0.01242
0.00358
0.00268
0.00550
QJ
4-1
f—
-0.00376
-0.00076
-0.00210
0.00076
0.00643
0.00221
0.01349
0.00736
-0.01670
-0.00944
-0.01806
-0.01259
-0.00176
-0.00785
0.01336
0.00319
0.00330
0.00440
Note 1: Coefficients selected for NPRM shown underlined.
Note 2: "MY2008-Based Fleet Projection" refers to market forecast developed using (a) MY2008 vehicle models and
characteristics, (b) AEO2011-based overall passenger car and light truck volumes, and (c) manufacturer- and segment-level
shares from forecast provided late 2009 by CSM (now owned by Global Insight).
Note 3: "MY2010-Based Fleet Projection" refers to market forecast developed using (a) MY2010 vehicle models and
characteristics, (b) AEO2012-based overall passenger car and light truck volumes, and (c) manufacturer- and segment-level
shares from forecast provided late 2011 by J.D. Power (automotive forecasting service now owned by LMC).
2-57
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
zed for Technology Differences
TO
E
o
2
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
Yes
Yes
4-1
tuO
'01
Ol
o
CL-
IO
Ol
u
CU
CU
it
Q
o
M—
T3
Ol
M
TO
E
o
2
No
No
No
No
Yes
Yes
Yes
Yes
No
No
No
No
Yes
Yes
Yes
Yes
No
No
H 1
- 1
zed for Differences in Weight/Footprint ^
TO
E
o
2
No
No
No
No
No
No
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
2-9 Fitted Coefficients (Slope in gpm/sf, Intercept in gpm), Light Trucks
4-1
10
TO
1/11/1 cu
TO TO ;:
<_> u O
cu cu u-
O O OJ
LL. LL. -^
OJ OJ ro
TO TO in -o
5 5 's 1/1
Ol .W) ~O "O -^ TO
T 4->
10
cu
TO
I/)
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
Yes
No
to
10
cu
tuO
CU
C£
OLS
OLS
MAD
MAD
OLS
OLS
OLS
OLS
MAD
MAD
OLS
OLS
OLS
OLS
OLS
OLS
OLS
OLS
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
iO 35 40 45 50 55 60 65 70 75 80
-sales-weighted OLS
-OLS
-sales-weighted MAO
-MAD
-sales-weighted OLS w/ pw norm
-OLS w/pw norm
-sales-weighted OLS w/pw andIh/'sf norm
-OLS w/ pw and ll>/sf norm
- sales-weighted MAD, no tech.
-MAD, no tech
-sales-weighted OLS, no tech.
-OLS, no tech.
"sales-weighted OLS, no tech., w/ pw norm
- OLS, no tech., w/ pw norm
sales-weigh ted OLS, no tech.. w/' pw and Ib/sf
norm
-OLS, no tech., w/ pw and Ib/sf norm
sales-weigh ted OLS w/ Ih/sf norm
OLS w/ Ib/sf norm
•coefficients selected for NPRM
Figure 2-20 Fitted Lines, Passenger Cars, NPRM Analysis
2-59
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
iO 35 40 45 50 55 60 65 70 75 80
-sales-weighted OLS
-OLS
-sales-weighted MAO
-MAD
-sales-weighted OLS w/ pw norm
-OLS w/pw norm
-sales-weighted OLS w/pw andIh/'sf norm
-OLS w/ pw and ll>/sf norm
- sales-weighted MAD, no tech.
-MAD, no tech
-sales-weighted OLS, no tech.
-OLS, no tech.
-sales-weighted OLS, no tech., w/ pw norm
- OLS, no tech., w/ pw norm
sales-weigh ted OLS, no tech.. w/' pw and Ib/sf
norm
-OLS, no tech., w/ pw and Ib/sf norm
sales-weigh ted OLS w/ Ih/sf norm
OLS w/ Ib/sf norm
•coefficients selected for NPRM
Figure 2-21 Fitted Lines, Passenger Cars, Corrected MY2008-Based Market Forecast
Note 1: Line based on coefficients selected for NPRM shown for comparison.
Note 2: "MY2008-Based Fleet Projection" refers to market forecast developed using (a) MY2008 vehicle models and
characteristics, (b) AEO2011-based overall passenger car and light truck volumes, and (c) manufacturer- and segment-level
shares from forecast provided late 2009 by CSM (now owned by Global Insight).
2-60
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Si 0.06
30 35 40 45 50 55 60 65 70 75 80
-sales-weighted OLS
-OLS
-sales-weighted MAD
-MAD
-sales-weighted OLS w/ pw noun
-OLS w/pw norm
- sales-weighted OLS w/' pw and Ib/sf norm
-OLS w/ pw and Ib/sf norm
- sales-weighted MAD, no tech.
-MAD, no tech
-sales-weighted OLS, no tech.
- OLS, no tech.
- sales-weighted OLS, no tech., w/ pw norm
-OLS, no tech., w/ pw norm
sales-weighted OLS, no tech., w/ pw and Ib/sf
norm
- OLS, no tech., w/ pw and Ib/sf norm
sales-weighted OLS w/ Ib/sf norm
OLS w/ H)/sf norm
•coefficients selected forNPRM
Figure 2-22 Fitted Lines, Passenger Cars, MY2010-Based Market Forecast
Note 1: Line based on coefficients selected for NPRM shown for comparison.
Note 2: "MY2010-Based Fleet Projection" refers to market forecast developed using (a) MY2010 vehicle models and
characteristics, (b) AEO2012-based overall passenger car and light truck volumes, and (c) manufacturer- and segment-level
shares from forecast provided late 2011 by J.D. Power (automotive forecasting service now owned by LMC).
2-61
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
^2 0.06
30 35 40 45 50 55 60 65 70 75 80
-sales-weighted OLS
-OLS
-sales-weighted MAD
-MAD
-sales-weighted OLS w/ pw noun
-OLS w/pw norm
-sales-weighted OLS w/' pw and Ib/sf norm
-OLS w/ pw and Ib/sf norm
- sales-weighted MAD, no tech.
-MAD, no tech
-sales-weighted OLS, no tech.
- OLS, no tech.
- sales-weighted OLS, no tech., w/ pw norm
-OLS, no tech., w/ pw norm
sales-weighted OLS, no tech., w/ pw and Ib/sf
norm
- OLS, no tech.. w/ pw and Ib/sf norm
sales-weigh ted OLS w/ Ib/sf norm
OLS w/ Ib/sf norm
•coefficients selected forNPRM
Figure 2-23 Fitted Lines, Light Trucks, NPRM Analysis
2-62
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Si 0.06
30 35 40 45 50 55 60 65 70 75 80
-sales-weighted OLS
-OLS
-sales-weighted MAD
-MAD
-sales-weighted OLS w/ pw noun
-OLS w/pw norm
- sales-weighted OLS w/' pw and Ib/sf norm
-OLS w/ pw and Ib/sf norm
- sales-weighted MAD, no tech.
-MAD, no tech
-sales-weighted OLS, no tech.
- OLS, no tech.
- sales-weighted OLS, no tech., w/ pw norm
-OLS, no tech., w/ pw norm
sales-weighted OLS, no tech., w/ pw and Ib/sf
norm
- OLS, no tech., w/ pw and Ib/sf norm
sales-weighted OLS w/ Ib/sf norm
OLS w/ H)/sf norm
•coefficients selected forNPRM
Figure 2-24 Fitted Lines, Light Trucks, Corrected MY2008-Based Market Forecast
Note 1: Line based on coefficients selected for NPRM shown for comparison.
Note 2: "MY2008-Based Fleet Projection" refers to market forecast developed using (a) MY2008 vehicle models and
characteristics, (b) AEO2011-based overall passenger car and light truck volumes, and (c) manufacturer- and segment-level
shares from forecast provided late 2009 by CSM (now owned by Global Insight).
2-63
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Si 0.06
•ill }5 40 45 50 55 60 65 70 75 80
Footprint (sf|
-sales-weighted OLS
-OLS
-sales-weighted MAD
-MAD
-sales-weighted OLS w/ pw noun
-OLS w/pw norm
- sales-weighted OLS w/ pw and Ib/sf norm
-OLS w/ pw and Ib/sf norm
- sales-weighted MAD, no tech.
-MAD, no tech
-sales-weighted OLS, no tech.
-• OLS, no tech.
- sales-weighted OLS, no tech., w/ pw norm
-OLS, no tech., w/ pw norm
sales-weighted OLS, no tech., w/ pw and Ib/sf
norm
- OLS, no tech., w/' pw and Ib/sf norm
sales-weighted OLS w/ Ib/sf norm
OLS w/ Ib/sf norm
•coefficients selected forNPRM
Figure 2-25 Fitted Lines, Light Trucks, MY2010-Based Market Forecast
Note 1: Line based on coefficients selected for NPRM shown for comparison.
Note 2: "MY2010-Based Fleet Projection" refers to market forecast developed using (a) MY2010 vehicle models and
characteristics, (b) AEO2012-based overall passenger car and light truck volumes, and (c) manufacturer- and segment-level
shares from forecast provided late 2011 by J.D. Power (automotive forecasting service now owned by LMC).
As discussed above, the selection of a calibrated functional form—in this case, a
specific line expressing a relationship between fuel consumption and footprint—upon which
to base attribute-based fuel economy and related GHG standards involves considering not just
the apparent range of the relevant technical relationship, but also the potential implications for
affected policy issues. The approaches described above provide a range of reasonable means
of estimating relationships between observed or adjusted fuel consumption and footprint.
Having made corrections to the MY 2008-based fleet projection, and having
developed a new MY 2010-based fleet projection, the agencies have obtained results
generally similar, albeit not identical, to those obtained for the NPRM analysis. For any given
method of estimating these lines, it is unlikely that the agencies could have obtained identical
results after changing inputs. Also, there is no reason to expect that the MY 2008- and MY
2010-based fleet projections should produce identical results. Still, these differences were
mostly small. Using both the corrected MY 2008-based passenger car market forecast and the
new MY 2010-based forecast, three techniques produced fitted passenger car lines very
close—in terms of average squared differences within the range of footprints between the
2-64
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
selected outpoints discussed above—to those selected for the NPRM: sales-weighted OLS
without normalizations for differences in power/weight or weight/footprint, sales-weighted
OLS with normalization for differences weight/footprint, and unweighted OLS with
normalizations for differences in both power/weight and weight/footprint. For light trucks,
two techniques did so for both the corrected MY 2008-based passenger car market forecast
and the post-proposal MY 2010-based forecast: unweighted OLS with normalizations for
differences in both power/weight and weight/footprint, and unweighted OLS with
normalization for differences weight/footprint. Without any normalizations applied to the set
of footprint and fuel economy values, unweighted OLS produced fitted slopes within 2% of
the values obtained through the corresponding unweighted OLS analysis conducted in support
of the NPRM. Also, as the above charts show, the resultant ranges (i.e.., areas in fuel
consumption - footprint space) spanned by these methods are similar across the NPRM
analysis and the updated analyses using the MY 2008- and MY 2010-based fleet projections.
Considering that the agencies have adopted an approach whereby regulatory
alternatives are developed by shifting fitted curves on a multiplicative basis, results of several
of the techniques evaluated here thus would produce regulatory alternatives virtually identical
to those developed for the NPRM. For the method that produced results selected for
development of the NPRM, relative adjustment of lines fitted to the corrected MY 2008-based
market forecast and the MY 2010-based market forecast produces lines that are, between the
footprint cutpoints discussed above (41-56 ft2 and 41-74 ft2 for passenger cars and light
trucks, respectively), very close to the lines fitted for the NPRM (FIGURE Label):
2-65
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
0.005
'coefficients selected forNPRM
-MY2008-Biise
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
0030
0.025
£ 0.015
0.010
0005
'coefficients selected forNPRM
-MY2008-Biise
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
vehicle attributes (e.g., power/weight or weight/footprint or, plausibly, seating capacity,
interior volume, towing capacity, etc.), and statistical techniques (e.g., unweighted, sales-
weighted, MAD, OLS). Considering (a) that the reasonable analytical techniques examined
by the agencies produce a range of fitted lines, (b) that the future composition of the light
vehicle market is subject to some uncertainty, and (c) that other aspects of the agencies'
analysis are informed by policy implications, in the agencies' judgment, there is no single
analytical method that is the sole "correct" way to establish the two fitted lines (one for
passenger cars, one for light trucks) the agencies use to specify final standards. The agencies'
updated analysis shows newly-fitted lines producing regulatory alternatives very close to the
corresponding regulatory alternatives considered in the NPRM. This confirms that the
standards are within the range of technically supportable possibilities.
While the agencies' analysis indicates that slopes spanning relatively wide ranges
could be technically supportable, the agencies note that the final car standard is very similar to
the slope of the MY 2016 standard, despite being based on a different analytical approach
than the previous rule. As explained above, the agencies have selected a truck curve differing
from that adopted for the previous rule (both slope and upper cut-point); the agencies expect
that doing so will account for the future characteristics of the larger (work) trucks, and the
manufacturers serving the future market for such trucks. The upper size cut-points for cars,
and the lower size cut-point for both cars and trucks, are the same as in the previous rule.
Without these adjustments, the agencies' believe that there would either be incentives for
manufacturers to reduce the utility of these trucks, or that the manufacturer's compliance
costs for reaching the targets would be disproportionately high (Preamble Sections III.C.5 and
HID).
Thus, in the agencies' judgment, the curves strike a reasonable and appropriate
balance between the affected policy considerations—better reflecting the reasonable
penetration rates of the technologies needed to achieve the standards and the lead time needed
for implementation of those technologies, minimizing the incentive for manufacturers to
respond to standards in ways that may either result in decreased utility or compromise safety
(by downsizing vehicles with footprints on the sloped portion of mathematical functions
defining fuel economy and GHG targets), and encouraging widespread penetration of
technologies throughout both the car and light truck fleets at reasonable cost while achieving
very significant energy and environmental benefits. Having repeated the analysis documented
in the NPRM, and having done so based on two fleets (the corrected MY 2008-based market
forecast, and the MY 2010-based market forecast), the agencies have demonstrated that, as
proposed, the passenger car and light truck curves are well within technically supportable
ranges. Slightly flatter standards would directionally have a potentially compromising effect
on the safety-related incentives reflected by the promulgated curves, and potentially force
more aggressive penetration of advanced technologies into work trucks in a way that raises
issues of both increased cost and consumer acceptance. Conversely, slightly steeper
standards would tend to increase the potential that manufacturers would respond to the
standards by increasing vehicle size beyond levels the market would otherwise demand, in
lieu of applying some fuel-saving technologies. For these reasons, the agencies are today
promulgating standards using lines matching those used to develop proposed standards for the
NPRM.
2-68
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
Additional discussion of the feasibility of the final standards is available in Preamble
section HID and IV.F.
2-69
-------
Chapter 2: What are the Attribute-Based Curves the Agencies are Adopting
References:
UQU.S.C. 32902(a)(3)(A).
2 69 FR 38958 (June 29, 2004).
3 76 FR 57106, 57162-64, (Sept. 15,2011).
4 See 74 FR at 14359 (Mar. 30, 2009).
5 75 FR at 25362.
6 See generally 74 FR at 49491-96; 75 FR at 25357-62.
7 68 FR 74920-74926.
8 74 FR 14359.
9 See 75 FR at 25458
10 75 FR at 25363
1' See 75 FR at 25359.
12 Mat 25362-63.
13 Mat 25363.
14 75 FR at 25362 and n. 64
15 75 FR at 25632/3.
16 See 75 FR at 25354
17 74 Fed. Reg. at 14370 (Mar. 30, 2009).
2-70
-------
Technologies Considered in the Agencies' Analysis
Chapter 3: Technologies Considered in the Agencies' Analysis
This Chapter of the joint TSD describes the technologies NHTSA and EPA evaluated
as potential inputs in their respective models and provides estimates of the technologies'
costs, effectiveness and availability. This Chapter also describes, in general terms, how the
agencies use these inputs in their respective models.
The agencies assume, in this analysis, that manufacturers will add a variety of
technologies to each of their vehicle model platforms in order to improve their fuel economy
and GHG performance. In order to evaluate CAFE and GHG standards and regulatory
alternatives, it is essential to understand what is feasible within the timeframe of the final rule.
Determining the technological feasibility of the MYs 2017-2025 standards requires a
thorough study of the technologies available to the manufacturers during that timeframe. This
chapter includes an assessment of the cost, effectiveness, and the availability, development
time, and manufacturability of the technologies within either the normal redesign periods of a
vehicle line or in the design of a new vehicle. As we describe below, when a technology can
be applied can affect the costs as well as the technology penetration rates (or phase-in caps)
that are assumed in the analysis.
The agencies considered technologies in many categories that manufacturers could use
to improve the fuel economy and reduce CC>2 emissions of their vehicles during the MYs
2017-2025 timeframe. Many of the technologies described in this chapter are available today,
are well known, and could be incorporated into vehicles once product development decisions
are made. These are "nearer-term" technologies and are identical or very similar to those
considered in the MYs 2012-2016 final rule analysis (of course, many of these technologies
will likely be applied to the light-duty fleet in order to achieve the 2012-2016 CAFE and
GHG standards; such technologies would be part of the 2016 reference case for this
analysis51). Other technologies considered may not currently be in production, but are under
development and are expected to be in production in the next five to ten years. Examples of
these technologies are downsized and turbocharged engines operating at combustion pressures
even higher than today's turbocharged engines, and an emerging hybrid architecture mated
with an 8 speed dual clutch transmission (DCT)—a combination that is not available today.
These are technologies which the agencies believe can, for the most part, be applied both to
cars and trucks, and which are expected to achieve significant improvements in fuel economy
and reductions in CC>2 emissions at reasonable costs in the MYs 2017 to 2025 timeframe. The
agencies note that we did not consider in our analysis technologies that are currently in an
initial stage of research because of the uncertainties involved in estimating their costs and
effectiveness and in assessing whether the technologies will be ready to implement at
significant penetration rates during the timeframe of the MY 2017-2025 standards. Examples
a The technologies in the 2016 reference fleet are projections made by EPA's OMEGA model and NHTSA's
CAFE model respectively. Some technologies may be significantly represented in this reference fleet and these
details can be found in each agency's respective RIAs.
3-2
-------
Technologies Considered in the Agencies' Analysis
of such technologies would be camless valve actuation and fuel cell vehicles.b The agencies
acknowledge that due to the relatively long period between the date of this final rule and the
timeframe of the MY 2017-2025 standards, the possibility exists that new and innovative
technologies not considered in this analysis will make their way into the fleet (perhaps even in
significant numbers). The agencies plan to assess these technologies afresh, along with all of
the technologies considered in this final rule, as part of our mid-term evaluation.
3.1 What Technologies did the agencies consider for the final 2017-2025 standards?
The technologies considered for this final rulemaking (FRM) analysis by NHTSA and
EPA are briefly described below. They fit generally into five broad categories: engine,
transmission, vehicle, electrification/accessory, and hybrid technologies. A more detailed
description of each technology, and the technology's costs and effectiveness, is described in
greater detail in section 3.4 of this TSD.
Types of engine technologies applied in this FRM analysis, consistent with the
proposal, analysis to improve fuel economy and reduce CC>2 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, and other 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.
• 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 engine 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.
b Fuel cell vehicles may be especially useful in lieu of full battery electric technology for the larger trucks.
However, the agencies are not including this technology in the final rule due to the maturity level of the
technology.
3-3
-------
Technologies Considered in the Agencies' Analysis
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 profile lobe heights.
Continuous variable valve lift - is 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
FRM, the agencies considered three levels of boosting, 18 bar brake mean
effective pressure (BMEP), 24 bar BMEP and 27 bar BMEP, 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, 24 bar BMEP is applied
with 50 percent downsizing and 27 bar BMEP is applied with 56 percent
downsizing. 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. Accordingly, with 56 percent downsizing,
the manifold absolute pressure range increases up to 2.3 bar. Ricardo states in
their 2011 vehicle simulation project report that advanced engines in the 2020-
2025 timeframe can be expected to have advanced boosting systems that increase
the pressure of the intake charge up to 3 bar1. Refer to Section 3.3.1.2.24.2 for
examples of Ricardo-modeled displacements used for turbocharged and downsized
engines in each vehicle class.
Exhaust-gas recirculation boost - increases the exhaust-gas recirculation used in
the combustion process to increase thermal efficiency and reduce pumping losses.
Levels of exhaust gas recirculation approach 25% by volume in the highly boosted
engines modeled by Ricardo (this, in turn raises the boost requirement by
approximately 25%). This technology is only applied to 24 bar and 27 bar BMEP
engines in this FRM.
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.
3-4
-------
Technologies Considered in the Agencies' Analysis
Types of transmission technologies applied in this FRM, consistent with the proposal,
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.
• Six- and seven-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.
• Eight-speed automatic transmissions - the transmission gear ratios are optimized
to enable the engine to operate in a more efficient operating range over a broader
range of vehicle operating conditions. This technology is applied after 2016.
• Shift Optimization - tries to keep the engine operating near its 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.
• 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 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 FRM analysis, consistent with the
proposal, 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 FRM analysis targeting at 10 percent
and 20 percent rolling resistance reduction respectively.
• Low-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.
• Front or 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.
3-5
-------
Technologies Considered in the Agencies' Analysis
• Aerodynamic drag reduction - is achieved by changing vehicle shape or reducing
frontal area, including skirts, air dams, underbody covers, and more aerodynamic
side view mirrors. There are two levels of aerodynamic drag reduction considered
in this FRM analysis targeting 10 percent and 20 percent aerodynamic drag
reduction respectively.
• Mass reduction- 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 to downsized engines and/or ancillary systems
(transmission, steering, brakes, suspension, etc.). The maximum mass reduction
level considered in this FRM is 20 percent.
Types of electrification/accessory and hybrid technologies applied in this FRM
include:
• Electric power steering (EPS) and electro-hydraulic power steering (EHPS) - is
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 FRM
analysis, consistent with the proposal. 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 CC>2 emissions and fuel economy when the A/C is operating.
These technologies are covered separately in Chapter 5 of this joint TSD.
• 12-volt Stop-start - also known as idle-stop or 12V micro hybrid and commonly
implemented as a 12-volt belt-driven integrated starter-generator, this is the most
basic hybrid system that facilitates idle-stop capability. Along with other enablers,
this system replaces a common alternator with an enhanced power starter-
alternator, both belt driven, and a revised accessory drive system.
• Higher Voltage Stop-Start/Belt Integrated Starter Generator (BISG) - sometimes
referred to as a mild hybrid, BISG provides idle-stop capability and launch
assistance and uses a high 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-
3-6
-------
Technologies Considered in the Agencies' Analysis
alternator which can recover braking energy while the vehicle slows down
(regenerative braking). An example of a BISG system is the GM eAssist
introduced in MY 2012. This technology was not included in the analysis for the
proposal because we had incomplete information on the technology at that time.
Since the proposal, the agencies have obtained better data on the costs and
effectiveness of this technology (see 3.4.3.5 of this joint TSD). Therefore, the
agencies have revised their technical analysis on both and found that the
technology is now competitive with the others in the CAFE model technology
decision trees and EPA's technology packages. Further, this technology has been
used for "game changing" credit for pick-up trucks and can act as a bridge
technology for strong hybrid. For these reasons, the technology is now included in
the analysis.
• 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 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 and based on simulation, when combined with a
DCT transmission, provides similar or improved fuel efficiency to other strong
hybrid systems with reduced cost.
• 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.
• Electric vehicles (EV) - are vehicles with all-electric drive and with vehicle
systems powered by energy-optimized batteries charged primarily from grid
electricity. EVs with 75 mile, 100 mile and 150 mile ranges have been included as
potential technologies.
Types of accessory/hybridization/electrification technologies discussed but not applied
in this FRM analysis, consistent with the proposal, include:
• Integrated Motor Assist (IMA)/Crank integrated starter generator (CISG) -
provides idle-stop capability and uses a high 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
wiring harness. This system replaces a standard alternator with an enhanced
power, higher voltage and higher efficiency starter-alternator that is crankshaft
mounted and can recover braking energy while the vehicle slows down
3-7
-------
Technologies Considered in the Agencies' Analysis
(regenerative braking). The IMA technology is not included as an enabling
technology in this analysis as the industry trends toward more cost effective hybrid
configurations, although it is included as a baseline technology because it exists in
the baseline fleet.
• Power-split Hybrid (PSHEV) - is a hybrid electric drive system that 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 power-split hybrid
technology is not included as an enabling technology in this analysis as the
industry is expected to trend toward more cost-effective hybrid configurations,
although it is included as a baseline technology because it exists in the baseline
fleet.
• 2-Mode Hybrid (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 CC>2 emissions at highway speeds relative to
other types of hybrid electric drive systems. The 2-mode hybrid technology is not
included as an enabling technology in this analysis as the industry is expected to
trend toward more cost effective hybrid configurations, although it is included as a
baseline technology because it exists in the baseline fleet.
3.2 How did the agencies determine the costs of each of these technologies?
3.2.1 Direct Costs0
3.2.1.1 Costs from Tear-down Studies
There are a number of technologies in this analysis that have been cost using the
rigorous tear-down method described in this section. As a general matter, the agencies
believe that the best 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
0 Note that only battery pack and non-battery costs for HEVs, EVs and PHEVs have changed since proposal.
All other direct costs are unchanged except for adjustments from 2009 to 2010 dollars. Battery pack and non-
battery cost changes are detailed in Section 3.4.3.6.
3-8
-------
Technologies Considered in the Agencies' Analysis
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. 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.
Over the past several years, EPA has contracted with FEV, Inc. and its subcontractor
Munro & Associates 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 technology configurations. FEV's methodology was
documented in a report published as part of the MY 2012-2016 rulemaking process, detailing
the costing of the first tear-down conducted in this work (#1 in the below list).2 This report
was peer reviewed by experts in the industry and revised by FEV in response to the peer
review comments.3 Subsequent tear-down studies (#2-5 in the below list) were documented
in follow-up FEV reports made available in the public docket for the MY 2012-2016
rulemaking.4
Since then, FEV's work under this contract has continued. Additional cost studies
have been completed for mild hybrid technology and are available for public review.5 The
most extensive study, performed after the MY 2012-2016 Final Rule, involved whole-vehicle
tear-downs of a 2010 Ford Fusion power-split hybrid and a conventional 2010 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
cost another hybrid technology, the P2 hybrid, which employs similar hardware. 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 and costed to provide supplemental battery costing
information to that associated with the NiMH battery in the Fusion, because we think
automakers are moving to Li-ion battery technologies due to the higher energy and power
density of these batteries. This HEV cost work, including the extension of results to P2
HEVs, has been extensively documented in a new report prepared by FEV.6 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 final report. The peer review comments and responses are available in the
3-9
-------
Technologies Considered in the Agencies' Analysis
rulemaking docket.d'eOver 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.
9. Mild hybrid with stop-start technology (Saturn Vue with 14 engine), replacing a
conventional 14 engine.
10. Fiat Multi-Air engine technology. (Although results from this cost study are
included in the rulemaking docket, they were not used by the agencies in this
rulemaking's technical analyses because the technology is under a very recently
awarded patent and we have chosen not to base our analyses on its widespread use
across the industry 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.
The agencies have relied on the findings of FEV for estimating the cost of the
technologies covered by the tear-down studies. However, we note that FEV based their costs
on the assumption that these technologies would be mature when produced in large volumes
d ICF, "Peer Review of FEV Inc. Report Light Duty Technology Cost Analysis, Power-Split and P2 Hybrid
Electric Vehicle Case Studies", EPA-420-R-11-016, November 2011.
e FEV and EPA, "FEV Inc. Report 'Light Duty Technology Cost Analysis, Power-Split and P2 Hybrid Electric
Vehicle Case Studies', Peer Review Report - Response to Comments Document", EPA-420-R-11-017,
November 2011.
3-10
-------
Technologies Considered in the Agencies' Analysis
(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 2017-2025 timeframe for fully mature, high sales volumes, we have
had FEV perform supplemental analysis to consider potential stranded capital costs, and have
included these in our primary analyses of program costs. The issue of stranded capital is
discussed in detail in Section 3.2.2.3 of this TSD.
3.2.1.2 Costs of HEV, PHEV, EV, and FCEVs
The agencies have also reconsidered the costs for HEVs, PHEVs, EVs, and FCEVs
since the MY 2012-2016 rulemaking and the Technical Assessment Report (TAR) as the
result of two issues. The first issue is that electrified vehicle technologies are developing
rapidly and we sought to capture the results from the most recent analyses. The second issue
is that the analysis for the MYs 2012-2016 final rule employed a single $/kWh ($ per
kilowatt-hour) estimate, and did not consider the specific vehicle and technology application
for the battery when we estimated the cost of the battery.g Specifically, batteries used in
HEVs (high power density applications) versus EVs (high energy density applications) need
to be considered appropriately to reflect the design differences, the chemical material usage
differences, and the differences in cost per kWh as the power to energy ratio of the battery
changes for different applications. To address these issues for this final rule, consistent with
the proposal, the agencies have used a battery cost model 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.7 The model developed by ANL
allows users to estimate unique battery pack costs using user customized input sets for
different types of electrified powertrains, such as strong hybrid, PHEV and EV. Since the
publication of the TAR, ANL's battery cost model has been peer-reviewed and ANL has
updated the model to incorporate suggestions from peer-reviewers.8 Further updates have
been made to the model since the NPRM and this newly updated model is used in this FRM
analysis.9 We discuss our updated battery costs in section in Section 3.4.3.9. As done in the
proposal, the agencies developed costs and effectiveness values for the mild and P2 HEV
configuration, two different all-electric mileage ranges for PHEVs (20 and 40 in-use miles)
and three different mileage ranges for EVs (75, 100 and 150 in-use miles). Details regarding
these vehicle technologies are discussed in sections 3.4.3.6.4 and 3.4.3.6.5.
f The potential for stranded capital occurs when manufacturing equipment and facilities cannot be used in the
production of a new technology.
8 However, we believe that this had little impact on the results of the cost analyses in support of the MYs 2012-
2016 final rule, as the agencies projected that the standards could be met with an increase of less than 2 percent
penetration of hybrid technology and no increase in plug-in or full electric vehicle technology.
3-11
-------
Technologies Considered in the Agencies' Analysis
3.2.1.3 Direct Manufacturing Costs Used in the Rulemaking Analysis
Building on the MYs 2012-2016 final rule, for the NPRM analysis, the agencies took a
fresh look at technology cost and effectiveness values. For this final rule analysis, the direct
manufacturing costs employed in the NPRM have been largely retained, although they were
updated to 2010$, and revisions were made to the costs of Li-ion batteries. The battery costs
have been updated for the final rule using the latest ANL BatPaC model as discussed above.
For costs, the agencies considered both the direct or "piece" costs and indirect costs of
individual components of technologies. For the direct costs that were not developed through
the FEV tear-down studies, the agencies generally followed a bill of materials (BOM)
approach. A bill of materials, in a general sense, is a list of components that make up a
system—in this case, an item of fuel economy-improving technology. In order to determine
what a system costs, one of the first steps is to determine its components and what they cost.
NHTSA and EPA estimated these components and their costs based on a number of
sources for cost-related information. The objective was to use those sources of information
considered to be most credible for projecting the costs of individual vehicle technologies. For
those cost estimates that are fundamentally unchanged since the 2012-2016 final rule and/or
the 2010 TAR (we make note of these in Section 3.4, below), we have a full description of the
sources used in Chapters of the final joint TSD supporting that rule.10'11 For those costs that
have been updated since those analyses (e.g., battery pack cost, costs based on more recent
tear down analyses, etc.), we note their sources in Section 3.4, below. We have also
considered input from manufacturers and suppliers gathered either through meetings
following the 2010 TAR or in comment submitted in response to the 2010 TAR, some of
which cannot be shared publicly in detailed form but, where used, we make note of it while
protecting its confidentiality. In this final rule analysis, the agencies have not updated the
costs based on any confidential information. Note that a summary of comments on the 2010
TAR, with the agencies' responses, was published as a "Supplemental Notice of Intent" in
December of 2010.12 As discussed throughout this chapter, the agencies have reviewed,
revalidated or updated cost estimates for individual components based on the latest
information available.
Once costs were determined, they were adjusted to ensure that they were all expressed
in 2010 dollars (the NPRM was in 2009 dollars) using the GDP price deflator as described in
section 3.2.4. Indirect costs were accounted for using the ICM approach developed by EPA
and explained below. NHTSA and EPA also considered how costs should be adjusted to
reflect manufacturer learning as discussed below. Additionally, costs were adjusted by
modifying or scaling content assumptions to account for differences across the range of
vehicle sizes and functional requirements, and the associated material cost impacts were
adjusted to account for the revised content, although these adjustments were different for each
agency due to the different vehicle subclasses used in their respective models.
h The conversion to 2010 dollars has very little impact on costs (the conversion factor to convert from 2009 to
2010 dollars is 1.01).
3-12
-------
Technologies Considered in the Agencies' Analysis
3.2.2 Indirect Costs'
3.2.2.1 Indirect Cost Multiplier Changes since the 2012-2016 FRM and 2010 TAR
As discussed in greater detail below, the agencies have revised the markups used to
estimate indirect costs. The first change was to normalize the ICM values to be consistent
with the historical average retail price equivalent (RPE) of 1.5, rather than the single year that
the RTI study examined. This was done by applying a factor of .57.46 to all indirect cost
elements. The second change was to re-consider the markup factors and the data used to
generate them. The result on this new thinking is to increase the markup in all cases. The
final change is the way in which the ICM factors are applied. In previous analyses ICMs
were applied to the learned value of direct costs. However, since learning influences direct
costs only, the agencies were concerned that this could overstate the impact of learning on
total costs. Indirect costs are thus now established based on the initial value of direct costs
and held constant until the long-term ICM is applied. This is done for all ICM factors except
warranties, which are influenced by the learned value of direct costs.
3.2.2.2 Cost markups to account for indirect costs
To produce a unit of output, auto manufacturers incur direct and indirect costs. Direct
costs include the cost of materials and labor costs. Indirect costs 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 goods sold. Although it is possible to account for direct costs allocated to each unit of
goods 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 EPA and NHTSA have frequently
used these multipliers to estimate the resultant impact on costs associated with manufacturers'
responses to regulatory requirements. The best approach 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.
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.
Note that our approach to estimating indirect costs remains unchanged since the proposal.
3-13
-------
Technologies Considered in the Agencies' Analysis
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 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, EPA has developed modified multipliers. These multipliers
are referred to as indirect cost multipliers (ICMs). In contrast to RPE multipliers, ICMs
assign unique incremental changes to each indirect cost contributor
ICM = (direct cost + adjusted indirect cost + profit)/(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. This
methodology was used in the cost estimation for the MYs 2012-2016 final rule. The ICMs
were developed in a peer-reviewed report from RTI International and were subsequently
discussed in a peer-reviewed journal article.13 Note that the cost of capital (reflected in profit)
is included because of 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. The capital costs are those associated with the incremental costs of the new
technologies.
As noted above, for the analysis supporting this final rulemaking, consistent with the
proposal, the agencies are again using the ICM approach but have made some changes to both
the ICM factors and to the method of applying those factors to arrive at a final cost estimate.
The first of these changes was done in response to continued thinking among the EPA-
NHTSA team about how past ICMs have been developed and what are the most appropriate
data sources to rely upon in determining the appropriate ICMs. The second change has been
done both due to staff concerns and public feedback suggesting that the agencies were
inappropriately applying learning effects to indirect costs via the multiplicative approach to
applying the ICMs.
Regarding the first change - to the ICM factors themselves - a little background must
first be provided. In the original work done under contract to EPA by RTI International,14
EPA staff with extensive experience in the auto industry had undertaken a consensus
approach to determining the impact of specific technology changes on the indirect costs of a
company. Subsequent to that effort, EPA staff, again with extensive experience in the auto
industry, conducted a blind survey to make this determination on a different set of technology
changes. This subsequent effort, referred to by EPA as a modified-Delphi approach, resulted
in slightly different ICM determinations. This effort is detailed in a memorandum contained
in the docket for this rule.15 Upon completing this effort, the EPA team determined that the
original RTI values should be averaged with the modified-Delphi values to arrive at the final
3-14
-------
Technologies Considered in the Agencies' Analysis
ICMs for low and medium complexity technologies and that the original RTI values would be
used for high complexity level 1 while the modified-Delphi values would be used for high
complexity level 2. These final ICMs as described were used in the MYs 2012-2016 light-
duty GHG/CAFE rulemaking.
More recently, EPA and NHTSA decided that the original light-duty RTI values,
because of the technologies considered for low and medium complexity, should no longer be
used and that we should rely solely on the modified-Delphi values for these complexity levels.
The original light-duty RTI study used low rolling resistance tires as a low complexity
technology example and a dual clutch transmission as a medium complexity technology.
Upon further thought, the technologies considered for the modified Delphi values (passive
aerodynamic improvements for low complexity and turbocharging with downsizing for
medium complexity) were considered to better represent the example technologies. As a
result, the modified-Delphi values became the working ICMs for low and medium complexity
rather than averaging those values with the original RTI report values. NHTSA and EPA staff
also re-examined the technology complexity categories that were assigned to each light-duty
technology and modified these assignments to better reflect the technologies that are now
used as proxies to determine each category's ICM value.
A secondary-level change was also made as part of this ICM recalculation to the light-
duty ICMs. That change was to revise upward the RPE level reported in the original RTI
report from an original value of 1.46 to 1.5 to reflect the long term average RPE. The original
RTI study was based on 2007 data. However, an analysis of historical RPE data indicates
that, although there is year to year variation, the average RPE has remained roughly 1.5.
ICMs will be applied to future year's data and therefore NHTSA and EPA staff believe that it
would be appropriate to base ICMs on the historical average rather than a single year's result.
Therefore, ICMs in this final rulemaking, consistent with the proposal, were adjusted to
reflect this average level. As a result, the High 1 and High 2 ICMs have also changed.
Table 3-1 shows both the ICM values used in the MYs 2012-2016 final rule and the
new ICM values used for the analysis supporting these final rules. 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 3-1 Indirect Cost Multipliers Used in this Analysis"
Complexity
Low
Medium
Highl
High2
20 12-20 16 Rule
Near term
1.17
1.31
1.51
1.70
Long term
1.13
1.19
1.32
1.45
This Final rule
Near term
1.24
1.39
1.56
1.77
Long term
1.19
1.29
1.35
1.50
a 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
3-15
-------
Technologies Considered in the Agencies' Analysis
Multipliers," Draft Report prepared by RTI International and Transportation Research
Institute, University of Michigan, July 2010
The second change made to the ICMs has to do with the way in which they are
applied. To date, we have applied the ICMs, as done in any analysis that relied on RPEs, as a
pure multiplicative factor. This way, a direct manufacturing cost of, say, $100 would be
multiplied by an ICM of 1.24 to arrive at a marked up technology cost of $124. However, as
learning effects (discussed below) are applied to the direct manufacturing cost, the indirect
costs are also reduced accordingly. Therefore, in year two the $100 direct manufacturing cost
might reduce to $97, and the marked up cost would become $120 ($97 x 1.24). As a result,
indirect costs would be reduced from $24 to $20. Given that indirect costs cover many things
such as facility-related costs, electricity, etc., it is perhaps not appropriate to apply the ICM to
the learned direct costs, at least not for those indirect cost elements unlikely to change with
learning. The EPA-NHTSA team believes that it is appropriate to allow only warranty costs
to decrease with learning, since warranty costs are tied to direct manufacturing costs (since
warranty typically involves replacement of actual parts which should be less costly with
learning). The remaining elements of the indirect costs should remain constant year-over-
year, at least until some of those indirect costs are no longer attributable to the rulemaking
effort that imposed them (such as R&D).
As a result, the ICM calculation has become more complex with the analysis
supporting this final rule, consistent with the proposal. We must first establish the year in
which the direct manufacturing costs are considered "valid." For example, a cost estimate
might be considered valid today, or perhaps not until high volume production is reached—
which will not occur until MY 2015 or later. That year is known as the base year for the
estimated cost. That cost is the cost used to determine the "non-warranty" portion of the
indirect costs. For example, the non-warranty portion of the medium complexity ICM in the
short-term is 0.343 (the warranty versus non-warranty portions of the ICMs are shown in
Table 3-2).
For the dual cam phasing (DCP) technology on an 14 engine we have estimated a
direct manufacturing cost of $70 in MY 2015. So the non-warranty portion of the indirect
costs would be $24.01 ($70 x 0.343). This value would be added to the learned direct
manufacturing cost for each year through 2018, the last year of short term indirect costs.
Beginning in 2019, when long-term indirect costs begin, the additive factor would become
$18.13 ($70 x 0.259). Additionally, the $70 cost in 2015 would become $67.90 in MY 2016
due to learning ($70 x (1-3%)). So, while the warranty portion of the indirect costs would be
$3.15 ($70 x 0.045) in 2015, indirect costs would decrease to $3.06 ($67.90 x 0.045) in 2016
as warranty costs decrease with learning. The resultant indirect costs for the DCP-I4
technology would be $27.16 ($24.01+$3.15) in MY 2015 and $27.07 ($24.01+$3.06) in
MY2016, and so on for subsequent years.
Table 3-2 Warranty and Non-Warranty Portions of ICMs
Complexity
Low
Near term
Warranty
0.012
Non-warranty
0.230
Long term
Warranty
0.005
Non-warranty
0.187
3-16
-------
Technologies Considered in the Agencies' Analysis
Medium
Highl
High2
0.045
0.065
0.074
0.343
0.499
0.696
0.031
0.032
0.049
0.259
0.314
0.448
There is some level of uncertainty surrounding both the ICM and RPE markup factors.
The ICM estimates used in this final rule, consistent with the proposal, 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.16 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. 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.
The International Council on Clean Transportation (ICCT) and the National
Automobile Dealers Association (NADA) commented on our use of ICMs. ICCT supported
the ICM approach as presented in the proposal, but argued for removal of sensitivity analyses
examining RPEs in NHTSA's FRIA. NADA argued that the ICM approach is not valid and
should be replaced with an RPE approach. Further, it argued that the RPE factor should be 2x
rather than the 1.5x approach that is supported by filings to the Securities and Exchange
Commission. We have conducted a thorough analysis of the NADA comments on the RPE
vs. ICM approach. We disagree with NADA's arguments for both using the RPE approach
and a 2x RPE factor, for the following reasons.
NADA's objections to the ICM approach include:
1. There is no evidence that the RPE method is flawed.
2. The ICMs do not include the total costs of complying with the standards,
because it does not include all the costs included in the RPE.
3-17
-------
Technologies Considered in the Agencies' Analysis
3. The ICMs use a subjective judgment to adjust indirect costs for different
technologies, while the RPE uses one value for all components and does not
rely on "nearly perfect foreknowledge."
4. The ICMs do not incorporate dealer and OEM profits.
NADA's arguments for the RPE of 2x include:
5. Several scholarly papers support the use of RPEs in the 2.0 range.
6. A case study comparison of the added content of a 1971 Chevrolet Vega and
2011 Cruze shows that an RPE of 2.0 accounts for the change in retail price.
The discussion above provides background on the issue of RPEs and ICMs, and on the
agencies' decision to use ICMs to estimate indirect costs for this rulemaking. Our responses
here address the specific points raised by NADA.
First, the RPE approach applies the same average indirect cost markup across all
technologies in the redesigned vehicle fleet, regardless of the source of the direct cost (i.e.
whether a technology is simple or complex; whether the source of the additional cost is a new
or a mature technology). The RPE methodology also assumes that an indirect cost is
associated with the rule, even if no relation is apparent. For instance, the RPEs (until recent
union contract changes) would have included the costs to the domestic auto companies of the
health insurance for retired auto workers. Because the rulemaking would not affect the
current retiree health care costs, (which account for about 1.5% of the RPE), they are
irrelevant to the rulemaking. The ICM approach differs in that it allows indirect costs to vary
with the complexity of the technology and the time frame.17 It is a reasonable assumption that
simple technologies are expected to have fewer indirect costs per dollar than complex
technologies. For instance, the use of low-rolling-resistance tires, considered by the
EPA/NHTSA team to be a low-complexity technology, adds costs, but, because they require
significantly less vehicle integration effort than for example, adding a hybrid powertrain
would, the additional indirect costs per dollar of direct manufacturing costs may be very low.
In contrast, converting a conventional vehicle to a hybrid-electric is a far more complex
activity, involving increases in indirect costs such as research and development
disproportionate to its direct costs. Shortly after product introduction, indirect costs for
components such as warranty and research may be relatively high, but auto makers are
expected to be able to reduce the costs of any specific technology over time, as they gain
experience with them and, thus, redirect those expenditures to other areas of their choosing.
Second, the ICM approach excludes some costs included in the RPE when those costs
are expected not to be affected by the standards. The ICM approach, as discussed above,
begins with the RPE and includes all the relevant cost categories. ICMs reflect the indirect
costs judged by the EPA panel (see above for further explanation) to be incurred for each
technology in response to regulatory imposed changes. Any "omissions", or instances where
the ICM carries no costs for a given technology, are cases where the indirect costs are
considered by the EPA panel not to be impacted by regulatory imposed changes for that
technology. For instance, the costs of switching from a standard tire to a low-rolling-
resistance tire (the example of a low-complexity technology in Rogozhin et al. (2009)) are not
expected to lead to an increase in transportation costs (i.e., costs for transporting finished
3-18
-------
Technologies Considered in the Agencies' Analysis
vehicles from production site to retail site) because it is not expected to be any more
expensive to ship a new vehicle with the new tires than with the old tires.18
Third, the RPE approach relies on the assumption that applying the average RPE for
the vehicle fleet as a whole will produce a reasonable average indirect cost for all
technologies in the redesigned vehicle fleet resulting from these standards. The agencies
believe that using the professional judgment and expertise of EPA staff with extensive
experience in the auto industry provides useful insight into how a given regulation will impact
indirect costs and is an improvement over ignoring differences among technologies. The
agencies have therefore based their central analyses on the ICM method.
Fourth, it is incorrect that the ICMs do not include profit. Although the initial ICM
report reviewed by NRC did not include OEM profit, the ICM approach applied in this
rulemaking does incorporate an allowance for profit, at the average corporate profit rate of 6%
of sales. The inclusion of profit for the Joint NPRM is discussed in the draft Technical
Support Document, and the agencies have included profit as an element of the indirect costs
for the final rulemaking as well.19
Fifth, the papers cited to support the use of an RPE of 2x are only a subset of the literature.
r\r\
The National Research Council (NRC) discusses the four studies that NADA's Exhibit A
cites in its support of an RPE of 2.0. The NRC also notes that NHTSA used an RPE of 1.5 for
its MY 2011 fuel economy rule; the NRC in 2002 used an RPE of 1.4, as did the California
Air Resources Board; and EPA has used a markup factor of 1.3. The NRC report then
discusses work done for the committee itself, doing a detailed analysis of a Honda Accord and
a Ford F-150 truck; the former had an RPE of "1.39 to market transaction price and 1.49 to
MSRP," and the latter had an RPE of "1.52 for market price and 1.54 for MSRP." Most
significantly, the NRC does not recommend an RPE of 2.0. Rather, the NRC recommends,
for technologies where the primary manufacturer of the technology is the automotive supply
base, an RPE of 1.5, except for hybrid powertrain components from the automotive supply
base, where it recommends an RPE of 1.3 due to the inclusion of several indirect costs in their
91
base estimate. Only in the case of technologies where an automotive OEM is the primary
manufacturer does the NRC recommend an RPE of 2.0.J We note, without specifically
commenting on the quality of the studies, that none of the papers NAD A cites in support of an
RPE of 2x was published in a peer-reviewed journal, and none of the studies claim to have
been peer-reviewed. In contrast, the research in Rogozhin et al. (2009) was peer-reviewed
twice: as documented in the Peer Review Report, and when it was submitted (and accepted)
for publication in the InternationalJournal of Production Economics. A full reading of the
literature on RPEs thus shows little support for a value of 2x. Further support for an average
J Importantly, application of the 2.0s RPE in the "OEM as primary manufacturer" case would be done to a
smaller direct cost since the OEM has produced the part in-house and, thus, is not paying the full supplier-level
indirect costs that would be included in a part purchased from a supplier. The end result should be a total cost
roughly equivalent or less than a 1.5x RPE applied to the supplier-produced part. If not, the manufacturer should
probably not produce in-house and should, instead, purchase parts since they would be less costly (all other
considerations being equal).
3-19
-------
Technologies Considered in the Agencies' Analysis
RPE lower than 2.0 comes from an examination of industry financial statements. NHTSA
examined industry 10-K submissions to the Securities and Exchange Commission from the
period 1972-1997.k The cost information in these submissions represents all industry
operations, including both OEM and supplier-sourced technologies. During this period, the
RPE averaged 1.5 while varying slightly, but never dropped below 1.4 or exceeded 1.6. At no
time did the average RPE approach the 2.0 value advocated by NADA. The results are
shown, together with the 2007 results from Rogozhin et al in the following figure:
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
V
4-RPE
1970
1975
1980
1985
1990
1995
2000
2005
2010
Sixth, the comparison of the Vega and the Cruze uses circular logic; it assumes its
conclusion. The direct costs of the vehicles are calculated using an RPE of 2, and the NADA
analysis then calculates a quality difference based on the change in direct costs. The
magnitude of the quality difference is then discovered to correspond to an RPE of 2, although
it is also an inevitable result of the initial assumption of an RPE of 2. The analysis provided
can be replicated with any value of RPE. This argument thus provides no evidence on the
value of the RPE.
For these reasons, we do not accept NADA's request to use an RPE of 2x., and instead
continue with our use of ICMs as the basis for our central analysis. However, the agencies
recognize that there is uncertainty regarding the impact on indirect costs of regulatorily
imposed changes. For this reason, both agencies have conducted sensitivity analyses using
different indirect cost estimates. EPA presents its sensitivities in Chapter 3 of its final RIA.
For its part, NHTSA rejects the ICCT proposal to eliminate sensitivity analyses examining the
RPE and presents the impact of using the RPE as a basis for indirect costs in its analysis in
Chapters 7 and 10 of NHTSA's FRIA. In addition, RPEs are incorporated into the
Probabilistic Uncertainty analysis in Chapter 12 of NHTSA's FRIA.
Spinney, B.C., Faigin, B.M, Bowie, N.N, Kratzke, S.R., Advanced Air Bag Systems Cost, Weight, and Lead
Time Analysis Summary Report, Contract No. DTNH22-96-0-12003, Task Orders - 001, 003, and 005.
3-20
-------
Technologies Considered in the Agencies' Analysis
3.2.2.3 Stranded capital
Because the production of automotive components is capital-intensive, it is possible
for substantial capital investments in manufacturing equipment and facilities to become
"stranded" (where their value is lost, or diminished). This would occur when the capital is
rendered useless (or less useful) by some factor that forces a major change in vehicle design,
plant operations, or manufacturer's product mix, such as a shift in consumer demand for
certain vehicle types. It can also be caused by new standards that phase-in at a rate too rapid
to accommodate planned replacement or redisposition of existing capital to other activities.
The lost value of capital equipment is then amortized in some way over production of the new
technology components.
It is difficult to quantify accurately any capital stranding associated with new
technology phase-ins under the final standards because of the iterative dynamic involved -
that is, the new technology phase-in rate strongly affects the potential for additional cost due
to stranded capital, but that additional cost in turn affects the degree and rate of phase-in for
the same or other individual competing technologies. In addition, such an analysis is very
company-, factory-, and manufacturing process-specific, particularly in regard to finding
alternative uses for equipment and facilities. Nevertheless, in order to account for the
possibility of stranded capital costs, the agencies asked FEV to perform an analysis, using
conservative assumptions, of the potential stranded capital costs associated with rapid phase-
in of technologies due to new standards, using data from FEV's primary teardown-based cost
analyses.22 Since the direct manufacturing costs developed by FEV assumed a 10 year
production life (i.e., capital costs amortized over 10 years) the agencies applied the FEV
derived stranded capital costs whenever technologies were replaced prior to being utilized for
the full 10 years. The other option would have been to assume a 5 year product life (i.e.,
capital costs amortized over 5 years), which would have increased the direct manufacturing
costs. It seems only reasonable to account for stranded capital costs in the instances where the
fleet modeling performed by the agencies replaced technologies before the capital costs were
fully amortized. The agencies did not derive or apply stranded capital costs to all
technologies only the ones analyzed by FEV. While there is uncertainty about the possible
stranded capital costs (i.e., understated or overstated), their impact would not call into
question the overall results of our cost analysis or otherwise affect the stringency of the
standards, since costs of stranded capital are a relatively minor component of the total
estimated costs of the rules.
The assumptions made in FEV's stranded capital analysis with potential for major
impacts on results are:
• All manufacturing equipment was bought brand new when the old technology
started production (no carryover of equipment used to make the previous
components that the old technology itself replaced).
10-year normal production runs: Manufacturing equipment used to make old
technology components is straight-line depreciated over a 10-year life.
• Factory managers do not optimize capital equipment phase-outs (that is, they are
assumed to routinely repair and replace equipment without regard to whether or
not it will soon be scrapped due to adoption of new vehicle technology).
3-21
-------
Technologies Considered in the Agencies' Analysis
• Estimated stranded capital is amortized over 5 years of annual production at
450,000 units (of the new technology components). This annual production is
identical to that assumed in FEV's primary teardown-based cost analyses. The 5-
year recovery period is chosen to help ensure a conservative analysis; the actual
recovery would of course vary greatly with market conditions.
FEV assembled a team of manufacturing experts to perform the analysis, using a
methodology with the following key steps for each vehicle technology scenario:
1) Identify all of the old technology components that are no longer used or that are
modified in the new technology vehicles (from the comparison bills of materials
developed in the primary teardown-based analyses).
2) For each of these components identify the manufacturing equipment and tooling
needed to make it.
3) Estimate the new-purchase $ value of each item identified in step 2.
4) Assign an "Investment Category" to each equipment item identified in step 2,
based on an assessment by FEV's experts of recoverable value:
• Flexible: Equipment can be used to manufacture new technology or other parts
(0% stranded)
• Re-Useable: Equipment can be used in alternative industries, sold at 50% of its
remaining value (50% stranded)
• Semi-Dedicated: Estimate that 50% of equipment is flexible (50% stranded)
• Dedicated: Custom manufacturing equipment (100% stranded)
5) Assign an "Investment Category" to each tooling item identified in step 2, based
on an assessment by FEV's experts of recoverable value:
• Flexible: Can be used for manufacturing new technology parts (0% stranded)
• Perishable: Frequent replacement of tooling (0% stranded)
Semi-Dedicated Tooling: Estimate that 50% of tooling is dedicated (50%
stranded)
• Dedicated: Commodity-specific (100% stranded)
6) Multiply the % stranding values from steps 4 and 5 by the $ values from step 3.
7) Multiply the results in step 6 by 70%, 50%, and 20% for 3-, 5-, and 8-year
stranding scenarios, respectively. That is, an old technology, for which production
is truncated prematurely after only 8 years, will experience the stranding of 20%
(the last 2 years of its 10-year normal production run) of its associated remaining
capital value.
8) Sum the results in step 7 to obtain overall stranded capital costs.
3-22
-------
Technologies Considered in the Agencies' Analysis
9) Divide the results in step 8 by 2,250,000 (5 years x 450,000 units/year) to obtain
$/vehicle values, applicable to new technology vehicles for the 1st 5 years of their
production due to the assumed 5-year recovery period.
The stranded capital analysis was performed for three transmission technology
scenarios, two engine technology scenarios, and one hybrid technology scenario, as shown in
Table 3-3. The methodology used by EPA in applying these results to the technology costs is
described in Chapter 3 of EPA's RIA. The methodology used by NHTSA in applying these
results to the technology costs is described in NHTSA's RIA section V.
Table 3-3 Stranded Capital Analysis Results (2010 dollars /vehicle)
Replaced
technology
6-speed AT
6-speed AT
6-speed DCT
Conventional V6
Conventional V8
Conventional V6
New
technology
6-speed DCT
8-speed AT
8-speed DCT
DSTGDI 14
DSTGDI V6
Power-split HEV
Stranded capital cost per vehicle
when replaced technology's production is ended
after:
3 years
$56
$48
$28
$57
$61
$112
5 years
$39
$34
$20
$40
$43
$80
8 years
$16
$14
$8
$16
$17
$32
DSTGDI=Downsized, turbocharged engine with stoichiometric gasoline direct injection.
3.2.3 Cost reduction through manufacturer learningl
For this final rule, consistent with the proposal, we have not changed our estimates of
learning and how learning will impact costs going forward from what was employed in the
analysis for the MYs 2012-2016 light-duty vehicle rule. However, we have updated our
terminology in an effort to clarify that we consider there to be one learning effect—learning
by doing—which results in cost reductions occurring with every doubling of production.™ In
the past, we have referred to volume-based and time-based learning. Our terms were meant
only to denote where on the volume learning curve a certain technology was—"volume-based
learning" meant the steep portion of the curve where learning effects are greatest, while
"time-based learning" meant the flatter portion of the curve where learning effects are less
pronounced. Unfortunately, our terminology led some to believe that we were implementing
two completely different types of learning—one based on volume of production and the other
based on time in production. Our new terminology—steep portion of the curve and flat
portion of curve—is simply meant to make more clear that there is one learning curve and
some technologies can be considered to be on the steep portion while others are well into the
Note that our approach to accounting for cost reduction through manufacturer learning is unchanged since the
proposal.
m Note that this new terminology was described in the recent heavy-duty GHG final rule (see 76 FR 57320). The
learning approach used in this analysis is entirely consistent with that used and described for the heavy-duty
analysis.
3-23
-------
Technologies Considered in the Agencies' Analysis
flatter portion of the curve. These two portions of the volume learning curve are shown in
Figure 3-1.
Volume Learning Curve - Steep & Flat Portions
120%
100%
20%
0%
Steep portion of volume learning curve
Flat portion of volume learning curve
Cumulative Production
Figure 3-1 Steep & Flat Portions of the Volume Learning Curve
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 like the light duty vehicle production industry that utilize many
common technologies and component supply sources. Both 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. We refer to this phenomenon as the
manufacturing learning curve.
NHTSA and EPA included a detailed description of the learning effect in the MYs
2012-2016 light-duty rule and the more recent heavy-duty rule.23 Most studies of the effect of
experience or learning on production costs appear to assume that cost reductions begin only
after some initial volume threshold has been reached, but not all of these studies specify this
3-24
-------
Technologies Considered in the Agencies' Analysis
threshold volume. The rate at which costs decline beyond the initial threshold is usually
expressed as the percent reduction in average unit cost that results from each successive
doubling of cumulative production volume, sometimes referred to as the learning rate. Many
estimates of experience curves do not specify a cumulative production volume beyond which
cost reductions would no longer occur, instead depending on the asymptotic behavior of the
effect for learning rates below 100 percent to establish a floor on costs.
In past rulemaking analyses, as noted above, both agencies have used a learning curve
algorithm that applied a learning factor of 20 percent for each doubling of production volume.
NHTSA has used this approach in analyses supporting recent CAFE rules. In its analyses,
EPA has simplified the approach by using an "every two years" based learning progression
rather than a pure production volume progression (i.e., after two years of production it was
assumed that production volumes would have doubled and, therefore, costs would be reduced
by 20 percent).11
In the MYs 2012-2016 light-duty rule and the heavy-duty GHG final rule, the agencies
employed an additional learning algorithm to reflect the volume-based learning cost
reductions that occur further along on the learning curve. This additional learning algorithm
was termed "time-based" learning in the MYs 2012-2016 rule simply as a means of
distinguishing this algorithm from the volume-based algorithm mentioned above, although
both of the algorithms reflect the volume-based learning curve supported in the literature. As
described above, we are now referring to this learning algorithm as the "flat portion" of the
learning curve. This way, we maintain the clarity that all learning is, in fact, volume-based
learning, and that the level of cost reductions depend only on where on the learning curve a
technology's learning progression is. We distinguish the flat portion of the curve from the
steep portion of the curve to indicate the level of learning taking place in the years following
implementation of the technology (see Figure 3-1). The agencies have applied learning
effects on the steep portion of the learning curve for those technologies considered to be
newer technologies likely to experience rapid cost reductions through manufacturer learning,
and learning effects on the flat portion learning curve for those technologies considered to be
more mature technologies likely to experience only minor cost reductions through
manufacturer learning. As noted above, the steep portion learning algorithm results in 20
n To clarify, EPA has simplified the steep portion of the volume learning curve by assuming that production
volumes of a given technology will have doubled within two-years time. This has been done largely to allow for
a presentation of estimated costs during the years of implementation, without the need to conduct a feedback
loop that ensures that production volumes have indeed doubled. If we were to attempt such a feedback loop, we
would need to estimate first year costs, feed those into OMEGA, review the resultant technology penetration rate
and volume increase, calculate the learned costs, feed those into OMEGA (since lower costs would result in
higher penetration rates, review the resultant technology penetration rate and volume increase, etc., until an
equilibrium was reached. To do this for all of the technologies considered in our analysis is simply not feasible.
Instead, we have estimated the effects of learning on costs, fed those costs into OMEGA, and reviewed the
resultant penetration rates. The assumption that volumes have doubled after two years is based solely on the
assumption that year two sales are of equal or greater number than year one sales and, therefore, have resulted in
a doubling of production. This could be done on a daily basis, a monthly basis, or, as we have done, a yearly
basis.
3-25
-------
Technologies Considered in the Agencies' Analysis
percent lower costs after two full years of implementation {i.e.., the MY 2016 costs would be
20 percent lower than the MYs 2014 and 2015 costs). Once two steep portion learning steps
have occurred, flat portion learning at 3 percent per year becomes effective for 5 years.
Beyond 5 years of learning at 3 percent per year, 5 years of learning at 2 percent per year,
then 5 at 1 percent per year become effective.
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 we therefore assume
that learning impacts have already occurred. The steep portion learning algorithm 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, the lower flat portion learning algorithm has been applied.
The learning algorithms applied to each technology and the applicable timeframes are
summarized in Table 3-4.
Table 3-4 Learning Effect Algorithms Applied to Technologies Used in this Analysis
Technology
Engine modifications to accommodate low
friction lubes
Engine friction reduction - level 1 & 2
Lower rolling resistance tires - level 1
Low drag brakes
Secondary axle disconnect
Electric/Plug-in vehicle battery charger
installation labor
Variable valve timing
Variable valve lift
Cylinder deactivation
Stoichiometric gasoline direct injection
Aggressive shift logic - level 1 & 2
Early torque converter lockup
5/6/7/8 speed auto transmission
6/8 speed dual clutch transmission
High efficiency gearbox
Improved accessories - level 1 & 2
Electronic/electro-hydraulic power steering
Aero improvements - level 1 & 2
Conversion to DOHC without reducing # of
cylinders
Air conditioner related hardware
Air conditioner alternative refrigerant
Cooled EGR
Conversion to Atkinson cycle
Turbocharging & downsizing
Mass reduction
Steep learning
2016-2020
Flat learning
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
2012-20205
2021-2025
2012-2025
2012-2025
2012-2025
2012-2025
No learning
2012-2025
2012-2025
2012-2025
2012-2025
2012-2025
3-26
-------
Technologies Considered in the Agencies' Analysis
Advanced diesel
Hybrid/Electric/Plug-in vehicle non-battery
components
P2 Hybrid vehicle battery-pack components
Electric/Plug-in vehicle battery -pack
components
Electric/plug-in vehicle battery charger
components
Stop-start
Lower rolling resistance tires - level 2
2012-2016
2012-20253
2012-2025a
2012-2015
2017-2021
2012-2025
2012-2025
2017-2025
2016-2025
2022-2025
Note that the steep learning effects have for EV and PHEV battery packs and charger components have been
carried through 5 learning cycles but at a decelerated pace as described in the text.
The learning effects discussed here impact the technology costs in that those
technology costs for which learning effects are considered applicable are changing throughout
the period of implementation and the period following implementation. For example, some of
the technology costs considered in this analysis are taken from the MYs 2012-2016 light-duty
rule. Many of the costs in the MYs 2012-2016 light-duty rule were considered "applicable"
for the 2012 model year. If flat-portion learning were applied to those technologies, the 2013
cost would be 3 percent lower than the 2012 cost, and the 2014 model year cost 3 percent
lower than the 2013 cost, etc. As a result, the MYs 2017-2025 costs for a given technology
used in this analysis reflect those years of flat learning and would not be identical to the 2012
model year cost for that same technology presented in the MYs 2012-2016 light-duty rule.
Because of the nature of battery pack development (i.e., we are arguably still in the
research phase for the types of batteries considered in this final rule, and cost reduction
through manufacturer-based learning has only just begun), the agencies have carried the
learning curve through five steep learning steps although at a somewhat slower pace than
every two years. This has been done in an effort to maintain the shape of a traditional
learning curve. This curve was developed by using the ANL BatPaC model costs as direct
manufacturing costs applicable in the 2025 MY. We have then unlearned those costs back to
2012 using the curve shown in Figure 3-2. This is the same curve used in the 2010 TAR (see
2010 TAR at page B-22). This allows the agencies to estimate costs in MYs 2017 through
2025, as well as those costs in each year back to MY 2012, if desired. As noted, this learning
curve consists of 5 full learning steps on the steep portion of the learning curve, each of which
results in costs being reduced 20 percent relative to the prior step. These learning steps are
shown occurring every two years beginning in 2012 until 2020, at which time a 5 year gap is
imposed until 2025 when the fifth steep learning step occurs. Beyond 2025, learning on the
flat portion of the curve begins at 3 percent per year cost reductions. The smooth line shows a
logarithmic curve fit applied to the learning curve as the agencies' cost model would apply
learning.
3-27
-------
Technologies Considered in the Agencies' Analysis
4.000
3.500
0.000
2012
2016
•Cost Model
2020 2024
•Log. (Cost Model)
2028
Figure 3-2 Learning Curve used for EV & PHEV Battery-Packs and In-Home Charger Costs
Note that the effects of learning on individual technology costs can be seen in the cost
tables presented in section 3.3, below. For each technology, we show direct manufacturing
costs for the years 2017 through 2025. The changes shown in the direct manufacturing costs
from year-to-year reflect the cost changes due to learning effects.
3.2.4 Costs Updated to 2010 Dollars0
This change is simply to update any costs presented in earlier analyses to 2010 dollars
using the GDP price deflator as reported by the Bureau of Economic Analysis on February 9,
2012. The factors used to update costs from 2007, 2008 and 2009 dollars to 2010 dollars are
shown below.
Price Index for Gross Domestic Product
Factor applied to convert to 2010 dollars
2007
106.2
1.04
2008
108.6
1.02
2009
109.7
1.01
2010
111.0
1.00
Source: Bureau of Economic Analysis, Table 1.1.4. Price Indexes for Gross Domestic Product,
downloaded 2/9/2012, last revised 1/27/2012.
0 Note that costs in the proposal were in terms of 2009 dollars.
3-28
-------
Technologies Considered in the Agencies' Analysis
3.3 How did the agencies determine effectiveness of each of these technologies?
The agencies determined the effectiveness of each individual technology with a
process similar to the one used for the 2012-2016 light duty vehicle GHG and CAFE
standards. The individual effectiveness of several technologies discussed in this rule that
were present in the earlier rule were left largely unchanged while others were updated. EPA
and NHTSA reviewed recent confidential manufacturer estimates of technology effectiveness
and found them to be generally consistent with our estimates. Additionally, EPA used vehicle
simulation modeling to gain further insight on existing and new technologies for this
rulemaking. EPA conducted a vehicle simulation project (described in 3.3.1) that included a
majority of the technologies, the results of which:
• informed existing individual technology effectiveness values,
• provided data for newly introduced technologies, and
• most importantly, provided an interactive data source with which to update and
calibrate the new LP model
The lumped parameter model then served as the primary tool in evaluating the
individual technology effectiveness estimates the combined effectiveness of groups of
technologies (or packages) and synergy factors, as described in 3.3.2. The effectiveness
values, in conjunction with costs, were then applied to vehicles across the fleet for use in the
Agencies' respective compliance models. For the final rule, NHTSA conducted a vehicle
simulation project with Argonne National Laboratory (ANL), as described in NHTSA's FRIA
that performed additional analyses on mild hybrid technologies and advanced transmissions to
help NHTSA develop effectiveness values better tailored for the CAFE model's incremental
structure. The effectiveness values that were developed by ANL for the mild hybrid vehicles
were applied by both agencies for the final rule. Additionally, NHTSA updated the
effectiveness values of advanced transmissions coupled with naturally-aspirated engines
based on ANL's simulation work for the final rule.
3.3.1 Vehicle simulation modeling
3.3.1.1 Background
For regulatory purposes, the fuel economy of any given vehicle is determined by
placing the vehicle on a chassis dynamometer (akin to a large treadmill that puts the vehicle's
wheels in contact with one or more rollers, rather than with a belt stretched between rollers) in
a controlled environment, driving the vehicle over a specific driving cycle (in which driving
speed is specified for each second of operation), measuring the amount of carbon dioxide
emitted from the vehicle's tailpipe, and calculating fuel consumption based on the density and
carbon content of the fuel.
One means of determining the effectiveness of a given technology as applied to a
given vehicle model would be to measure the vehicle's fuel economy on a chassis
dynamometer, install the new technology, and then re-measure the vehicle's fuel economy.
However, most technologies cannot simply be "swapped out," and even for those that can,
3-29
-------
Technologies Considered in the Agencies' Analysis
simply doing so without additional engineering work may change other vehicle characteristics
(e.g., ride, handling, performance, etc.), producing an "apples to oranges" comparison.
Some technologies can also be more narrowly characterized through bench or engine
dynamometer (i.e., in which the engine drives a generator that is, in turn, used to apply a
controlled load to the engine) testing. For example, engine dynamometer testing could be
used to evaluate the brake-specific fuel consumption (e.g., grams per kilowatt-hour) of a
given engine before and after replacing the engine oil with a less viscous oil. However, such
testing does not provide a direct measure of overall vehicle fuel economy or changes in
overall vehicle fuel economy.
For a vehicle that does not yet exist, as in the agencies' analyses of CAFE and GHG
standards applicable to future model years, even physical testing can provide only an estimate
of the vehicle's eventual fuel economy. Among the alternatives to physical testing,
automotive engineers involved in vehicle design make use of computer-based analysis tools,
including a powerful class of tools commonly referred to as "full vehicle simulation." Given
highly detailed inputs regarding vehicle engineering characteristics, full vehicle simulation
provides a means of estimating vehicle fuel consumption over a given drive cycle, based on
the explicit representation of the physical laws governing vehicle propulsion and dynamics.
Some vehicle simulation tools also incorporate combustion simulation tools that represent the
combustion cycle in terms of governing physical and chemical processes. Although these
tools are computationally intensive and required a great deal of input data, they provide
engineers involved in vehicle development and design with an alternative that can be
considerably faster and less expensive than physical experimentation and testing.
Properly executed, methods such as physical testing and full vehicle simulation can
provide reasonably (though not absolutely) certain estimates of the vehicle fuel economy of
specific vehicles to be produced in the future. However, when analyzing potential CAFE and
GHG standards, the agencies are not actually designing specific vehicles. In this rulemaking
analysis, the agencies have considered the implications of new standards that will apply to the
average performance of manufacturers' entire production lines. For this type of analysis,
precision in the estimation of the fuel economy of individual vehicle models is not essential;
although it is important that the agency avoid systematic upward or downward bias,
uncertainty at the level of individual models is mitigated by the fact that compliance with
CAFE and GHG standards is based on average fleet performance.
DOT's CAFE model and EPA's OMEGA are not full vehicle simulation models.
Both models use higher-level estimates of the efficacy of different technologies or technology
packages. Both models apply methods to avoid potential double-counting of efficacy
addressing specific energy loss mechanisms (e.g., pumping losses), and for this FRM,
consistent with the proposal, both agencies applied estimates using EPA's lumped parameter
model, which was updated using results of full vehicle simulation performed by Ricardo,
PLC. Although full vehicle simulation could, in principle, be fully integrated into the
agencies' model-by-model analyses of the entire fleet to be projected to be produced in future
model years, this level of integration would be infeasible considering the size and complexity
of the fleet. Also, considering the forward-looking nature of the agencies' analyses, and the
amount of information required to perform full vehicle simulation, this level of integration
3-30
-------
Technologies Considered in the Agencies' Analysis
would involve misleadingly precise estimates of fuel consumption and CC>2 emissions. Still,
while the agencies have used results of full vehicle simulation to inform the development of
model inputs for performing fleet-level analysis, information from other sources (e.g., vehicle
testing) could be considered when developing such model inputs.
3.3.1.2 2011 Ricardo Simulation Study
For this rule EPA built upon its 2008 vehicle simulation project24 used to support the
2012-2016 light duty vehicle GHG and CAFE standards. As in the initial project, the
technical work was conducted by the global engineering consulting firm, Ricardo, Inc. (under
subcontract to SRA Corporation), using its MSC.EASY5 dynamic vehicle simulation model.
This section is intended to supplement the main report which has been recently published and
peer-reviewed1. While this project represents a new round of full-scale vehicle simulation of
advanced technologies, the scope has also been expanded in several ways to broaden the
range of vehicle classes and technologies considered, consistent with a longer-term outlook
through model years 2017-2025. The expanded scope also includes a new analytical tool
(complex systems analysis tool) to assist in interpolating the response surface modeling
(RSM) data and visualizing technology effectiveness. This tool was especially useful in
isolating effectiveness trends during development of the updated Lumped Parameter model.
The agencies try to use publicly available information as the basis for technical
assessments whenever possible. Because these standards extend to MY 2025, and include
some technologies that are not currently in production and for which there is limited
information available in the literature, some of the technology inputs used to estimate
effectiveness are based on confidential business information. This includes the inputs related
to the technologies listed below which were based on confidential business information
belonging to Ricardo, Inc, and their expert judgment that contributed to projecting how these
technologies might improve in the future. The agencies have also considered information
which is in the public domain, in particular for turbo-charged, downsized GDI engines as
discussed in Section 3.4.1.8, as well as confidential information on engine and transmission
technologies from automotive suppliers which directionally was in line with the information
considered by Ricardo. In the draft TSD, the agencies encouraged commenters to submit
technical information, preferably that may be released publicly, related to these technologies,
particularly on their effectiveness and ability to be implemented in a way that maintains
utility. The agencies sought comment and data on the following technologies individually or
in combination: advanced turbocharged and downsized, atkinson, and advanced diesel (e.g.
projected BSFC maps) engines, hybrid powertrain control strategies, optimized transmission
shift control strategies, and transmission efficiency improvement. Few comments were
received specific to these technologies, although the Alliance emphasized that the agencies
should examine the progress in the development of powertrain improvements as part of the
mid-term evaluation and determine if researchers are making the kind of breakthroughs
anticipated by the agencies for technologies like high-efficiency transmissions. Additionally,
Volkswagen commented that while high BMEP (27-31) bar engines with cooled EGR are
currently the subject of research, Volkswagen believed that there are significant obstacles,
such as thermal and mechanical loads and their impacts on costs and durability, low-end
torque performance and part-load efficiency, which need to be overcome before these engines
represent a viable option for improving fuel economy while maintaining customer
3-31
-------
Technologies Considered in the Agencies' Analysis
satisfaction. The agencies recognize Volkswagen's comments, but note that the analysis for
this final rule considered only high BMEP engines up to 27 bar, and will be monitoring the
progress of this technology carefully and consider it at the mid-term evaluation. Moreover,
since this technology does not reach significant levels in our modeling analyses of the final
standards until after MY 2021, the agencies will evaluate industry experience with this
technology at the mid-term evaluation and can adjust assumptions as appropriate.
Below is a summary of the significant content changes from the 2008 simulation
project to the 2011 simulation project that supports the final rule, consistent with the proposal.
3.3.1.2.1 More Vehicle Classes
Two additional vehicle classes were considered, for a total of seven classes: a small
car (subcompact) and a medium/heavy duty truck class. The inclusion of the small car class
increased the fidelity of the results by capturing engineering differences unique to the smallest
vehicles in the market. The inclusion of the medium/heavy duty truck was meant primarily to
support EPA's analysis for the Heavy Duty GHG Rule25. It is worth noting that these vehicle
classes are for simulation purposes only and are not be confused with regulatory classes,
OMEGA classes, or NHTSA's technology subclasses for CAFE modeling.
3.3.1.2.2 More engine and vehicle technologies
The original 2008 project modeled several engine and transmission technologies that
were expected to become commercially available within the 2012-2016 time frame. These
technologies included advanced valvetrain technologies (such as variable valve timing and
lift, cylinder deactivation), turbocharged and downsized engines, as well as 6 speed automatic
transmissions, CVTsp and dual-clutch transmissions. The current project built on top of this
effort with the inclusion of several new engine and vehicle technologies. Highlighted
examples included:
• Advanced, highly downsized, high BMEPq turbocharged engines
• High efficiency transmissions with 8 speeds and optimized shift strategies to
maximize vehicle system efficiency
• Atkinson-cycle engines for hybrids
• Stop-start (or idle-off) technology
A discussion of these technologies is included Section 3.3.1.2, and also in the 2011
vehicle simulation report1.
p Continuously variable transmissions
q BMEP refers to brake mean effective pressure, a common engineering metric which describes the specific
torque of an engine, as a way of comparing engines of different sizes. It is usually expressed in units of bar, or
kPa, Current naturally aspirated production engines typically average 10-12 bar BMEP, while modern
turbocharged engines are now exceeding 20 bar BMEP with regularity. Simply put, a 20 bar BMEP
turbocharged engine will provide twice the torque of an equivalently-sized engine that achieves 10 bar BMEP.
3-32
-------
Technologies Considered in the Agencies' Analysis
3.3.1.2.3 Includes hybrid architectures
For the first time, this new work includes modeling of hybrid architectures for all
vehicle classes. Two main classes of hybrids were considered:
• Input powersplit hybrids. Examples of input powersplits in the market today
include the Ford Fusion HEV and the Toyota Prius.
• P2 hybrids. An example of the P2 hybrid is the Hyundai Sonata Hybrid.
While input powersplit hybrids remain a very likely hybrid architecture choice for
some manufacturers, the agencies focused solely on P2 hybrids compared to powersplit
hybrids due to their apparent cost-effectiveness advantage in future years.
Ricardo proprietary methodology was used to develop control strategies for each
architecture, the details of which can be found in section 6.8 of the 2011 project report1.
3.3.1.2.4 Complex systems tool for data analysis
In the original 2008 project, EPA staff selected unique technology packages, based on
engineering judgment, to cover a representative subset of possible vehicle options ending in
MY 2016. The expanded project time horizon (through MY 2025) and increased complexity
of potential vehicle technology interactions (including hybrids) made package selection much
more difficult. To account for unforeseen results and trends which might exist, EPA and
Ricardo adopted a complex systems approach, which is a rigorous computational strategy
designed to mathematically account for multiple input variables and determine the
significance of each (the complex systems approach is described in further detail in the 2011
Ricardo report). As a comparison, in the 2008 study, twenty-six unique technology packages
spanning five vehicle classes were selected by EPA staff and then modeled. For this project a
set of core technology packages were chosen for each vehicle class, constituting a total of 107
unique vehicle packages ("nominal runs"), which are shown as Table 3-5 and Table 3-6 in
3.3.1.2.8. A neural network Complex Systems approach to design of experiments (DOE) was
then applied to generate a set of response surface models (RSM), in which several input
parameters were varied independently over a specified range to identify the complex
relationship between these inputs and the vehicle performance. Using these methods, the
vehicle simulation was run for a set of discrete input variables chosen based on a full factorial
analysis, using a computationally efficient algorithm to select each input variable within the
design space, allowing for subsequent statistical regression of the output variables. This
approach resulted in an average of approximately two thousand independent simulation runs
for each of the 100+ vehicle packages, the outputs of which were interpolated in the data
analysis tool developed for this modeling activity. For each of these nominal and DOE runs
Ricardo provided detailed 10-hz output data csv files for reviewr.
r Stakeholders wishing to obtain this data may contact EPA to arrange for transfer of the data. Due to the
considerable size of the files (2 terabytes), stakeholders must supply their own storage media.
3-33
-------
Technologies Considered in the Agencies' Analysis
An interactive Complex Systems analysis and visualization tool was developed to
interpret the vast arrays of RSM data generated as part of the project. It was created to sample
a selected portion of the design space populated using the DOE approach described above,
and then interpret the RSM data set in a form that could be used to calibrate the lumped
parameter model (reference the equivalent-performance results in Section 3.3.1.2.18). For
more detail on the use of the RSM tool, refer to the 2011 Ricardo report1.
3.3.1.2.5 Process
The core technical work, completed in February 2011, consisted of the following
steps:
Definition of project scope
Selection of vehicle classes and baseline vehicle characteristics
Selection of vehicle architectures and individual technologies
Selection of swept variables for use in the RSM matrix
Selection of vehicle performance metrics
Review and revision of the input assumptions and modeling process
Build and run the baseline EASY5 vehicle models
Review of baseline runs and checking for errors
Build and run the nominal technology package EASY5 vehicle models
Review results and debug
Run complete DOE matrix for each technology package
Incorporation of DOE results into RSM tool
3.3.1.2.6 Definition of project scope
At project initiation, an advisory committee was formed and led by EPA to help guide
the analysis. The advisory committee consisted of technical experts from CARB and The
ICCT, the latter of which co-founded the project. A complete list of advisory committee
members is found in the vehicle simulation project report1. The committee agreed upon the
underlying ground rules, reviewed modeling assumptions and identified the desired vehicle
architectures and selected technologies for review. The boundaries for the project are
highlighted (quoted) below:
• A total of seven vehicle classes will be included: small car, standard car, large car,
small and large MPVs (multi-purpose vehicles), truck and HD truck
• LDV technologies must have the potential to be commercially deployed in the
MYs 2020-2025 timeframe
• Vehicle sizes (footprint and interior space) for each class will be largely
unchanged from MY 2010 to MYs 2020-2025
3-34
-------
Technologies Considered in the Agencies' Analysis
• Hybrid vehicles will use an advanced hybrid control strategy, focusing on battery
state-of-charge management, but will not compromise vehicle drivability
• Ricardo simulation study uses certification gasoline and 40 cetane pump diesel to
determine the effectiveness of engine technologies. The certification gasoline
typically has an RON of approximately 95 versus approximately 91 for regular
grade 87 anti-knocking index gasoline.
• It is assumed that MYs 2020-2025 vehicles will meet future California LEV III
requirements for criteria pollutants, approximately equivalent to current SULEV II
(or EPA Tier 2 Bin 2) emissions levels
• Changes in vehicle road loads including mass, aerodynamic drag, and rolling
resistance, will not be accounted for in any of the modeled technologies. Instead,
changes in vehicle road loads may be addressed through user-specified continuous
input variables in the Complex Systems tool.
The committee also decided that the following technologies fell outside the scope of
the project, either due to project resource limitations, lack of sufficient input data, or a low
potential to be commercially deployed in the timeframe considered:
• Charge-depleting powertrains (e.g. plug-in hybrids and electric range-extended
vehicles) and electric vehicles
• Fuel cell-powered vehicles
• Non-reciprocating internal combustion engines or external combustion engines
• Manual transmissions and single-clutch automated manual transmissions (AMTs)
• Kinetic energy recovery systems other than battery systems
• Intelligent vehicle-to-vehicle (V2V) and vehicle-to-infrastructure optimization
technology
• Bottoming cycles (such as organic Rankine cycles) for energy recovery
• Vehicle safety systems or structures will not be explicitly modeled for vehicles, as
it is beyond the scope of the study
The committee also selected a set of swept input variables (vehicle parameters) which
were considered most important to vehicle fuel economy and performance (swept variables
are continuously variable input values that affect vehicle output efficiency in a smooth
function for the response surface model). These variables consisted of engine displacement,
final drive ratio, electric drive motor size (for hybrids), as well as road load factors (vehicle
mass, aerodynamic drag, and rolling resistance). All of these input variables were
3-35
-------
Technologies Considered in the Agencies' Analysis
randomized in each vehicle design of experiment matrix and then incorporated into the post-
processing RSM data visualization tool.
3.3.1.2.7 Selection of vehicle classes and baseline vehicle characteristics
In order to estimate both technology costs and CO2 reduction estimates, it is necessary
to describe the baseline vehicle characteristics as the basis from which comparisons may be
drawn. In the MYs 2012-2016 light-duty vehicle rule the vehicle baseline was defined as
having a naturally aspirated gasoline engine with a port-fuel injection system, two intake and
two exhaust valves and fixed valve timing and lift; the baseline transmission was a
conventional 4-speed automatic, with no hybrid systems. These vehicles are referred to
throughout this section as the "2008 baselines." For the present study, EPA and Ricardo
elected to include a set of "2010 baseline" technology vehicles, which reflect MY 2010 trends
in engine and vehicle technology as well as some technologies that are expected to be
widespread within a few years. It is important to note that the 2010 baseline vehicles in the
Ricardo study do not reflect the technology content of the baseline fleet vehicles used by each
agency in their respective compliance modeling. The Ricardo 2010 baseline vehicles are only
used in the analysis required to establish effectiveness and synergies in the lumped parameter
model. The 2010 baseline vehicles all include an engine with dual overhead camshaft and
dual-independent intake/exhaust valve timing, a six-speed automatic transmission, 12-volt
idle off (stop-start) functionality and an alternator with partial energy regeneration capability.
There is no change in the engine displacement or vehicle road load coefficients between the
2008 baseline and the 2010 baseline vehicles. For a table showing the 2010 baseline vehicle
characteristics refer to Appendix 3 of the 2011 Ricardo report1.
In the Ricardo study, seven vehicle classes were selected for the analysis, in order to
more fully represent the broad groupings of a wide variety of products offered in the US
passenger car and light-duty truck market. The seven vehicle categories chosen were as
follows:
• Small car: a subcompact car typically powered by a small 4 cylinder engine.
• Standard car: a midsize car typically powered by a small 6 cylinder engine.
• Large car: a large passenger car typically powered by a large 6 cylinder engine.
• Small MPV: a small multi-purpose vehicle (MPV) or "crossover" vehicle typically
powered by a 4 cylinder engine
• Large MPV: a minivan or large MPV or "crossover" unibody constructed vehicle
with a large frontal area, typically powered by a 6 cylinder engine, capable of carrying
~ 6 or more passengers.
• Large truck (1/2 ton): large sports-utility vehicles and large pickup trucks, typically a
ladder-on-frame construction, and typically powered by an 8 cylinder engine.
• Class 2b/3 truck (3/4 ton): a large pickup truck (although with a GVW no greater than
8.500 pounds) with a heavier frame intended to provide additional utility (a.k.a.
"work" truck), typically powered by a larger 8 cylinder gasoline or diesel engine.
3-36
-------
Technologies Considered in the Agencies' Analysis
3.3.1.2.8 Technology selection
Ricardo presented the committee with an array of potential technologies that might
become commercially viable and present in the light-duty market by MY 2025. EPA and the
Advisory Committee suggested additional other technologies, e.g. Atkinson engines for
hybrids, fast engine warm-up strategies, etc, to consider in the selection process. The
complete set of potential technologies can be found in Appendix 2 of the 2011 Ricardo
report1. After further deliberation within the committee and by Ricardo, a subset of
technologies considered most promising (from a technical feasibility and cost effectiveness
standpoint) was selected by the committee and Ricardo for inclusion in the project test matrix.
The technologies were distributed among four distinct vehicle architectures. These
architectures represented unique EASY5 model structures, and are listed below:
• 2010 Baseline vehicles: intended to represent physical replicas of existing vehicle
models, although some minor additional content was included (as described in
Section 3.3.1.2.7)
• Conventional stop-start: vehicles for the MYs 2020-2025 timeframe that included
advanced engines but did not incorporate an electric drive or braking energy
recovery. These vehicles all contained a 12 volt stop-start (or idle-off) capability,
along with the following technologies further detailed in the 2011 Ricardo
simulation study8:
o higher efficiency gearbox (2020 timeframe)
o optimized shift strategy (best BSFC)
o alternator regeneration (during braking)
o high-efficiency alternator
o advanced engine warmup technologies
o engine friction reduction (+3.5% fuel consumption reduction over
2008 baseline)
• P2 hybrid: represent a class of hybrids in which the electric drive motor is coupled
via a clutch directly to the transmission input shaft. An existing vehicle in the
market which most closely represents this architecture is the 2011 Hyundai Sonata
Hybrid except that Ricardo recommended a P2 hybrid with a more efficient and
cost effective dual clutch transmission in lieu of an automatic transmission.
Additional examples of a P2 hybrid approach are the 2011 Volkswagen Touareg
Hybrid, the 2011 Porsche S Hybrid, and the 2012 Infmiti M35 Hybrid. Each of
these are examples of "first generation" P2 systems, as compared to for example
the powersplit hybrid systems offered by Ford, Toyota and or the IMA systems
s The technologies included in all of the conventional stop-start packages were expected to be widespread by
years 2017-2025. Some "anytime technologies" such as aerodynamic drag and rolling resistance reduction were
excluded from the nominal runs, but were incorporated in the complex systems portion of this project.
3-37
-------
Technologies Considered in the Agencies' Analysis
from Honda which are in their second, third or even fourth generation. The
agencies are aware of some articles in trade journals, newspapers and other
reviews that some first generation P2 hybrid vehicles with automatic transmissions
have trade-offs in NVH and drivability - though these reviews do not cover all of
the P2 systems available today, and a number of reviews are very positive with
respect to NVH and drivability. For this analysis we are projecting that these
issues with some first generation P2 systems can be addressed with no hardware
cost increase or reduction in efficiency for future generations of P2 systems
developed for the 2017-2025 time frame. The agencies sought comment on our
assumptions in this regard, and we requested comment on the applicability of
DCTs to P2 hybrid applications, including any challenges associated with NVH or
drivability. There were no comments submitted. Key technology assumptions
included:
o Lithium-ion battery
o DCT transmission
o Electric drive motor which provides, when combined with a less
powerful engine, equivalent 0-60 performance to the baseline
vehicle.
o Engine displacement for the P2 hybrids were assumed to be 20%
less than their conventional stop-start equivalents
• Input powersplit hybrid: represent a class of hybrids with both an electric drive
motor and a separate generator linked to a planetary gearset which effectively
controls the overall gear ratio and distribution of tractive and electrical power.
Example vehicles in the market include the Toyota Prius and the Ford Fusion
hybrid. Key technology assumptions are consistent with those for the P2 hybrid,
with the exception of the power split device, which functions as a CVT-type
transmission (as is the case in real world examples), and replaces the DCT
transmission in the P2 design. As stated previously while this technology was
simulated it was not used in this FRM analysis, consistent with the proposal.
Some architectures that seemed less appropriate for certain vehicle classes were
omitted. For example, in the Ricardo modeling of the medium/heavy duty truck (a Class 3
vehicle with a GVWR >10,000 pounds, and thus not subject to the final standards in this
rulemaking), no P2 or input powersplit hybrids were included. Other technologies that did
not seem reasonable for some vehicle classes (such as dry-clutch DCTs for Large MPVs and
Trucks) were also excluded in the Ricardo simulations.
In summary, 4 distinct vehicle architectures (including the baselines as an
"architecture"), across 7 vehicle classes, and a number of engine and transmission
combinations, represented the complete set of vehicle combinations. The test matrices1 can be
' For each vehicle class, each advanced engine option is combined with each advanced transmission. Baseline
runs are not combined with other transmissions.
3-38
-------
Technologies Considered in the Agencies' Analysis
found below in Table 3-5 (for 2010 baselines and conventional stop-start vehicles) and Table
3-6 (for hybrids).
Table 3-5: Nominal Package Matrix for Non-Hybrids
Vehicle Class
Small Car
Standard Car
Small MPV
Full Size Car
Large MPV
LOT
LHDT
o3
o
c* «
m a. .°.
~ to E
S o o
w o 3
ffl ry tff
X
X
X
X
X
X
X
o ._
V **
° E
— "3 .2
.S> TJ -E
Q O C
LJ o v>
0 Q. C
5 °? 2
CM U> 1—
X
Ac
o
A
3
1-
Q Q
•c O
** ^
o t:
** '^
OT g
X
X
X
X
X
X
X
Ivance
o
i_
3
1— (/)
Q O
C ^
CQ +j
X
X
X
X
X
X
X
d Engii
o
i^
H <
Q Q
' C
0 -^
UJ |
X
X
X
X
X
X
X
ie
o
<
X
Ivance
>»
o
^
o
o
Q. 1-
OT O
u> O
X
d Iran
_o
^ ts
o f=
o. o
oo <
X
X
X
X
X
X
smissic
>»
O
•o
o
o
Q. 1-
OT O
oo O
X
X
X
>n
^3
>
•o
o
o
Q. 1-
OT O
oo O
X
X
X
Table 3-6: Nominal package matrix for P2 and Input Powersplit hybrids
Vehicle Class
Small Car
Standard Car
Small MPV
Full Size Car
Large MPV
LOT
LHDT
Hybrid Architecture
4-1
g
O
T O
CM
Q. CM
X
X
X
X
X
X
Q.
W Q
< 5
X
X
X
X
X
X
3-39
-------
Technologies Considered in the Agencies' Analysis
3.3.1.2.9 Selection of the swept input variables and their ranges
The advisory committee agreed upon a set of continuous input variables to be swept in
each vehicle package response surface. These variables consisted of both powertrain
characteristics (engine displacement, final drive ratio, and electric machine size for hybrids)
and road load parameters (rolling resistance coefficient, aerodynamic drag force, and vehicle
mass). They were included in the DOE matrix for each vehicle architecture and powertrain
configuration, and also serve as inputs to the complex systems visualization tool. Table 3-7
and Table 3-8 show the swept variables used (and their ranges) for the conventional stop-start
and hybrid packages, respectively. The ranges represent a percentage of the default value
used in the nominal runs.
Table 3-7: Continuous input parameter sweep ranges for conventional stop-start vehicle
Parameter
Engine Displacement
Final Drive Ratio
Rolling Resistance
Aerodynamic Drag
Mass
DoE Range (%)
50 125
75 125
70 100
70 100
60 120
Table 3-8: Continuous input parameter sweep ranges for P2 and Powersplit hybrid vehicles
Parameter
Engine Displacement
Final Drive Ratio
Rolling Resistance
Aerodynamic Drag
Mass
Electric Machine Size
DoE Ra
P2 Hybrid
50 150
75 125
70 100
70 100
60 120
50 300
nge (%)
Powersplit
50 125
75 125
70 100
70 100
60 120
50 150
The ranges were intended to include both the (unknown) optimal value for each
technology case, but also wide enough to capture the range of values as they depart from the
optimal value (in engineering parlance this is often referred to as finding the "knee" in the
curve).
From these variables, a user can determine the sensitivity of each input variable to the
vehicle fuel economy and performance. For example, the effect of engine displacement on
fuel economy was evaluated for several packages. A more elaborate discussion of engine
displacement effects is provided in Section 3.3.1.2.24.2.
3-40
-------
Technologies Considered in the Agencies' Analysis
3.3.1.2.10 Selection of vehicle performance metrics
For both effectiveness and cost estimates in these rulemakings, the agencies are
assuming that vehicles will maintain utility (performance) comparable to the models in the
baseline fleet". It was therefore important to maintain equivalent performance in the vehicle
simulation modeling of future vehicle technology. The resulting effectiveness estimates were
in the context of equivalent performance, which carried over into the lumped parameter model
and into the OMEGA and CAFE model packages.
Consistent with the 2008 simulation project, a set of vehicle (acceleration)
performance metrics were selected by the advisory committee as a way of measuring
"equivalent" vehicle performance. When quantifying vehicle efficiency, it is important that
certain other vehicle performance metrics are maintained, such that there are no other
competing factors contributing or detracting from the vehicle efficiency. Other vehicle
characteristics that could impact or detract from vehicle efficiency (e.g., noise, vibration and
harshness (NVH), drivability, durability, etc) were also considered during the generation of
model inputs. However, they were not analyzed explicitly, with the expectation that
manufacturers would ultimately be able to meet vehicle refinement levels necessary for
commercial acceptability of these new technologies. These metrics, shown below in Table
3-9, include time at full load to reach given speeds (0-10 mph, 0-30 mph, etc), maximum
grade capability, and distance traveled at a given time (e.g., after 3 seconds). Ultimately, the
measure of equivalent performance is up to the reader or user of the Complex Systems tool.
For EPA's analysis baseline vehicle 0-30 mph and 0-60 mph acceleration times were used as
a benchmark for equivalent performance for the advanced vehicle packages. These estimated
acceleration times are included in Table 3-11 through Table 3-18. Detailed results that
include all performance metrics including those for baseline vehicles are provided in the full
2011 simulation report1.
u The only exception to this is a subset of hybrids explicitly listed as "non-towing" vehicles. For further details
and background, reference Section 1.3 of EPA's RIA.
3-41
-------
Technologies Considered in the Agencies' Analysis
Table 3-9: Vehicle performance metrics produced by the EASY5 model
Launch (WOT)
0-10mph
0-30mph
0-50mph
0-60mph
0-70mph
Distance @ 1.3 sec
Distance @ 3 sec
Speed @ 1.3 sec
Speed @ 3 sec
Passing (WOT)
30-50mph
50-70mph
Gradeability/
torque reserve
Max Speed @ 5%
grade
Max Speed @ 10%
grade
Max Grade @ 70
mph (non-towing)
Max Grade @ 60
mph (towing)
3.3.1.2.11 Review and revision of inputs
For any system modeling in which the results extend beyond the bounds of known
physical examples (and therefore direct data validation is impossible), it is imperative that the
inputs be carefully constructed and thoroughly examined to minimize the potential for
uncertainty-related errors. Prior to coding of the models, Ricardo presented the following
inputs for review and approval to EPA. For each topic, EPA reviewed the material
considering the rationale of Ricardo's technical experts, the appropriateness of the inputs in
relation to the assumed time horizon, the required emissions levels, and the known literature
in the field today. Listed below are several of the model inputs that were jointly reviewed by
Ricardo and EPA:
Engine maps
o
o
o
o
Stoichiometric GDI turbo
Lean-burn GDI turbo
Cooled EGR turbo
Advanced diesel maps
Transmission efficiency tables (by gear) including torque converter efficiency
Engine warm-up strategy (cold start modifiers)
Alternator regeneration strategy
Transmission shift optimizer
Engine friction reduction level
P2 hybrid controls
3-42
-------
Technologies Considered in the Agencies' Analysis
• Input powersplit hybrid controls
• Hybrid battery assumptions
• Hybrid motor/generator efficiency maps
EPA technical experts recommended several changes and iterated with Ricardo to
establish a consensus set of inputs that were plausible and met the ground rules of the project.
Some of these changes resulted in higher efficiencies, while others lowered efficiency.
Highlighted below are a few key examples, starting with development of the engine maps:
Engine maps carry perhaps the most significance of any of the sets of inputs needed to
build vehicle simulation models. They provide the brake specific fuel consumption, or BSFC
(typically in g/kWh) for a given engine speed and load. Typically these maps show an
optimum speed and load band (or minimum BSFC "island") that is the most efficient
condition in which to operate the engine. Ricardo generated engine maps for both the
baseline vehicles (through benchmarking data) and proposed future engine maps for the
various turbocharged and diesel engines. Figure 3-3 shows an example engine map for a
baseline vehicle. It was constructed from EPA's analysis of a baseline vehicle model run
output file. The contours represent lines of equivalent brake-specific fuel consumption/
3.3.1.2.12 Engine Technologies
Ricardo developed the engines for the 2012-2025 timeframe in two ways. The first
was to take current boosted SI research engines and project these would represent the level of
performance which could be achieved by production engines in the 2020-2025 timeframe.
The second method took current production Atkinson cycle SI and diesel engines and then
included 2020-2025 timeframe technology improvements. Both methods extrapolated current
engine design and development trend to the 2020-2025 timeframe. These current trends
include engine friction reduction, improved fuel injection systems (e.g., spray guided for the
SI, and higher injection pressures for the diesels), more advanced engine controls, and
improved engine design for faster engine warm-up. EPA reviewed the engine maps
recommended by Ricardo and generally concurred they were appropriate for the study time
frame based on EPA's review of maps for current production engines and for research engines
described in the literature.
v B SFC is measured in units of grams of fuel per kW-hour of energy and is an indicator of engine efficiency.
Lower numbers indicate more efficient operating regions. As in this case, an engine typically has an "island' or
region of best efficiency, in this case between 2000-3000 RPM and 150-180 Nm of torque. This island becomes
much larger with the advent of advanced technologies such as boosting and downsizing, as well as advanced
valvetrain technologies.
3-43
-------
Technologies Considered in the Agencies' Analysis
220
200
1BO
160 -
14O
12O -
1OO -
SO -
40
1OOO
2OOO
3OOO 4OOO
Speed (rpm)
5OOO
BOOO
Figure 3-3: Example baseline engine BSFC map
3.3.1.2.12.1 Stoichiometric GDI
The original Stoichiometric GDI map that Ricardo proposed was based on laboratory
data they had published in 2007, showing a peak brake-specific load of just under 20 bar
BMEP and a minimum BSFC of approximately 235 g/kWh, obtained using a compression
ratio of 10.5:1.26 However, based on input from manufacturers and from other, more recent
published data on developmental and research engines, EPA asked Ricardo to raise the load
_ 9*7 OQ OQ "20
capability of the engine to approximately 27 bar BMEP. ' ' ' This allowed a greater degree
of engine downsizing, which resulted in a downsizing of a 1.5 liter engine to a 0.74 liter
engine for the nominal small car and a 5.4 liter to a 1.94 liter engine for the nominal large
truck. A compression ratio of 10.5:1 was maintained for improved efficiency. At the same
time, EPA asked that Ricardo eliminate the use of high-load enrichment, since water-cooled
exhaust manifolds, in some cases integrated into the cylinder head, can be incorporated in
next-generation designs to mitigate the need for fuel enrichment in lowering turbine inlet
temperatures to 950 degrees C and thus avoid the added costs of high-temperature materials in
the turbocharger.31'32 By reducing the need for fuel enrichment fuel consumption is reduced
over the more aggressive portions of the drive cycle, and PM emissions control at high load is
improved.
3-44
-------
Technologies Considered in the Agencies' Analysis
3.3.1.2.13 Lean-burn GDI
Ricardo's initial lean-burn GDI map was based on their single-cylinder research
engine data, in which they operated in lean stratified charge mode at all speeds and loads,
without due consideration of the potential limitations in lean exhaust NOx aftertreatment
systems. To address concerns in this area, EPA examined the boundaries of operation of lean-
NOx catalysts, assuming that manufacturers would adopt either LNTs or metal-zeolite urea
SCR systems. EPA therefore asked Ricardo to place a constraint on the maximum allowable
catalyst space velocity (at high engine power) and exhaust gas temperature entering the
catalyst (at high load, low engine speed conditions) to maintain catalyst efficiency at high
load and to reduce thermal sintering of PGM under high-temperature, lean operating
conditions. More specifically, EPA recommended that engine operation switch away from
lean operation (at air/fuel equivalence ratios up to approximately X=1.5) to stoichiometric
operation at turbine outlet temperatures above 600C, and at total exhaust flows corresponding
to space velocities of 60,000/hour, assuming a catalyst volume of 2.5 times engine
displacement. This marginally diminished the engine brake thermal efficiency to
stoichiometric GDI levels over this region of the map, but it provided more certainty that the
engine would be able to adhere to the emissions levels as assumed in the project ground rules
by the Advisory Committee. Figure 3-4 shows the engine speed and load region EPA
proposed as suitable for lean stratified operation.
3-45
-------
Technologies Considered in the Agencies' Analysis
-700-
- 40000 -
Temperature
NOx Catalyst Space Velocity
Transition from lean to A=l operation
0 1000 2000 3000 4000 5000 6000 7000
Engine Speed [rpm]
Figure 3-4 Proposed lean/stoichiometric operating threshold for lean-burn GDI engines
3.3.1.2.13.1 Cooled EGR GDI
EPA provided technical information from the literature which enabled Ricardo to
assume a dual loop (both low pressure and high pressure EGR loops), cooled EGR system in
addition to the stoichiometric turbocharged engine. The development of engine maps for this
engine configuration was heavily informed by recently published data.30'31'32'33. Cooled EGR
allowed the use of "X=l" operation at the same compression ratio with more aggressive spark
timing at high load and reduced pumping losses at part load while maintaining acceptable
turbocharger inlet temperatures.
3.3.1.2.13.2 Motor/generator and power inverter efficiency maps
EPA recommended that Ricardo update the efficiency maps of the motor and
generator (referred to as "electric machines" throughout the project), which they had proposed
3-46
-------
Technologies Considered in the Agencies' Analysis
based on current best-in-class technology. The baseline motor/generator+inverter efficiency
map is taken from a 2007 Camry and shown in Figure 3-5 below.
300
250
200
a>
g.150
100
50
92
90
88
86
84
82
80
78
500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000
Speed (RPM)
Figure 3-5: 2007 Camry Hybrid motor-inverter efficiency map (Burress, et al, 200834)
EPA requested that Ricardo provide their assessment of where they believed
efficiency improvements might be made, based upon trends in research and development for
both electric machines and power electronics. Ricardo and EPA generally agreed that these
efficiency improvements were likely to be modest, particularly given the competitive
pressures on manufacturers to reduce the cost of hybrid components. However, EPA and
Ricardo assumed that today's best-in-class efficiency would likely be marginally improved
through continuous incremental reductions in parasitic losses. To account for this, EPA and
Ricardo agreed to reduce the losses in the motor/generator by 10% (in other words, raising the
efficiency of a 90% efficient motor to 91%) and to reduce the losses in the power electronics
by 25% (mainly through continued improvements in inverter development and electronic
control systems).
3.3.1.2.13.3 Battery
Battery packs were assumed to consist of spinel LiMnO2 cathode chemistry, which is
consistent with the current state of technology. EPA recommended a maximum usable state of
charge of 40% (from 30% charge to 70% charge) be incorporated as an operating window in
Ricardo's hybrid control logic. This range may increase in subsequent real world examples as
manufacturers gain more field experience with long term battery durability. Additionally
there will likely be more advances in battery construction and chemistry by 2025, so EPA
considers these assumptions as conservative in view of the long term research currently
underway in many battery research companies.
3-47
-------
Technologies Considered in the Agencies' Analysis
3.3.1.2.14 Additional Technologies Modeled by Ricardo for 2011 Report
The previous section discusses in detail those areas of the Ricardo simulation inputs
which EPA provided recommendations to Ricardo on and which Ricardo agreed and made
modifications to their initial suggestions. EPA did review modeling inputs for many other
technologies modeled by Ricardo, but for which we generally agreed with the reasonableness
of Ricardo's approach and did not request any changes. This section summarizes at a high
level some of the additional technologies considered by Ricardo. Additional detail on these
technologies is contained in the 2011 Ricardo final report.
Diesel engines - Ricardo started with existing production engines and identified
technology advances that would lead to further advances in fuel consumption. These included
many of the same technologies considered for advanced gasoline engines, such as engine
friction reduction, improved fuel injection systems with higher injection pressures and more
advanced controls, and better engine design to improve engine warm-up rate.
Transmission Technologies - Taking a systems approach in the vehicle simulation
modeling, Ricardo also introduced additional transmission and driveline oriented technologies
that may be pathways to increased efficiency. Some of these key technological enablers
include: shift optimization schedules, advanced clutches, torque converter design and lockup
schedules.
Automatic and Dual Clutch Transmissions - For the study timeframe, Ricardo
assumed that eight-speed automatic transmissions will be in common use, as this supports
more efficient operation, except for small cars, with energy losses expected to be about 20-
33% lower than in current automatic transmissions. Energy losses in both wet clutch and dry
clutch DCTs are expected to be 40-50% lower than in current automatic transmissions.
Transmission Shift Optimization - This advanced transmission shift optimization
strategy tries to keep the engine operating near its most efficient point for a given power
demand in effort to emulate a CVT. To protect against operating conditions out of normal
range, several key parameters were identified, such as maximum engine speed, minimum
lugging speed, and minimum delay between shifts. During development of this strategy,
Ricardo estimated that fuel economy benefits of up to 5% can be obtained when compared to
typical MY 2010 shift maps.
Torque Converter Technology - Ricardo utilized a lockup clutch model with a multi-
damper system to provide earlier torque converter clutch engagement. The advanced
automatic transmission applications allow torque converter lockup in any gear except first
gear, up to sixth for the Small Car or eighth for the other LDV classes.
Shifting Clutch Technology - Shift clutch technology improves the thermal capacity
of the shifting clutch to reduce plate count and lower clutch losses during shifting. Reducing
the number of plates for the shifting process and reducing the hydraulic cooling requirements
will increase the overall transmission efficiency for similar drivability characteristics.
Dry Sump Technology - A dry sump lubrication system provides benefits by keeping
the rotating members out of oil, which reduces losses due to windage and churning. This
approach will provide a GHG emissions benefit across all vehicle classes, with the best
benefits at higher speed.
3-48
-------
Technologies Considered in the Agencies' Analysis
3.3.1.2.15 Baseline models built and run
Once all of the inputs were established, Ricardo built the baseline models: For these
new (2010) baseline models Ricardo added a group of minor technologies, most of which
already exist today in the market. The technologies included 12V stop-start, 6-speed
automatic transmission, a high efficiency (70% efficient) alternator, and a strategy -
"alternator regen" - that charges the 12V battery more aggressively by increasing the
alternator field upon vehicle deceleration .
In the 2008 study Ricardo validated their baseline models with 2008 MY certification
data. Ricardo's 2010 baseline model results provided effectiveness data for EPA to calibrate
the lumped parameter model for some of the newly applied technologies. These technologies
included alternator regeneration, high efficiency alternator, and stop-start.
For all model runs - the baselines and each of the advanced package nominal runs -
EPA reviewed an extensive set of detailed intermediate output data for each model run. The
parameters that were reviewed are shown in Table 3-10.
Table 3-10: Vehicle simulation output data reviewed
Ricardo outputs
vehicle speed
throttle position
engine torque
engine power
transmission input shaft torque
wheel torque
transmission gear
torque converter slip ratio
current engine BSFC
accessory power
engine speed
road load
N/V
electric power of motor generator
mechanical power of motor generator
motor generator speed
motor generator torque
motor generator current
motor generator voltage
power flow through battery
battery state of charge
battery voltage
regenerative braking power
vehicle foundation braking power
driver braking force
fuel mass flow rate
transmission mechanical loss power
idle off status
EPA-calculated outputs
engine operating point distribution
engine load (BMEP)
total accessory energy
round-trip battery loop losses
torque converter lockup time
total road load
total engine brake thermal energy
EPA-calculated metrics
cycle-average BSFC
average brake thermal efficiency
average engine power
average engine speed
average engine torque
#of idle-off events
% of engine time off
average accessory power
time in each gear
average gear efficiency
average torque converter efficiency
battery state-of-charge statistics
battery efficiency
% of vehicle braking energy recovered
average motor efficiency
average generator efficiency
average motor and generator operating speeds
average motor and generator operating torque
total vehicle tractive energy
3-49
-------
Technologies Considered in the Agencies' Analysis
From this data, a set of summary statistics was generated to compare each baseline and
nominal package run as a quality check. This information was used as the starting point in the
dialogue between EPA and Ricardo to identify technical issues with the models. An example
summary table (or "snapshot") for the 2010 Standard Car baseline is provided in Figure 3-6.
3-50
-------
Technologies Considered in the Agencies' Analysis
Vehicle
C02 Emissions (g/mi)
Fuel Economy (mpg)
2007 Base Vehicle CO2 (g/mi)
% CO2 Reduct on
Engine
Avg Brake Thermal Efficiency
Cycle Avg BSFC (g/kWh)
Avq Engine Power (HP)
Avg Engine Speed (RPM)
Avg Load (BMEP-barJ
Avg Torque (Nm)
Total Fuel (g)
Idle Off Events
% Time Off
Accessory Loss
Avg accessory power (W)
Avg BSFC temp mult (20F)
Avg BSFC temp mult (75F)
Transmission
Time in gear 1
Time in gea 2
Time in gea 3
Time in gea 4
Time in gea 5
Time in gea 6
Time in gea 7
Time in gea 8
Avg. n (gear)
Avg. n (TC)
Avg. n (driveline)
lattery
SOCAvg
Std Deviat on
Max SOC
MinSOC
Max SOCSwng
Battery Efficiency (%)
Average Voltage (V)
Std Dev Voltage (V)
Battery Energy Change (kWh)
% ofbraking energy recovered
%batt charge via brake recov
%batt charge via eng ne
MG1
Test-Avg Motor Power (hp)
Avg Motor Eff
Avg Generator Eff
Avg Torque-Motor (N-m)
Avg Torque-Generator (N-m)
Avg RPM-Motor
Avg RPM-Generator
Mech Energy-Motor (kWh)
Mech Energy-Gen (kWh)
MG2
Avg Motor Power (hp)
Avg Motor Eff
Avg Generator Eff
Avg Torque-Motor (N-m)
Avg Torque-Generator (N-m)
Avg RPM-Motor
Avg RPM-Generator
Mech Energy-Motor (kWh)
Mech Energy-Gen (kWh)
Round-trip MG efficiency'
Buck/Boost Converter
Avg Discharge Efl
Avg Charging Eff
Avg Bus Voltage (V)
.HV (fuel)
SG(fiel)
Speciic C02
Vehicle Energy Audit (kWh)
Total fuel energy
Total indicated energy
Engine pumping energy
Engine friction energy
Engine brakinq enerqy
Total accessory energy
Net brake thermal energy
Torque converter losses
Transmission losses
Battery loop losses
PE losses
Losses to MG devices
Total driveline losses
Vehicle tractive energy
Total road load energy
Foundation braking energy
Alternator regen decel energy
Total reqd. braking energy
FTP
303.8
29.9
337.8
10.1%
FTP
21.7%
376
7.0
1993
2.21
42.1
1026.4
20
18.0%
0.0%
8.2
1.32
1.20
FTP
30%
9%
16%
27%
9%
9%
0%
0%
87.4%
88.9%
77.7%
FTP
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.0%
#DPV/0!
#DIV/0!
FTP
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.00
FTP
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.00
SDIV/0!
FTP
n/a
n/a
n/a
44
0.739
9087
FTP
12.54
4.48
0.69
0.86
0.20
0.00
2.73
0.30
0.31
0.00
0.00
0.00
0.61
2.12
1.29
0.50
0.32
0.82
Hwy
209.0
43.5
217.5
3.9%
Hwy
27.8%
295
14.1
1833
3.27
62.5
657.8
1
0.5%
0.5%
198.0
n/a
n/a
Hwy
2%
1%
2%
6%
35%
54%
0%
0%
88.0%
97.8%
86%
Hwy
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.0%
#DIV/0!
#DIV/0!
Hwy
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.00
Hwy
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.00
#DIV/0!
Hwy
n/a
n/a
n/a
kJ/g
g/gal
Hwy
8.04
3.38
0.57
0.48
0.03
0.04
2.23
0.05
0.26
0.00
0.00
0.00
0.31
1.92
1.76
0.11
0.06
0.16
Combined
261.2
34.8
283.7
7.9%
Combined
23.8%
344
10.2
1921
2.69
51.3
860.5
n/a
10.1%
0.3%
93.6
n/a
n/a
Combined
17%
5%
10%
18%
21%
29%
0%
0%
87.7%
92.9%
81.5%
Combined
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.0%
#DIV/0!
#DIV/0!
Combined
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.00
Combined
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.00
#DIV/0!
Combined
n/a
n/a
n/a
Combined
10.52
3.98
0.63
0.69
0.12
0.02
2.50
0.19
0.29
0.00
0.00
0.00
0.47
2.03
1.50
0.32
0.20
0.53
US06 Power!
312.2
29.1
US06
30.6%
267
23.0
2453
5.19
99.1
764.8
5
6.5%
0.0%
12.4
n/a
n/a
US06
13%
5%
7%
8%
10%
57%
0%
0%
87.9%
95.4%
83.8%
US06
n/a
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.0%
#DPV/0!
#DIV/0!
US06
n/a
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.00
US06
n/a
n/a
n/a
n/a
n/a
n/a
0.00
0.00
" SDIV/0!
n/a
n/a
n/a
US06
9.35
4.22
0.76
0.52
0.07
0.00
2.86
0.13
0.33
0.00
0.00
0.00
0.46
2.40
1.75
0.49
0.12
0.62
tngme tngme Irans
Disp Torque Type
L Nm
2.4 220 base auto
Perfo
[ain Archi ecti
# M
of s
gears k
re
G1 MG2 Battery
ze size size
W kW kWh
6 n/a n/a n/a
rmance Metrics
| 0-10mph | 0-30mph | 0-60mph | base 0-60 |30-50mph| 50-70mph| dist @ 3s
1.0 3.1 8.3
for using Ricardo maps
% of FC
s.s :
Shift Optimizer Evaluation
Gear Avg BMEP (bar)
FTP Hwy
1 1.7 2.3
2 3.0 3.9
3 2.4 4.5
4 1.6 3.1
5 2.7 3.7
6 2.3 2.8
7 #DIV/0! #DPV/0!
8 #DIV/0! #DIV/0!
US 06 F
4.2 1
7.1 2
6.5 2
6.7 2
6.7 2
4.0 1
#DIV/0!
#DIV/0!
Gear Avg BSFC (g/kWh)
FTP Hwy
1 338 330
2 328 282
3 359 268
4 482 298
5 361 279
6 388 31 1
700
800
MG1=sun on planetary
Recovered energy returned to whee s
Gross recovered braking energy
MG2=carrier (tractive)
From alt regen braking (extra alternator load) x
US 06 F
256 1
255 1
264 2
265 2
251 1
279 1
0 0
0 t
^
.2 5.1 20.5
Tables
Avg RPM
TP Hwy US06
121 1710 2155
309 2463 2881
188 2395 2974
60 1978 3209
D28 1869 2561
27 1737 2137
000
000
Total Energy (%)
TP Hwy US06
3% 1% 8%
5% 1% 9%
% 3% 10%
4% 7% 10%
2% 42% 16%
% 46% 49%
% 0% 0%
% 0% 0%
Figure 3-6 Sample output summary sheet for Standard Car (Camry) baseline
3-51
-------
Technologies Considered in the Agencies' Analysis
Summary statistics were used as a first-order quality check on the model. Sample
checks included:
• were average engine speed and load within or close to the best BSFC region for
the vehicle's engine map?
• was transmission gear distribution reasonable and consistent between engine
types?
3.3.1.2.16 Nominal runs
The Ricardo "nominal" runs refer to the initial set of vehicle simulation models built
for each vehicle architecture and vehicle class. These runs were used by EPA to assess the
validity of the detailed model outputs (and hence the models themselves) prior to proceeding
with the full design of experiment runs. Table 3-11 shows the summary results from the raw
nominal runs for the conventional stop-start vehicles (including 12V stop-start, 70% efficient
alternator, shift optimizer and alternator regen, as well as a 3.5% improvement due to engine
friction reduction). Conventional automatic transmissions are assumed in all nominal runs.
No road load reductions are included in these results. GHG reductions are in reference to the
2008 baseline vehicles.
Table 3-11: Nominal Conventional Stop-Start modeling results
Vehicle Engine
Class Type
Small Car
Std Ca r
La rge Ca r
Small MPV
LargeMPV
Truck
HD Truck
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
2020 Diesel
Displ.
L
0.74
0.74
0.74
1.23
1.04
1.04
1.04
1.41
1.41
1.41
2.85
1.13
1.13
1.13
1.31
1.31
1.31
2.61
1.94
1.94
1.94
4.28
2.3
2.3
2.3
6.6
Torque
Nm
157
157
157
221
220
220
220
298
298
298
503
239
239
239
277
277
277
460
410
410
410
694
486
486
486
895
Trans
Type
AT6
AT6
AT6
AT6
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
FTP
mpg
53.2
55.1
55.1
55.8
44.8
46.6
46.4
37.1
38.8
38.6
38.2
38.8
40.3
40.3
34.8
36.0
36.2
37.3
23.8
24.6
24.8
26.4
16.5
16.8
17.2
19.8
HW
mpg
55.1
56.0
57.4
59.4
54.5
55.5
56.7
43.2
44.0
44.9
46.5
42.6
43.1
44.4
39.2
39.8
40.9
43.3
26.6
27.0
27.7
30.4
18.3
18.4
19.1
21.5
Comb
mpg
54.0
55.5
56.1
57.4
48.7
50.2
50.5
39.6
41.0
41.2
41.5
40.4
41.5
42.0
36.7
37.6
38.2
39.8
25.0
25.6
26.0
28.1
17.3
17.5
18.0
20.5
0-30mph
s
4.0
4.0
4.0
3.7
3.1
3.1
3.1
3.0
3.0
3.0
2.9
3.3
3.3
3.3
3.2
3.2
3.2
3.0
3.0
3.0
3.0
2.9
3.2
3.2
3.2
2.9
0-60mph
s
10.0
10.0
10.0
9.8
8.5
8.5
8.5
7.4
7.4
7.4
7.5
8.9
8.9
8.9
8.6
8.6
8.6
8.6
8.1
8.1
8.1
8.0
9.8
9.8
9.8
8.8
%GHG
Reduction
20%
22%
23%
16%
28%
31%
31%
31%
33%
33%
27%
25%
27%
28%
31%
33%
34%
30%
26%
28%
29%
26%
27%
28%
30%
31%
3-52
-------
Technologies Considered in the Agencies' Analysis
Table 3-12 shows the results from the nominal runs for the P2 hybrid vehicles. Dual-
clutch transmissions are assumed in all nominal runs. No road load reductions are included in
these results. GHG reductions are in reference to the 2008 baseline vehicles.
Table 3-12: Nominal P2 Hybrid modeling results
Vehicle Engine
Class Type
Small Car
Std Car
La rge Ca r
Small MPV
LargeMPV
Truck
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
Displ.
L
0.59
0.59
0.59
1.66
1.66
0.83
0.83
0.83
2.4
2.4
1.13
1.13
1.13
3.8
3.8
0.9
0.9
0.9
2.6
2.6
1.05
1.05
1.05
3.15
3.15
1.55
1.55
1.55
4.6
4.6
Torque
Nm
124
124
124
138
138
176
176
176
200
200
238
238
238
317
317
190
190
190
217
217
221
221
221
263
263
327
327
327
384
384
EM size
kW
14
14
14
14
14
24
24
24
24
24
28
28
28
28
28
20
20
20
20
20
25
25
25
25
25
50
50
50
50
50
Battsize
kWh
0.70
0.70
0.70
0.70
0.70
1.00
1.00
1.00
1.00
1.00
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.15
1.15
1.15
1.15
1.15
1.50
1.50
1.50
1.50
1.50
Trans
Type
DCT6
DCT6
DCT6
DCT6
DCT6
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
FTP
mpg
68.2
68.4
70.2
70.8
71.7
61.9
62.9
65.1
64.6
65.9
49.8
50.4
51.7
49.9
51.1
50.1
50.8
52.0
52.9
54.1
47.7
47.4
47.6
48.3
48.8
32.5
33.0
33.8
33.2
33.9
HW
mpg
57.3
57.7
59.9
59.0
60.5
57.2
58.0
59.7
59.7
61.0
46.5
46.8
48.3
46.2
47.4
44.2
44.5
46.1
45.5
46.8
42.2
42.6
43.0
42.4
43.5
28.4
28.6
29.6
29.0
29.7
Comb
mpg
62.8
63.2
65.2
64.9
66.2
59.7
60.6
62.5
62.3
63.6
48.2
48.7
50.1
48.1
49.4
47.2
47.8
49.2
49.3
50.5
45.0
45.1
45.4
45.4
46.2
30.5
30.9
31.8
31.2
31.8
0-30mph
s
3.8
3.8
3.8
3.7
3.7
3.6
3.6
3.6
3.4
3.4
3.4
3.4
3.4
3.0
3.0
3.9
3.9
3.9
3.7
3.7
3.8
3.8
3.8
3.6
3.6
3.3
3.3
3.3
3.1
3.1
0-60mph
s
9.6
9.6
9.6
10.0
10.0
8.6
8.6
8.6
8.6
8.6
7.7
7.7
7.7
7.1
7.1
9.4
9.4
9.4
9.3
9.3
9.1
9.1
9.1
8.8
8.8
7.9
7.9
7.9
7.8
7.8
%GHG
Reduction
31%
31%
33%
33%
35%
42%
42%
44%
44%
45%
43%
44%
45%
43%
44%
36%
36%
38%
38%
40%
44%
44%
44%
45%
45%
39%
40%
42%
40%
42%
Table 3-13 shows the results from the nominal runs for the input powersplit vehiclesw.
No road load reductions are included in these results. GHG reductions are in reference to the
2008 baseline vehicles.
w While input powersplit hybrids remain a very likely hybrid architecture choice for some manufacturers, the
Agencies focused on P2 hybrids compared to powersplits due to their apparent cost-effectiveness advantage in
future years. As a result the powersplit nominal runs did not receive the same level of engineering scrutiny as
the P2 hybrid nominal runs.
3-53
-------
Technologies Considered in the Agencies' Analysis
Table 3-13: Nominal Powersplit hybrid modeling results
Vehicle Engine
Class Type
Small Car
Std Car
La rge Ca r
Small MPV
LargeMPV
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
Displ.
L
0.59
0.59
0.59
1.66
1.66
0.83
0.83
0.83
2.4
2.4
1.13
1.13
1.13
3.8
3.8
0.9
0.9
0.9
2.6
2.6
1.05
1.05
1.05
3.15
3.15
Torque
Nm
124
124
124
138
138
176
176
176
200
200
238
238
238
317
317
190
190
190
217
217
221
221
221
263
263
EM size
kW
14
14
14
14
14
80
80
80
80
80
28
28
28
28
28
20
20
20
20
20
25
25
25
25
25
Battsize
kWh
0.70
0.70
0.70
0.70
0.70
1.00
1.00
1.00
1.00
1.00
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.15
1.15
1.15
1.15
1.15
Trans
Type
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
PS
FTP
mpg
64.7
65.8
67.7
64.2
67.3
55.6
57.9
58.0
53.3
56.4
46.6
48.0
47.9
40.3
43.0
49.1
50.8
51.3
44.3
49.3
44.8
45.7
47.0
41.7
44.3
HW
mpg
57.2
57.4
60.1
59.5
60.0
51.7
53.5
54.8
51.7
53.3
42.0
41.8
43.6
38.7
40.8
42.2
42.7
44.9
39.6
42.3
39.3
40.6
41.5
38.6
39.6
Comb
mpg
61.1
61.7
64.0
62.0
63.8
53.8
55.8
56.5
52.6
55.0
44.4
45.0
45.9
39.6
42.0
45.8
46.8
48.2
42.1
45.9
42.1
43.3
44.4
40.3
42.0
0-30mph
s
4.8
4.8
4.8
4.7
4.7
3.7
3.7
3.7
3.6
3.6
3.2
3.2
3.2
3.2
3.2
4.7
4.7
4.7
4.6
4.6
4.3
4.3
4.3
4.2
4.2
0-60mph
s
10.4
10.4
10.4
9.8
9.8
8.7
8.7
8.7
8.0
8.0
7.8
7.8
7.8
7.1
7.1
10.3
10.3
10.3
9.1
9.1
9.7
9.7
9.7
8.8
8.8
%GHG
Reduction
29%
30%
32%
30%
32%
35%
38%
38%
34%
37%
38%
39%
40%
31%
35%
33%
35%
37%
28%
34%
40%
42%
43%
37%
40%
3.3.1.2.17 Response Surface Model matrix runs
After the nominal runs were completed according to the agreed-upon methodology,
Ricardo set up a design of experiment matrix for each vehicle architecture. The continuously
swept variables were randomized in a Latin hypercube fashion to achieve a representative
sample within each matrix (reference the Ricardo report for more details on the complex
systems modeling approach used). After a data review and removal of runs with errorsx (as
needed) Ricardo then generated Response Surface Models (RSM) for use in the complex
systems tool. EPA used the tool to evaluate a range of potential engine displacements, final
drive ratios and electric motor sizes (hybrids only) for each vehicle package, in an effort to
find the combination that would provide the greatest effectiveness while meeting EPA's
definition of "equivalent performance".
x e.g., model runs in which the vehicles were underpowered to the point where they could not follow the
prescribed vehicle speed trace, rendering an invalid test or "error". These configurations were then excluded
from the data sets.
3-54
-------
Technologies Considered in the Agencies' Analysis
3.3.1.2.18 Equivalent performance definition
The Ricardo output data provides several performance metrics, as discussed in
3.3.1.2.10. For simplicity, EPA assumed that a range of acceleration times for both a 0-60
mph test and also a 0-30 mph test (emphasizing launch character) would provide a simple yet
representative measure of a vehicle's equivalent performance. A range was chosen rather
than assuming a single point value equal to the baseline. This provided more acceptable data
points and reduced error due to "noise" in the datasets. The acceptable acceleration times
were as follows with respect to the baseline:
0-60 mph: 5 percent slower to 15 percent faster as compared to baseline
0-30 mph: 10 percent slower to 20 percent faster as compared to baseline
The range above reflects a deviation from the actual baseline value that is well within
the normal variation of acceleration times for different vehicle models within a given vehicle
class.
3.3.1.2.19 Treatment of "turbo lag" in performance runs for turbocharged engines
A common critique of comparisons of the modeled performance of highly
turbocharged engines with naturally-aspirated engines is that consideration must be given to
the delay in producing full engine load associated with the turbocharger, commonly referred
to as "turbo lag". In technical discussions, Ricardo's engine experts assured EPA that the
dual-sequential designs of the turbocharger systems in the engines in this study should
mitigate most of this phenomenon often seen on older-model vehicles. However, due to the
heavy reliance on turbocharged engines as a significant source of motive force for the high
BMEP engines evaluated in this project, EPA took this sensitivity further into account.
Ricardo's initial model of WOT operation was based on a steady-state model of
engine torque, assuming that the engine would be able to instantaneously reach a desired level
of output torque, without consideration of the intake manifold filling dynamics or the
mechanical inertia of the engine. EPA raised this as an issue, more in terms of properly
representing vehicle performance than for effectiveness differences. EPA reviewed its own
engine development data and proposed a somewhat conservative time constant for both the
naturally aspirated engines (0.3 s) and the turbocharged engines (1.5 s), to apply to the engine
torque response in the vehicle performance runs (these are shown below in Figure 3-7). In
turn, Ricardo recalculated the acceleration times for the 0-30 and 0-60 mph runs to reflect the
slower time constants. As a result, EPA used these two performance metrics exclusively in
determining "equivalent performance". A transient engine/turbo model would have improved
the accuracy of the model somewhat; however, it was beyond the scope of this project.
3-55
-------
Technologies Considered in the Agencies' Analysis
100 0%
>? en no/
gine Output Torque (
* CTl C
3 O 1
D 0 i
c^ c
1U
7O n%
0 0%
-0
//
r
^
1
/
7
5 0.5 1
^
/
^
^~~~~
_ •
— Ricarc
— EPA-h
EPA-B
5 2.5 3.5 4.5 5
Elapsed Time from WOT Command (sec)
'
Jo-assumec
at. Asp.
oosted
^=^^=^^»
I
5 6.5 7.5
Figure 3-7: EPA proposed time constants and resulting effect on torque rise time for turbocharging
3.3.1.2.20 Treatment of engine response and "turbo lag" in cycle simulations and
control logic algorithms
The EASY5 model used in the Ricardo simulations included engine and driveline
inertia effects which account for some of the real-world transient torque delays. However, the
simulation modeling did not include an adjustment to account for transient engine response
delays (e.g. inclusion of time constant offsets), to simulate naturally aspirated and
turbocharged engine response delays associated with intake manifold gas dynamics and
turbocharger response delay. Consideration of engine response delay might affect how
transmission shift optimization control logic and advanced HEV control logic is structured,
and potentially affect GHG and fuel economy projections, particularly for boosted and
downsized engines. EPA and Ricardo believe that the impact is small over the city and
highway fuel economy test cycles. The agencies sought comment on the fuel economy impact
of transient delays over the test cycles not accounted for in the Ricardo modeling, but there
were no comments received, so the agencies have made no changes in this respect for the final
rule analysis.
3.3.1.2.21 "Equivalent performance" results for conventional stop-start vehicles
The following tables show the results from the complex systems tool, when
displacement, final drive ratio and electric motor size are varied to optimize GHG and fuel
consumption reduction effectiveness at equivalent performance for conventional stop-start, P2
and powersplit hybrids. Most of the vehicles show little change in performance between the
nominal runs and the equivalent performance results from the complex systems tool. Table
3-56
-------
Technologies Considered in the Agencies' Analysis
3-14 through Table 3-18 illustrate the various effects of changing road loads on the various
vehicle package configurations. Table 3-14, Table 3-16, and Table 3-16, respectively, show
the equivalent performance results for the conventional stop-start (for both automatic
transmissions and DCTs) and the P2 hybrid vehicles (modeled only as DCTs). No road load
reductions are included in Table 3-14 through Table 3-16. For comparison, a second set of
tables (Table 3-17 and Table 3-18) give equivalent performance results for conventional stop-
start vehicles and P2 hybrids, each including example road load reductionsy of 20% mass
reduction, 20% aerodynamic drag reduction and 10% rolling resistance reduction.
The package effectiveness results from the equivalent performance runs were used in
the datasets to calibrate the individual technology effectiveness values within the lumped
parameter model. The development of the lumped parameter model is described in detail in
Section 1.5 of EPA'sRIA.
Table 3-14: Equivalent performance results for conventional-stop start vehicles (no road load reductions)
Vehicle Engine
Class Type
Small Car
StdCar
La rge Ca r
Small MPV
LargeMPV
Truck
HD Truck
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
2020 Diesel
Displ.
L
0.86
0.90
0.72
1.19
1.13
1.26
1.09
1.48
1.50
1.56
2.57
1.32
1.41
1.40
1.57
1.51
1.47
2.74
2.30
2.06
2.28
4.12
2.72
2.69
2.71
5.64
Torque
Nm
183
190
154
213
240
266
230
314
317
330
454
280
297
296
332
319
312
483
486
435
482
669
575
568
573
764
Trans
Type
AT6
AT6
AT6
AT6
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
FTP
mpg
53.1
56.3
55.2
57.3
44.4
47.0
46.2
37.0
39.2
38.6
39.1
38.9
41.1
40.0
34.8
36.2
36.4
36.7
24.0
25.0
24.8
26.8
16.6
17.2
17.3
21.0
HW
mpg
56.5
57.5
59.1
64.2
54.5
56.0
57.0
43.4
44.3
45.0
47.1
42.4
43.9
45.1
39.5
40.6
40.9
44.0
26.8
26.9
28.1
31.2
18.6
18.8
19.4
24.6
Comb
mpg
54.6
56.9
56.9
60.2
48.5
50.6
50.5
39.6
41.3
41.2
42.3
40.4
42.3
42.1
36.8
38.0
38.3
39.7
25.2
25.8
26.2
28.6
17.4
17.9
18.2
22.5
0-30mph
s
4.1
4.1
4.1
3.8
2.9
2.8
3.1
3.0
2.9
3.0
3.0
3.2
3.2
3.2
2.9
3.0
2.9
3.0
2.8
2.9
2.9
2.9
3.0
2.9
2.9
3.2
0-60mph
s
9.1
8.9
10.1
10.0
7.9
7.2
8.3
7.2
7.1
7.0
8.1
8.0
7.7
7.7
7.4
7.7
7.6
8.4
7.0
7.6
7.2
8.3
8.4
8.4
8.4
10.3
%GHG
Reduction
21%
24%
24%
20%
28%
31%
31%
31%
34%
34%
28%
25%
28%
28%
31%
34%
34%
29%
26%
28%
29%
28%
27%
29%
30%
37%
y Note that in the regulatory fleet analysis, levels of road load reduction technologies (e.g., mass reduction) will
vary by vehicle class. These tables are illustrative in nature.
3-57
-------
Technologies Considered in the Agencies' Analysis
Table 3-15: Equivalent performance results for conventional-stop start vehicles with DCT transmissions
(no road load reductions)
Vehicle Engine
Class Type
Small Car
Std Car
La rge Ca r
Small MPV
LargeMPV
Truck
HD Truck
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
2020 Diesel
Displ.
L
0.91
0.92
0.89
1.13
1.08
1.29
1.17
1.53
1.66
1.48
2.44
1.30
1.32
1.33
1.53
1.56
1.56
2.42
2.23
2.26
2.25
3.78
2.55
2.62
2.58
5.45
Torque
Nm
193
196
188
204
229
273
248
324
352
313
431
276
280
282
324
330
330
427
472
478
475
613
538
554
544
739
Trans
Type
dry DCT6
dry DCT6
dry DCT6
dry DCT6
dry DCT8
dry DCT8
dry DCT8
dry DCT8
dry DCT8
dry DCT8
dry DCT8
dry DCT8
dry DCT8
dry DCT8
wet DCT8
wet DCT8
wet DCT8
wet DCT8
wet DCT8
wet DCT8
wet DCT8
wet DCT8
wet DCT8
wet DCT8
wet DCT8
wet DCT8
FTP
mpg
55.0
58.0
57.2
61.4
46.4
48.7
48.1
38.4
40.5
40.0
41.0
40.1
42.1
41.7
36.0
38.0
37.6
39.2
24.8
25.9
25.8
28.1
17.3
17.8
18.0
21.8
HW
mpg
58.8
59.8
61.3
69.4
55.0
57.5
57.6
44.0
45.4
45.6
48.4
43.6
44.7
45.6
40.2
41.1
41.8
45.2
27.1
27.7
28.1
32.1
18.1
18.7
19.0
24.2
Comb
mpg
56.7
58.8
59.0
64.8
49.9
52.3
51.9
40.7
42.6
42.3
44.0
41.6
43.2
43.3
37.8
39.4
39.4
41.7
25.8
26.7
26.8
29.8
17.6
18.2
18.4
22.8
0-30mph
s
3.9
3.9
3.9
3.9
3.1
3.0
3.0
2.9
2.9
3.0
3.0
3.1
3.2
3.1
3.1
3.0
3.0
3.1
3.0
3.0
3.0
3.0
3.1
3.1
3.1
3.3
0-60mph
s
8.6
8.5
8.7
10.4
8.0
7.1
7.6
6.8
6.5
7.0
8.1
7.7
7.7
7.6
7.4
7.3
7.3
9.0
7.1
7.0
7.0
8.6
8.5
8.4
8.5
10.3
%GHG
Reduction
23%
26%
27%
26%
30%
33%
33%
33%
36%
35%
31%
27%
30%
30%
33%
36%
36%
33%
28%
31%
31%
31%
28%
30%
31%
38%
3-58
-------
Technologies Considered in the Agencies' Analysis
Table 3-16: Equivalent performance results for P2 hybrids (no road load reductions)
Vehicle Engine
Class Type
Small Car
Std Car
La rge Ca r
Small MPV
LargeMPV
Truck
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
Displ.
L
0.68
0.68
0.67
1.72
1.68
1.00
0.95
1.04
2.54
2.31
1.39
1.37
1.38
3.73
3.33
1.40
1.39
1.41
3.87
3.59
1.31
1.30
1.29
3.13
3.00
1.87
1.92
1.92
5.34
5.34
Torque
Nm
144
144
142
143
140
213
202
219
212
193
292
289
291
311
278
295
293
297
322
299
276
274
272
262
250
394
404
405
445
445
EM size
kW
21
21
21
17
19
26
27
26
27
28
29
29
29
30
30
34
37
38
38
39
30
31
32
34
34
50
48
48
53
56
Battsize
kWh
0.70
0.70
0.70
0.70
0.70
1.00
1.00
1.00
1.00
1.00
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.15
1.15
1.15
1.15
1.15
1.50
1.50
1.50
1.50
1.50
Trans
Type
DCT6
DCT6
DCT6
DCT6
DCT6
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
FTP
mpg
68.9
70.1
72.0
72.0
74.4
62.2
63.2
64.8
64.6
65.7
50.6
51.3
52.6
48.6
50.7
52.3
53.0
54.4
53.6
55.2
48.5
49.0
49.2
48.0
48.5
33.3
33.6
34.6
32.3
32.7
HW
mpg
58.7
59.2
61.2
60.8
62.0
57.7
58.3
60.4
59.5
60.7
47.3
47.9
49.0
46.1
47.7
45.5
45.9
47.2
46.2
47.4
42.3
42.6
42.7
42.3
43.0
29.0
29.3
30.2
28.8
29.4
Comb
mpg
63.9
64.7
66.7
66.5
68.2
60.1
60.9
62.7
62.2
63.4
49.1
49.7
50.9
47.5
49.3
49.0
49.6
50.9
50.0
51.4
45.5
45.9
46.0
45.3
45.9
31.2
31.5
32.4
30.6
31.1
0-30mph
s
3.7
3.7
3.7
3.9
3.8
3.4
3.4
3.4
3.4
3.4
3.3
3.4
3.4
3.2
3.3
3.6
3.5
3.4
3.6
3.7
3.2
3.2
3.2
3.2
3.2
3.3
3.4
3.3
3.1
3.0
0-60mph
s
8.5
8.5
8.5
9.6
9.6
7.9
8.0
7.8
8.6
8.7
7.2
7.3
7.2
7.5
8.0
8.1
8.0
7.9
9.0
9.3
7.4
7.4
7.5
8.2
8.3
7.3
7.2
7.2
7.2
7.1
%GHG
Reduction
32%
33%
35%
35%
36%
42%
43%
44%
44%
45%
44%
45%
46%
42%
44%
38%
39%
40%
39%
41%
45%
45%
45%
44%
45%
40%
41%
43%
39%
40%
3-59
-------
Technologies Considered in the Agencies' Analysis
Table 3-17 Equivalent performance results for conventional-stop start vehicles (with 20% mass, 20%
aerodynamic drag and 10% rolling resistance reductions)
Vehicle Engine
Class Type
Small Car
Std Car
La rge Ca r
Small MPV
LargeMPV
Truck
HD Truck
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
2020 Diesel
STDI
LBDI
EGRB
2020 Diesel
Displ.
L
0.68
0.89
0.69
0.91
1.04
1.27
0.98
1.00
1.49
1.00
2.05
1.20
1.40
1.13
1.00
1.26
1.02
1.98
1.44
1.89
1.44
3.20
2.21
2.24
2.19
4.45
Torque
Nm
145
189
146
164
220
268
207
212
315
212
362
253
296
238
212
266
216
349
303
399
305
518
466
473
463
603
Trans
Type
AT6
AT6
AT6
AT6
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
ATS
FTP
mpg
65.0
68.9
67.6
71.8
53.9
57.3
56.2
46.5
48.4
48.5
48.5
46.3
49.1
48.4
42.4
44.2
44.2
46.4
29.4
30.2
30.5
32.8
20.0
20.5
20.9
25.3
HW
mpg
70.0
72.4
73.1
83.2
67.6
70.6
70.1
53.8
55.0
55.9
59.7
51.8
53.5
53.6
46.8
48.1
48.7
54.0
32.1
32.9
33.6
38.8
22.2
22.6
23.1
30.1
Comb
mpg
67.2
70.4
70.0
76.5
59.3
62.6
61.7
49.5
51.2
51.6
53.0
48.6
51.0
50.6
44.3
45.9
46.2
49.6
30.6
31.3
31.8
35.3
20.9
21.4
21.8
27.3
0-30mph
s
4.1
4.2
4.1
3.7
2.9
2.8
3.0
3.1
3.0
3.1
3.0
3.2
3.3
3.2
3.2
2.9
3.2
3.0
3.1
2.8
3.1
3.0
3.0
3.0
3.0
3.2
0-60mph
s
9.2
8.4
9.2
10.4
7.2
6.4
7.6
8.1
6.5
8.1
8.1
7.4
6.9
7.7
8.8
7.3
8.7
9.0
8.6
7.0
8.6
8.6
8.4
8.4
8.4
10.3
%GHG
Reduction
35%
38%
38%
37%
41%
44%
43%
45%
46%
47%
42%
37%
40%
40%
43%
45%
45%
43%
39%
41%
42%
41%
39%
41%
42%
48%
3-60
-------
Technologies Considered in the Agencies' Analysis
Table 3-18: Equivalent performance results for P2 hybrids (with 20% mass, 20% aerodynamic drag and
10% rolling resistance reductions)
Vehicle Engine
Class Type
Small Car
Std Car
La rge Car
Small MPV
LargeMPV
Truck
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
STDI
LBDI
EGRB
ATKCS
ATKDVA
Displ.
L
0.68
0.68
0.68
1.60
1.52
0.90
0.91
0.92
2.36
2.03
1.21
1.25
1.25
3.52
3.29
1.25
1.22
1.24
3.71
3.44
1.01
1.04
1.02
2.91
2.84
1.57
1.60
1.58
4.16
4.15
Torque
Nm
143
144
143
133
127
191
194
194
197
169
254
263
263
293
274
265
257
262
309
287
213
219
215
243
237
330
337
334
347
346
EM size
kW
11
11
11
11
11
18
18
18
18
18
22
21
21
21
21
21
22
21
21
21
28
28
26
21
22
41
38
40
38
39
Battsize
kWh
0.70
0.70
0.70
0.70
0.70
1.00
1.00
1.00
1.00
1.00
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.10
1.15
1.15
1.15
1.15
1.15
1.50
1.50
1.50
1.50
1.50
Trans
Type
DCT6
DCT6
DCT6
DCT6
DCT6
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
DCT8
FTP
mpg
85.8
87.6
89.5
89.4
93.9
78.1
79.7
81.4
82.2
83.5
63.2
64.9
65.7
61.1
63.9
63.9
65.2
66.5
65.0
67.5
59.5
61.0
60.6
58.9
60.1
39.4
40.3
41.0
39.9
41.4
HW
mpg
72.2
73.1
75.4
74.9
76.9
71.1
72.2
74.2
73.8
76.2
57.3
58.5
59.8
57.0
59.3
53.4
53.9
55.7
55.1
56.7
50.2
50.9
51.6
51.1
52.4
34.4
35.0
36.0
34.9
35.9
Comb
mpg
79.1
80.4
82.5
82.2
85.4
74.8
76.2
78.0
78.2
80.0
60.4
61.9
62.9
59.2
61.7
58.7
59.5
61.1
60.1
62.1
54.9
56.0
56.2
55.1
56.3
37.0
37.7
38.6
37.5
38.7
0-30mph
s
3.7
3.7
3.7
3.8
3.8
3.2
3.3
3.2
3.1
3.3
3.1
3.1
3.1
3.0
3.0
3.5
3.5
3.5
3.4
3.6
3.2
3.2
3.2
3.2
3.2
3.2
3.3
3.2
3.0
3.0
0-60mph
s
7.9
7.9
8.0
8.9
9.0
7.2
7.2
7.1
7.5
8.3
6.6
6.5
6.6
6.7
6.8
7.7
7.7
7.8
8.2
8.7
7.4
7.3
7.3
7.5
7.7
7.0
7.0
7.0
7.1
7.2
%GHG
Reduction
45%
46%
47%
47%
49%
53%
54%
55%
55%
56%
55%
56%
56%
54%
56%
48%
49%
50%
49%
51%
54%
55%
55%
54%
55%
50%
51%
52%
50%
52%
3.3.1.2.22 Validation of vehicle simulation results
Ricardo described the process used to validate the baseline vehicles in its reportl.
Ideally it would be desirable to validate the simulation results with actual vehicle certification
test data. However, due to the nature and intended time frame (10+ years into the future) of
the technologies modeled within the vehicle classes, it is difficult to find many real-world
examples of specific technologies at the level of development reflected within the latest
simulation models. Furthermore, there are no current vehicles in production that contain all
(or even a majority) of the multiple advanced technologies embedded within the models so it
is difficult to make meaningful direct comparisons between actual vehicles and model results.
Finally, there is no direct way to disaggregate the various advanced technologies and isolate
only the relevant pieces for evaluation (e.g., an advanced turbocharged engine at an interim
BMEP level with a baseline-level transmission without stop-start): the lumped parameter
model was developed for this very analytical capability. A full description of the lumped
parameter model (including example comparisons of existing vehicle models to lumped
parameter estimates) is provided in 3.3.2.
3-61
-------
Technologies Considered in the Agencies' Analysis
3.3.1.2.23 The "efficient frontier" capability in Complex Systems tool
A powerful feature of the Complex Systems tool is the "efficient frontier" function,
which provides a graphical representation of the RSM data for the vehicle configuration of
interest. The user can identify the combination of various attributes (engine displacement,
final drive ratio, motor size, etc) that project the best model effectiveness. Figure 3-8 below is
an example of the efficient frontier for a Standard Car with a cooled EGR turbocharged
engine and a dry clutch DCT. The light red line along the top of the data set represents the
best fuel economy at each 0-60 mph acceleration time within the desired window. The solid
dark blue points represent the combinations that achieve both the desired 0-60 and 0-30 mph
criteria for equivalent performance. In this way, it is easy to quantify the best effectiveness
for a given technology package.
Efficient Frontier: Camry_ConvSS-Camry EGRB_DCT
59,9
59,8
59.7
59.6
59,5
59,4
59,3
59,2
59,1
59.0
5:3,9
53,3
53,7
58.6
: 58.5
58.4
53,3
58,3
53,1
58.0
57.9
57,8
5"
57,6
57.5
57,4
57.3
57,2
57.1
57,1
Acceptable 0-60 mph time window
8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0
0-60 mph
Figure 3-8: "Efficient Frontier" function in complex systems tool
3.3.1.2.24 Significance of the Complex Systems tool
The complex systems tool was used not only to identify the optimal combination of
input variables for each vehicle architecture, but also to analyze trends in the input variables
for quality assurance (i.e., to make sure the response surface models made engineering sense),
and to establish numerical relationships between these variables for the lumped parameter
3-62
-------
Technologies Considered in the Agencies' Analysis
model calibration. Shown below are a few examples of the types of inquiries made via the
complex systems tool:
3.3.1.2.24.1 Effects of motor size (HEVs)
EPA reviewed the effects of motor size on hybrids. As motor size is increased, there
is more opportunity to recapture energy during braking (because more powerful motors can
recover all of the energy in more severe braking events). However, oversized motors also
experience reduced efficiency as they operate in a less efficient operating region. This is
shown in Figure 3-9 below, which shows a sweep of motor size vs. fuel economy for both the
FTP/HWFE combined and also the high speed/load US06 cycle. Note that the optimum
motor size increases with respect to the US06 cycle due to more severe braking and
acceleration rates.
Monte Carlo Results
Plot 2: Monte Carlo Results
78
77
76
p75
| 74
•F 73
tl,72
"S 71
170!
69
68
67
60
49
48
47
46
45
44
43
4J
41
4n
0.5
0,6 0,7 0,8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7
EMSIZE
0.5 0.6 0,7 0.3 0.9 1,0 1.1 1.2 1.3 1.4 1.5 1.6 1.7
EMSIZE
Figure 3-9: Electric motor sweeps for Standard Car class, P2 hybrid with stoichiometric GDI engine (left
= FTP/HWFE test; right = US06 test)
3.3.1.2.24.2 Effects of engine displacement
EPA reviewed the effects of engine displacement at equivalent performance to
determine if there would be an "optimal" range of downsizing for best effectiveness.
Surprisingly, there was little benefit beyond downsizing the engine past a minimal point.
Shown in Figure 3-10 is an example complex systems tool graph with fuel economy plotted
against engine displacement multiplier (compared to the "nominal" engine displacement) for
the Truck class for three gasoline turbocharged engine packages and one diesel engine
package (note all packages included 20% weight reduction, 20% aerodynamic drag reduction,
and 10% rolling resistance reduction):
• The diesel engine result shows that the nominal engine in this case was originally
oversized because it was scaled on engine power not more accurately on engine
torque and continued displacement reduction would improve fuel economy. For
this package, the displacement for optimal fuel economy is smaller than 50% of
3-63
-------
Technologies Considered in the Agencies' Analysis
the nominal value; however, when considering equivalent vehicle performance, the
minimum diesel displacement increases to roughly 70% of the nominal value.
In contrast, the gasoline turbo engine results shown reflect a relative insensitivity
of displacement to fuel economy for these advanced vehicles.
391
33
37]
LU
^
i
LL
"
OJ
33-
32-
30-
29
28]
F-150 Moiite Carlo Results
Expected trend
0.5 0.6 0.7 0.8 0.9 1.0 1.1
DISPLACEMENT MULTIPLIER
1.2
1.3
Figure 3-10: Example displacement sweep for Truck class in complex systems tool
Figure 3-10 shows that as modeled, the swept displacement range is not large enough
for the advanced gasoline turbocharged engines. The displacement multiplier for these
engines must be greater than 1.3x the nominal displacement before the fuel economy would
degrade substantially. As the displacement drops below about 65% of the nominal (already
downsized) value, the efficiency decreases, as the engine load must be much higher to provide
the same required power. Regardless, the total fuel efficiency decrease from optimal is rather
small compared to today's engines. A 27-bar cooled EGR turbocharged GDI engine map for
3-64
-------
Technologies Considered in the Agencies' Analysis
a large car2 was reverse-engineered from the Ricardo 10 Hz output data, and is provided in
Figure 3-11. The efficiency of this family of engines is very robust to changes in engine
displacement because the highlighted BSFC region of interest (the second one out from the
minimum BSFC "island") spans a large speed and load range. As a result, significant changes
in displacement do not greatly reduce fuel efficiency. As displacement increases, the average
operating points for the engine over a given test cycle will trend towards the lower left (lower
speed, lower loadaa) portion of the map. In this case the points on the plot exist within the
same BSFC contour, so there is little degradation in engine efficiency with increasing
displacement (and drivetrain efficiency may improve at higher gears, potentially resulting in a
fuel economy increase). Were the displacement to be increased much further, the operating
region would cross the contour and fuel efficiency would begin to drop much more
dramatically.
z The 27 bar, cooled EGR turbocharged engine maps are similar for all classes as they originated from a common
reference map and scaled according to engine displacement, as described in Section 6.3 of the 2011 Ricardo
report.
aa Load decreases as it is reflective of a % of the maximum achievable torque and torque is increasing with
increased displacement. Speed decreases because of the greater torque available combined with the shift
optimizer algorithm (allowing for a greater propensity to operate in higher gears).
3-65
-------
Technologies Considered in the Agencies' Analysis
Avg. load (bmep) and
speed decrease with
increasing displacement
1000 2000 3000 4000
RPM
5000
6000
Figure 3-11: Advanced engine BSFC map (27-bar cooled EGR turbocharged GDI engine for large car)
3.3.1.2.25 Effects of mass reduction
With the complex systems tool EPA isolated the effectiveness of mass reduction on
advanced vehicle technology packages. Figure 3-12 below shows a mass reduction sweep
plot of the Large MPV class for a conventional STDI and P2 hybrid vehicle with an Atkinson
engine.
3-66
-------
Technologies Considered in the Agencies' Analysis
Large MPV - weight sweep at equivalent performance
53
5J
51
50
-49
48
£ 47
1 16
45
44
4?
42
41
40
39
38
37
36
35
34
Atkinson-P2 hybrid
-4.6% per 10% WR
StoichGDI engine
~5.2% per 10% WR
0.725 0.750 0.775 O.SOO 0.825 0.850 0.875 0.900 0.925 0.950 0.975 1.000 1.025 1.050 1,075 1.100 1.125
Weight Factor
Figure 3-12: Mass reduction sweep for Large MPV class at baseline equivalent performance. Engine
displacement and motor size (hybrids) held constant.
The mass reduction effectiveness, originally estimated at roughly 6% GHG reduction
for a 10% reduction in mass, has been revised to reflect data such as that shown above.
Isolated from benefits due to engine downsizing opportunities, the effectiveness of weight
reduction for the non-hybrid packages is on the order of 5% per 10% weight reduction, while
mass reduction for the P2 hybrid (or any hybrid) is reduced, on the order of 4.5% per 10%
reduction due to the synergies with brake energy recovery (less braking energy is recoverable
because the vehicle weighs less). The lumped parameter tool was also revised to incorporate
the synergies of weight reduction and hybrids.
3.3.1.2.26 Vehicle simulation report peer review process
As previously discussed, vehicle simulation modeling is a very detailed,
mathematically intensive approach which relies heavily on numerical engineering inputs.
These inputs (e.g., engine maps, transmission efficiency, control logic, etc.) are the heart of
the model and are derived directly from proprietary engineering knowledge of components
and subsystems. To simulate advanced engine and vehicle concepts, state-of-the-art
knowledge must be applied and converted into modeling inputs. Public domain information
is rarely at the forefront of technology, and of little use in modeling vehicles in the MYs
2017-2025 time frame.
Engineering details on advanced vehicle technologies are closely guarded in industry,
and engineering services companies which develop and generate this confidential information
rely on it to remain competitive in the marketplace. Therefore, it is difficult, if not
3-67
-------
Technologies Considered in the Agencies' Analysis
impossible, to be completely transparent with an advanced vehicle simulation model and
make all of the inputs available for public review. EPA commissioned an external peer
review of the 2011 Ricardo simulation project and report. The peer reviewers selected were
highly respected members of academia and industry, all with substantial backgrounds in
automotive technology. The list of peer reviewers and their credentials is provided in the
associated peer review report35.
EPA charged the peer reviewers to thoroughly evaluate the body of work with respect
to the following topics:
• Adequacy of the numerical inputs (engine technology selection, battery inputs,
accessory load assumptions, etc.) and highlight any caveats or limitations that
would affect the final results.
• Validity and applicability of the simulation methodology, and if it adequately
addresses synergies
• The results, and their validity and applicability to the light-duty vehicle fleet in the
2020-2025 timeframe.
• Completeness of the report (does it offer enough detail of the modeling process)
• The overall adequacy of the report for predicting the effectiveness of these
technologies, and suggest recommendations for improvement
The first round of comments was reflective of the reviewer's lack of access to model
inputs. Because the confidential inputs were initially withheld (for reasons described above),
"lack of transparency" was a consistent theme amongst the reviewers, so much that they
expressed frustration with their ability to evaluate the model methodology and the quality of
the inputs. Additionally, due to the lack of access to Ricardo proprietary input data the peer
reviewers expressed concern that they could not adequately judge the validity or accuracy of
the input information or the simulation results. EPA worked with Ricardo to provide the peer
reviewers with access to all of the detailed confidential modeling inputs under non-disclosure
agreements. With this necessary information, 3 of the 5 peer reviewers submitted a second
round of comments which were generally more specific. In turn, Ricardo modified the report
to address some of the comments, and they developed a response to comments document
which covered the comments from the peer review. One common theme called for increased
detail in how the inputs were generated. To address these requests, Ricardo provided the
detailed case studies that were used in the development of the engine maps for the cooled
EGR boosted engines and the Atkinson engines for hybrids. Ricardo also elaborated on the
hybrid control strategy, complete with state flow diagrams of operating modes, as well as a
discussion of how hybrid control strategy was optimized. Additional transmission input
details were provided, including an overview of the development of advanced gear
efficiencies and how the optimized shift strategy was applied.
The docket to this final rule contains Ricardo's response to comment document (which
includes the first version of the Ricardo report that was peer reviewed and both rounds of peer
review comments), and Ricardo's final report.36'37 The agencies sought comment on the all of
these references and on the responsiveness of the final report to the peer review comments.
3-68
-------
Technologies Considered in the Agencies' Analysis
3.3.1.3 Argonne National Laboratory Simulation Study
As discussed in the proposal, the U.S. D.O.T. Volpe Center has entered into a contract
with Argonne National Laboratory (ANL) to provide full vehicle simulation modeling support
for this MYs 2017-2025 rulemaking. While modeling was not complete in time for use in the
NPRM, the ANL results were available for the final rule and were used to define the
effectiveness of mild hybrids for both agencies, and NHTSA used the results to update the
effectiveness of advanced transmission technologies coupled with naturally-aspirated engines
for the CAFE analysis, as discussed above and more fully in NHTSA's RIA. This simulation
modeling was accomplished using ANL's full vehicle simulation tool called "Autonomie,"
which is the successor to ANL's Powertrain System Analysis Toolkit (PSAT) simulation tool,
and that includes sophisticated models for advanced vehicle technologies. The ANL
simulation modeling process and results are documented in multiple reports that can be found
in NHTSA's docket38.
3.3.2 Lumped parameter Modeling
3.3.2.1 Overview of the lumped parameter model
As a more practical alternative to full vehicle simulation, EPA developed a "lumped
parameter model" that estimates the effectiveness of various technology combinations or
"packages," in a manner that accounts for synergies between technologies. In the analysis
supporting the MYs 2012-2016 light duty vehicle GHG and CAFE rule, EPA built over 140
packages for use in its OMEGA model, which spanned 19 vehicle classes and over 1100
vehicle models. Vehicle simulation modeling performed for EPA by Ricardo, PLC, was used
to calibrate the lumped parameter model. Although DOT's analysis supporting the MYs
2012-2016 CAFE rule applied technologies incrementally, rather than specifying packages in
advance, DOT calibrated CAFE model inputs, using EPA's lumped parameter model, to
harmonize as fully as practical with estimates produced by EPA's lumped parameter model.
To support this rulemaking, EPA has updated its lumped parameter model and
calibrated it with updated vehicle simulation work performed for EPA by Ricardo, PLC. As
in the MYs 2012-2016 rulemaking, DOT has calibrated inputs, including synergy factors, to
the CAFE model to as fully as practical align with estimates produced by EPA's lumped
parameter model.
Both agencies have continued to conduct and sponsor vehicle simulation efforts to
improve inputs to the agencies' respective modeling systems. For the final rule, simulation
results for the mild hybrid technology have been incorporated into the modeling systems for
both agencies. Also, NHTSA updated the incremental effectiveness of advanced
transmissions as applied to naturally-aspirated engines, a change which was only
implemented in the CAFE model.
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
3-69
-------
Technologies Considered in the Agencies' Analysis
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,
• Vehicle road load (tire and aerodynamic) losses;
• Inertial losses (energy dissipated as heat in the brakes)
The remaining energy is available to propel the vehicle. It is assumed that the 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.
Figure 3-13 is an example spreadsheet used by EPA to estimate the package effectiveness and
the synergistic impacts of a technology package for a standard-size car.
bb For example, if only 4% of fuel energy is lost (in a baseline engine) to pumping work, leveraging multiple
technologies to theoretically eliminate all pumping losses would yield an aggregate reduction of no more than
15% in fuel consumption.
3-70
-------
Technologies Considered in the Agencies' Analysis
Vehicle Energy Effects Estimator
i :; r I "M .". '• '
: : : . . j i L : :
:.-^: t-. = r
'-:: -:::.:
IiU | Iri- ' :: :
Uni:.-.; .%;i 7;.T.;
-Sir:::.
Trj.-j
-u.^.:
T:--: I : ; 7. ;•;
-. - : - Trx =cr. 7 • jrr- r-l
Mar
L:i=T:
T: : j,.-.r
1--I3J
ITJVJX^XZJ
i;^"^.
".7, :*:•-
*•* :•:
K. X" >. OJ? 071 077
•-^3.-i| :.l:i. I i.i^c |3r.-.;T»r.|Z;..:iI T^i TiJt
T«ao Zto-»»a Qgi« r^ijari
-. : : : r - : : ; : :
- «." • .- --. : -• * '••'• .- ^ i
; r r vs : L: : : —£-.-.
Figure 3-13 Sample lumped parameter model spreadsheet
The LP model has been updated from the MYs 2012-2016 final rule to support the
MYs 2017-2025 final standards. Changes were made to include new technologies for 2017
and beyond, improve fidelity for baseline attributes and technologies, and better represent
hybrids based on more comprehensive vehicle simulation modeling. EPA RIA Chapter 1
provides details of the methodology used to update and refine the model.
3.3.2.2 Calibration of Lumped Parameter model to vehicle simulation data
The lumped parameter model includes a majority of the new technologies being
considered as part of this proposed rulemaking. The results from the Ricardo vehicle
simulation project (See section 3.3.1 for additional information) were used to successfully
calibrate the predictive accuracy and the synergy calculations that occur within the lumped
3-71
-------
Technologies Considered in the Agencies' Analysis
parameter model. When the vehicle packages Ricardo modeled are estimated in the lumped
parameter model, the results are comparable. All of the baselines for each vehicle class, as
predicted by the lumped parameter model, fall within 3% of the Ricardo-modeled baseline
results. With a few exceptions (discussed in Chapter 1 of EPA's RIA the lumped parameter
results for the 2020-2025 "nominal" technology packages are within 5% of the vehicle
simulation results. Shown below in Figure 3-14 through Figure 3-19 are Ricardo's vehicle
simulation package results (for conventional stop-start and P2 hybrid packages00) compared to
the lumped parameter estimates.
Small Car Nominal Results
Ricardo
LP results
Figure 3-14: Comparison of LP to simulation results for Small Car class
! Refer to 3.3. Ifor definitions of the baselines, "conventional stop-start" and "P2 hybrid" vehicle architectures.
3-72
-------
Technologies Considered in the Agencies' Analysis
Standard Car Nominal Results
Ricardo
LP results
Figure 3-15: Comparison of LP to simulation results for Standard Car class
Large Car Nominal Results
60
50
40
30
20
10
Illllllll
Illllllll
I Ricardo
I LP results
^
-------
Technologies Considered in the Agencies' Analysis
Small MPV Nominal Results
60
50
40
30
20
10
n • • • • • •
i • • • • • • •
i • • • • • • •
I Ricardo
I LP results
4? ^
c9»
| Conventional SS
P2 Hybrid
Figure 3-17: Comparison of LP to simulation results for Small MPV class
Large MPV Nominal Results
Ricardo
LP results
Figure 3-18: Comparison of LP to simulation results for Large MPV class
3-74
-------
Technologies Considered in the Agencies' Analysis
Truck Nominal Results
Figure 3-19: Comparison of LP to simulation results for Truck class
The recent ANL modeling results for mild hybrids largely confirmed the effectiveness
as originally predicted by the lumped parameter model, with minor differences for small cars
and large trucks. A comparison of the ANL results to the original lumped parameter results
(for comparable vehicle classes when modeled with a nominal 15 kW motor size) is shown
below in Table 3-19 and Table 3-20.
Table 3-19 ANL Effectiveness for Mild Hybrid
FC reduction
Compact
11.6%
Midsize
11.6%
Small SUV
10.2%
Midsize SUV
10.5%
Pickup
8.5%
Table 3-20 Lumped Parameter Model Effectiveness for Mild Hybrid
FC reduction
Small Car
14.1%
Std Car
11.8%
Small MPV
10.1%
Large MPV
10.1%
Truck
6.9%
The underlying structure of the lumped parameter model was not changed to
accommodate this new information; instead, the nominal 15 kW motor sizes for small cars
3-75
-------
Technologies Considered in the Agencies' Analysis
and pickup truck mild hybrids were adjusted (to 10 kW and 18 kW, respectively) to reflect the
updated effectiveness results provided by the ANL simulation work.
3.3.2.3 Comparison of results to real-world examples
To validate the lumped parameter model, representations of actual late-model
production vehicles exhibiting advanced technologies were created. Shown in Table 3-21 are
a set of select vehicle models containing a diverse array of technologies: included are the
pertinent technologies and vehicle specifications, along with actual vehicle certification fuel
economy test data compared to the lumped parameter fuel economy estimates. For the
vehicles and technologies shown, the predicted fuel economy is within about 3% of the actual
data.
Table 3-21: Production vehicle certification data compared to lumped parameter predictions
Vehicle
Vehicle Class
Engine
Transmission
HEV motor (kW)
ETW (Ibs)
City/HW FE (mpg)
LP estimate (mpg)
Key technologies
applied in LP model
20 11 Chevy Craze
ECO
Small Car
1.4L 14 Turbo GDI
6 speed auto
n/a
3375
40.3
40.2
GDI (stoich)
Turbo (30%
downsize)
Ultra low R tires
Active grill shutters
2011 Sonata Hybrid
Standard Car
2.4L 14 Atkinson
6 speed auto
30
3750
52.2
51.7
P2 hybrid
Aero improvements
20 11 Escape Hybrid
Small MPV
2.5L 14 Atkinson
CVT
67
4000
43.9
44.0
Powersplit hybrid
2011F-150
EcoBoost
Track
3.5LV6TurboGDI
6 speed auto
n/a
6000
22.6
21.9
GDI (stoich)
Turbo (37%
downsize)
3.4 What cost and effectiveness estimates have the agencies used for each technology?
As discussed in the previous sections, many the effectiveness estimates for this final
rule, consistent with the proposal, including the estimates for the technologies carried over
from the MYs 2012-2016 final rule, are derived from the 2011 Ricardo study and
corresponding updated version of the lumped-parameter model. It is important to note that
the agencies used the average of the range presented when referencing the effectiveness
3-76
-------
Technologies Considered in the Agencies' Analysis
estimates from the MYs 2012-2016 final rule. If, for example, the effectiveness range for
technology X was determined to be 1 to 2 percent, the agencies used a value of 1.5 percent in
their respective analyses. However, the effectiveness ranges that are presented for the MYs
2017-2025 analysis, as informed by the Ricardo 2011 study, define the range of estimates
used by the agencies for the different vehicle types. Again using technology X as an example,
if the range is now defined as 2.0 to 2.5 percent then for small passenger cars (subcompact or
compact) the estimated effectiveness might be 2.0 percent but for large cars an estimate of 2.5
percent might be used.
As noted in section 3.1.3, the effects of learning on individual technology costs can be
seen in the cost tables presented throughout this section 3.3. For each technology, we show
direct manufacturing costs for the years 2017 through 2025. The changes shown in the direct
manufacturing costs from year-to-year reflect the cost changes due to learning effects.
3.4.1 Engine technologies
As indicated in the cost tables that found in this section, the agencies updated the
costing approach for some technologies in an effort to provide better granularity in our
estimates. This is reflected in Table 3-23, among others, listing costs for technologies by
engine configuration—in-line or "I" versus "V"—and/or by number of cylinders. In the MYs
2012-2016 final rule, we showed costs for identified vehicle classes such as small car, large
car, large truck, etc. The identified challenges inherent with that approach are that different
vehicle classes can have many different sized engines. This condition may become more
prominent going forward as more turbocharged and downsized engines enter the fleet. For
example, the agencies project that many vehicles in the large car class, have large
displacement V8 or V6 engines would move to highly turbocharged 14 engines under the final
rule, consistent with the proposal. As such, we would not want to estimate the costs of engine
friction reduction for large cars—which have always and continue to be based on the number
of cylinders—by assuming that all large cars have V8 or V6 engines.
3.4.1.1 Low Friction Lubricants
A basic method of reducing fuel consumption in gasoline engines is using of lower
viscosity engine lubricants. Advanced multi-viscosity engine oils are available today which
yield improved performance in a wider temperature band and with better lubricating
properties. These advances are 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 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, connecting rod and main crankshaft bearing designs and/or materials along with
the mechanical tolerances of engine components may be required. In all cases, durability
testing would be required to ensure that durability is not compromised. Shifting to lower
viscosity and lower friction lubricants can also improve the management of valvetrain
technologies such as cylinder deactivation or variable valve timing, which rely on a minimum
oil temperature (viscosity) for operation.
3-77
-------
Technologies Considered in the Agencies' Analysis
Several manufacturers have previously commented confidentially that low friction
lubricants could have an effectiveness value between 0 to 1 percent. The agencies used the
average effectiveness of 0.5 in the MYs 2012-2016 final rule. For purposes of this final rule,
consistent with the proposal, the agencies relied on the lumped parameter model and the range
for the effectiveness of low friction lubricant is 0.5 to 0.8 percent.
In the MYs 2012-2016 final rule, the 2010 TAR and the MYs 2014-2018 Medium and
Heavy Duty GHG and Fuel Efficiency final rule, EPA and NHTSA used a direct
manufacturing cost (DMC) of $3 (2007$) and considered that cost to be independent of
vehicle class since the engineering work required should apply to any engine size. The
agencies continue to believe that this cost is appropriate and, having adjusted for 2010$, the
cost remains the same for this analysis. No learning is applied to this technology so the DMC
remains $3 (2010$) year-over-year. The agencies have used a low complexity ICM of 1.24
for this technology through 2018 and 1.19 thereafter. The resultant costs are shown in Table
Note that low friction lubes are expected to exceed 85 percent penetration by the 2017
3-22.
MY.
dd
Table 3-22 Costs for Engine Modifications to Accommodate Low Friction Lubes (2010$)
Cost
type
DMC
1C
TC
Engine
type
All
All
All
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
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are
incremental to the baseline.
3.4.1.2 Engine Friction Reduction
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.39 Example improvements include 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 where minute improvements in several components can result in a
dd
Note that the costs developed for low friction lubes for this analysis reflect the costs associated with any
engine changes that would be required as well as any durability testing that may be required.
3-78
-------
Technologies Considered in the Agencies' Analysis
measurable fuel economy improvement. In the MYs 2012-2016 final rule, the agencies relied
on the 2002 NAS, NESCCAF and EEA reports, as well as, confidential manufacturer data that
suggested a range of effectiveness for engine friction reduction (EFR1) to be between 1 to 3
percent. Because of the incremental technology application capability of the CAFE model,
NHTSA used the narrower range of 1 to 2 percent, which resulted in an average effectiveness
of 1.5 percent. Based on the 2011 Ricardo study results, the agencies have revised the
effectiveness for engine friction reduction range to 2.0 to 2.7 percent for this analysis.
For this final rule, consistent with the proposal, the agencies added a second level of
incremental improvements in engine friction reduction (EFR2) applicable over multiple
vehicle redesign cycles. This second level of engine friction reduction forecasts additional
improvements to low friction lubricants relative to the low friction lubricant technology
discussed above and is considered to be mature only after MY 2017. The effectiveness for this
second level, relative to the base engine, is 3.4 to 4.8 percent based on the lumped parameter
model. Because of the incremental technology application capability of the CAFE model,
NHTSA used the effectiveness range of 0.83 to 1.37 percent incremental to the first level of
engine friction reduction and low friction lubricants for a total effectiveness of 2.83 to 4.07
percent.
In the MYs 2012-2016 rule, the 2010 TAR and the MYs 2014-2018 Medium and
Heavy Duty GHG and Fuel Efficiency final rule, NHTSA and EPA used a EFR1 cost estimate
of $11 (2007$) per cylinder DMC, or $12 (2010$) per cylinder in this analysis. No learning is
applied to this technology so the DMC remains $12 (2010$) year-over-year. The agencies
have used a low complexity ICM of 1.24 for this technology through 2018 and 1.19
thereafter. The resultant costs are shown in Table 3-23. Note that EFR1 is expected to
exceed 85 percent penetration by MY 2017.
Table 3-23 Costs for Engine Friction Reduction - Level 1-EFR1 (2010$)
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
Engine
type
13
14
V6
V8
13
14
V6
V8
13
14
V6
V8
2017
$36
$48
$71
$95
$9
$11
$17
$23
$44
$59
$89
$118
2018
$36
$48
$71
$95
$9
$11
$17
$23
$44
$59
$89
$118
2019
$36
$48
$71
$95
$7
$9
$14
$18
$43
$57
$85
$113
2020
$36
$48
$71
$95
$7
$9
$14
$18
$43
$57
$85
$113
2021
$36
$48
$71
$95
$7
$9
$14
$18
$43
$57
$85
$113
2022
$36
$48
$71
$95
$7
$9
$14
$18
$43
$57
$85
$113
2023
$36
$48
$71
$95
$7
$9
$14
$18
$43
$57
$85
$113
2024
$36
$48
$71
$95
$7
$9
$14
$18
$43
$57
$85
$113
2025
$36
$48
$71
$95
$7
$9
$14
$18
$43
$57
$85
$113
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are
incremental to the baseline.
The agencies have estimated the DMC of the second level of friction reduction and
low friction lubricants at double the combined DMCs of EFR1 (double the DMC relative to
the baseline). As a result, the costs of EFR2 are as shown in Table 3-24. For EFR2 the
agencies have used a low complexity ICM of 1.24 through 2024 and 1.19 thereafter.
3-79
-------
Technologies Considered in the Agencies' Analysis
Table 3-24 Costs for Engine Friction Reduction - Level 2 - EFR2 (2010$)
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
Engine
type
13
14
V6
V8
13
14
V6
V8
13
14
V6
V8
2017
$78
$102
$149
$197
$19
$25
$36
$48
$97
$126
$185
$244
2018
$78
$102
$149
$197
$19
$25
$36
$48
$97
$126
$185
$244
2019
$78
$102
$149
$197
$19
$25
$36
$48
$97
$126
$185
$244
2020
$78
$102
$149
$197
$19
$25
$36
$48
$97
$126
$185
$244
2021
$78
$102
$149
$197
$19
$25
$36
$48
$97
$126
$185
$244
2022
$78
$102
$149
$197
$19
$25
$36
$48
$97
$126
$185
$244
2023
$78
$102
$149
$197
$19
$25
$36
$48
$97
$126
$185
$244
2024
$78
$102
$149
$197
$19
$25
$36
$48
$97
$126
$185
$244
2025
$78
$102
$149
$197
$15
$20
$29
$38
$93
$121
$178
$234
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are
incremental to the baseline.
3.4.1.3 Cylinder Deactivation
In conventional spark-ignition engines, throttling the intake 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 (usually) half of the cylinders when the load is less than half of the
engine's total torque capability. Cylinder deactivation is achieved by keeping specific
cylinder valves closed and stopping fuel flow to the specified cylinder. 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 almost double
the load required if all of the cylinders were operating. Overall engine 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 ranges where it is acceptable to deactivate engine cylinders.
Noise and vibration issues reduce the operating range where 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 may choose to
adopt active engine mounts and/or active noise cancellation systems to address NVH concerns
and allow a greater operating range of activation which is also shown in the cost estimates for
this technology. Most manufacturers have legitimately stated that use of DEAC on 4 cylinder
engines would cause unacceptable NVH; therefore, as in the MYs 2012-2016 rule and the
2010 TAR, the agencies are not applying cylinder deactivation to 4-cylinder engines in
evaluating potential emission reductions/fuel economy improvements and associated costs.
Cylinder deactivation has seen a recent resurgence thanks to better valvetrain designs
and engine controls. General Motors and Chrysler Group have incorporated cylinder
deactivation across a substantial portion of their V8-powered vehicles and Honda offers V6
models with cylinder deactivation.
3-80
-------
Technologies Considered in the Agencies' Analysis
Effectiveness improvements scale roughly with engine displacement-to-vehicle weight
ratio: the higher displacement-to-weight vehicles, operating at lower relative loads for normal
driving, have the potential to operate in part-cylinder mode more frequently.
NHTSA and EPA reviewed estimates from the MYs 2012-2016 final rule, 2010 TAR,
the RIA for the MYs 2014-2018 Medium and Heavy Duty GHG and Fuel Efficiency final
rule. The lumped parameter model applied a 6 percent reduction in CC>2 emissions depending
on vehicle class. The CAFE model, due to its incremental technology application capability,
used a range depending on the engine valvetrain configuration. For example, DOHC engines
already equipped with DCP and DVVLD achieve little benefit, 0.5 percent for DEACD, from
adding cylinder deactivation since the pumping work has already been minimized and internal
Exhaust Gas Recirculation (EGR) rates are maximized. However, SOHC engines, which
have CCP and DWLS applied, achieve effectiveness ranging from 2.5 to 3 percent for
DEACS. And finally, OHV engines, without VVT or VVL technologies, achieved
effectiveness for DEACO ranging from 3.9 to 5.5 percent.
For this final rule, consistent with the proposal, the agencies, taking into account the
additional review and the work performed for the 2011 Ricardo study, have revised the
effectiveness estimates for cylinder deactivation. The effectiveness relative to the base engine
is 4.7 to 6.5 percent based on the lumped parameter model. Because of the incremental
technology application capability of the CAFE model, NHTSA used the effectiveness range
of 0.44 to 0.66 percent incremental for SOHC and DOHC applications. For OHV applications
having no incremental application of VVT or VVL, the effectiveness was increased to a range
of 4.66 to 6.30 percent.
In the MYs 2012-2016 final rule and the 2010 TAR, the agencies used a DMC
estimate of $140 (2007$) and $157 (2007$) for cylinder deactivation technology on V6 and
V8 engines, respectively. Adjusted for 2010$, the DMCs become $146 (2010$) and $165
(2010$) for this analysis and are considered applicable in MY 2015. This technology is
considered to be on the flat-portion of the learning curve. The agencies have applied a low
complexity ICM of 1.24 to this technology through 2018 and 1.19 thereafter. The resultant
costs are shown in Table 3-25.
Table 3-25 Costs for Cylinder Deactivation (2010$)
Cost
type
DMC
DMC
1C
1C
TC
TC
Engine
type
V6
V8
V6
V8
V6
V8
2017
$139
$157
$56
$63
$196
$220
2018
$136
$153
$56
$63
$193
$217
2019
$134
$150
$42
$47
$176
$198
2020
$131
$147
$42
$47
$173
$195
2021
$128
$144
$42
$47
$170
$191
2022
$126
$142
$42
$47
$168
$189
2023
$123
$139
$42
$47
$165
$186
2024
$121
$136
$42
$47
$162
$183
2025
$118
$133
$42
$47
$160
$180
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are
incremental to the baseline.
There is potential that, on engines already equipped with the mechanisms required for
cylinder deactivation capability, the cost of DEAC as applied to SOHC and DOHC engines
could be as low as $32 in MY 2017. This $32 accounts for the potential additional
3-81
-------
Technologies Considered in the Agencies' Analysis
application of active engine mounts on SOHC and DOHC engines that, while having the
potential to apply cylinder deactivation, may or would require these additional NVH
improving devices for consumer acceptance. For this analysis, this additional expanded
application and expense is only applied on 50 percent of the vehicles. Further, this SOHC and
DOHC engine estimate is relevant to the CAFE model only because the OMEGA model does
not apply technologies in the same incremental fashion as the CAFE model.
3.4.1.4 Variable Valve Timing (WT)
Variable valve timing (VVT) encompasses 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 MY 2011, approximately 93.8
percent of all new cars and light trucks had engines with some method of variable valve
timing.40 Manufacturers are currently using many different types of variable valve timing,
which have a variety of different names and methods. Manufacturers are currently using
many different types of variable valve timing, which have a variety of different names and
methods. Therefore, the degree of further improvement across the fleet is limited by the level
of valvetrain technology already implemented on the vehicles. Information found in the 2008
and 2010 baseline vehicle fleet files is used to determine the degree to which VVT
technologies have already been applied to particular vehicles to ensure the proper level of
VVT technology, if any, is applied. The three major types of VVT are listed below.
Each of the three implementations of VVT 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.
3.4.1.4.1 Intake Cam Phasing (ICP)
Valvetrains with Intake Cam Phasing (ICP), which is the simplest of the cam phasing
technologies, 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.
In the MYs 2012-2016 final rule and 2010 TAR, NHTSA and EPA assumed an
effectiveness range of 2 to 3 percent for ICP. Based on the additional information from the
3-82
-------
Technologies Considered in the Agencies' Analysis
2011 Ricardo study and updated lumped parameter model the agencies have been able to fine-
tuned the effectiveness range to be 2.1 to 2.7 percent for this analysis.
In the MYs 2012-2016 rule and the 2010 TAR, the agencies estimated the DMC of a
single cam phaser for ICP at $37 (2007$). This DMC, adjusted for 2010$, becomes $39
(2010$) for this analysis and is considered applicable in the 2015 MY. This cost would be
required for each cam shaft controlling intake valves. As such an OHC 14 and OHV V6 or V8
would need one cam phaser while an OHC V6 or V8 would need two cam phasers. This
technology is considered to be on the flat-portion of the learning curve. The agencies have
applied a low complexity ICM of 1.24 to this technology through 2018 and 1.19 thereafter.
The resultant costs are shown in Table 3-26.
Table 3-26 Costs for WT-Intake Cam Phasing - ICP (2010$)
Cost
type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
Engine type
OHC-I4
OHC-V6/V8
OHV-V6/V8
OHC-I4
OHC-V6/V8
OHV-V6/V8
OHC-I4
OHC-V6/V8
OHV-V6/V8
2017
$37
$74
$37
$9
$19
$9
$46
$93
$46
2018
$36
$72
$36
$9
$19
$9
$46
$91
$46
2019
$35
$71
$35
$7
$15
$7
$43
$86
$43
2020
$35
$70
$35
$7
$15
$7
$42
$84
$42
2021
$34
$68
$34
$7
$15
$7
$42
$83
$42
2022
$33
$67
$33
$7
$15
$7
$41
$82
$41
2023
$33
$65
$33
$7
$15
$7
$40
$80
$40
2024
$32
$64
$32
$7
$15
$7
$40
$79
$40
2025
$31
$63
$31
$7
$15
$7
$39
$78
$39
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; OHC=overhead cam;
OHV=overhead valve; all costs are incremental to the baseline.
3.4.1.4.2 Coupled 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 an overhead valve (OHV) engine. For SOHC 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 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.ee
The agencies' MYs 2012-2016 final rule estimated the effectiveness of CCP to be
between 1 to 4 percent. Due to the incremental technology application capability of the
ee It is also noted that coaxial camshaft developments would allow other VVT options to be applied to OHV
engines. However, since they would potentially be adopted on a limited number of OHV engines NHTSA did
not include them in the decision tree.
3-83
-------
Technologies Considered in the Agencies' Analysis
CAFE model, NHTSA estimated the effectiveness for CCP to be 1 to 3 percent for a SOHC
engine and 1 to 1.5 percent for an overhead valve engine.
For this final rule, consistent with the proposal, the agencies, have revised the
estimates for CCP taking into account the additional review and the work performed for the
2011 Ricardo study. The effectiveness relative to the base engine is 4.1 to 5.5 percent based
on the lumped parameter model. Because of the incremental nature of the CAFE model,
NHTSA used the incremental effectiveness range of 4.14 to 5.36 percent for SOHC
applications; an increase over the MYs 2012-16 final rule and 2010 TAR. For OHV
applications, CCP was paired with discrete variable valve lift (DVVL) to form a new
technology descriptor called variable valve actuation (VVA). Effectiveness values for this
new descriptor is discussed later in Section 3.4.1.6.
In regard to CCP costs, the same cam phaser has been assumed for intake cam phasing
as for coupled cam phasing, thus the DMCs for CCP is identical to those presented for ICP in
Table 3-26.
3.4.1.4.3 Dual Cam Phasing (DCP)
The most flexible VVT design is dual (independent) cam phasing (DCP), where the
intake and exhaust valve opening and closing events are controlled independently. This
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 CC>2 emissions. Increased internal EGR also results in lower engine-out
NOx emissions. Fuel consumption and CC>2 emissions improvements enabled by DCP are
dependent on the residual tolerance of the combustion system. Additional improvements are
observed at idle, where low valve overlap could result in improved combustion stability,
potentially reducing idle fuel consumption. For forward looking technology application, DCP
is only applicable to dual overhead cam (DOHC) engines.ff
For the MYs 2012-2016 final rule and 2010 TAR, the EPA and NHTSA assumed an
effectiveness range for DCP of 3 to 5 percent relative to a base engine or 2 to 3 relative to an
engine with ICP. The agencies have updated this range, based on the updated lumped-
parameter model, to be 4.1 to 5.5 percent relative to a base engine or 2.0 to 2.7 percent
relative to an engine with ICP.
ff The agencies note at least one production implementation of an OHV dual cam phasing is included in the
baseline fleet. This consisted of a single concentric camshaft (a "camshaft within a camshaft") and a single dual
vane phaser assemblies enabling independent phasing of the intake and exhaust camshaft profiles. However, this
technology was applied to a limited production sports car versus a mass market application with significant sales
volume. The agencies are not aware of any similar application moving forward.
3-84
-------
Technologies Considered in the Agencies' Analysis
The costs for DCP are the same per phaser as described above for ICP. However, for
DCP, an additional cam phaser is required for each camshaft controlling exhaust valves. As a
result, a dual overhead cam 14 would need two phasers and a dual overhead cam V6 or V8
would need four phasers, and an overhead valve V engine would need two.gg
This technology is considered to be on the flat-portion of the learning curve. The
agencies have applied a medium complexity ICM of 1.39 to this technology through 2018 and
1.29 thereafter. The resultant costs are shown in Table 3-27.
Table 3-27 Costs for WT-Dual Cam Phasing (2010$)
Cost
type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
Engine type
OHC-I4
OHC-V6/V8
OHV-V6/V8
OHC-I4
OHC-V6/V8
OHV-V6/V8
OHC-I4
OHC-V6/V8
OHV-V6/V8
2017
$68
$146
$74
$27
$59
$30
$95
$205
$104
2018
$66
$143
$72
$27
$59
$30
$94
$202
$102
2019
$65
$140
$71
$20
$44
$22
$86
$184
$93
2020
$64
$137
$70
$20
$44
$22
$84
$181
$92
2021
$62
$134
$68
$20
$44
$22
$83
$178
$90
2022
$61
$132
$67
$20
$44
$22
$82
$176
$89
2023
$60
$129
$65
$20
$44
$22
$80
$173
$88
2024
$59
$127
$64
$20
$44
$22
$79
$170
$86
2025
$58
$124
$63
$20
$44
$22
$78
$168
$85
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; OHC=overhead cam;
OHV=overhead valve; all costs are incremental to the baseline.
3.4.1.5 Variable Valve Lift (VVL)
Varying and controlling the amount of cylinder valve lift across and engine operating
range 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 processes. 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 also
incur increased parasitic losses associated with their actuation mechanisms. A number of
manufacturers have already implemented VVL into their fleets (Toyota, Honda, and BMW),
but overall this technology is still available as an efficiency improving technology for most of
the fleet. There are two major classifications of variable valve lift, described below:
-Ibid.
3-85
-------
Technologies Considered in the Agencies' Analysis
3.4.1.5.1 Discrete Variable Valve Lift (DVVL)
Discrete variable valve lift (DVVL) systems allow the selection between two or three
discrete cam profiles by means of a hydraulically-actuated mechanical system. These cam
profiles consist of a low and a high-lift lobe, and may 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. DWL is also known as Cam Profile Switching (CPS).
DVVL is a mature technology with low technical risk.
The effectiveness of DVVL has been estimated to range from 1 to 4 percent in addition to that
realized by VVT systems. These values were based on the research supporting MYs 2012-16
final rule, confidential manufacturer data, and a research conducted by the Northeast States
Center for a Clean Air Future (NESCCAF). Based on additional information contained in the
2011 Ricardo study, NHTSA and EPA have revised the effectiveness range of DVVL systems
to 2.8 to 3.9 percent above that realized by VVT systems.
In the MYs 2012-2016 rule and the 2010 TAR, the agencies estimated the DMC of
DVVL at $116 (2007$), $169 (2007$) and $241 (2007$) for an 14, V6 and V8 engine,
respectively. Adjusted for 2010$, these DMCs become $122 (2010$), $177 (2010$) and $253
(2010$) for this analysis all of which are considered applicable in MY 2015. This technology
is considered to be on the flat-portion of the learning curve and is applicable only to engines
with overhead cam configurations. The agencies have applied a medium complexity ICM of
1.39 to this technology through 2018 and 1.29 thereafter. The resultant costs are shown in
Table 3-28.
Table 3-28 Costs for Discrete Variable Valve Lift - DWL (2010$)
Cost
type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
Engine type
OHC-I4
OHC-V6
OHC-V8
OHC-I4
OHC-V6
OHC-V8
OHC-I4
OHC-V6
OHC-V8
2017
$116
$168
$240
$47
$68
$97
$163
$236
$338
2018
$114
$165
$235
$47
$68
$97
$161
$233
$333
2019
$111
$161
$231
$35
$51
$73
$146
$212
$303
2020
$109
$158
$226
$35
$51
$72
$144
$209
$298
2021
$107
$155
$222
$35
$51
$72
$142
$206
$294
2022
$105
$152
$217
$35
$50
$72
$140
$202
$289
2023
$103
$149
$213
$35
$50
$72
$137
$199
$285
2024
$101
$146
$209
$35
$50
$72
$135
$196
$280
2025
$99
$143
$204
$35
$50
$72
$133
$193
$276
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; OHC=overhead cam;
OHV=overhead valve; all costs are incremental to the baseline.
3.4.1.5.2 Continuously Variable Valve Lift (CVVL)
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 sold port-injected
"Valvetronic" engines since 2001. Fiat is now offering "MultiAir" engines enabling precise
control over intake valve lift. CVVL allows the airflow into the engine to be regulated by
3-86
-------
Technologies Considered in the Agencies' Analysis
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.
Variable valve lift gives a further reduction in pumping losses compared to that which
can be obtained with cam phase control only, with CVVL providing 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.
The MYs 2012-2016 final rule estimated the effectiveness for CVVL at 1.5 to 3.5
percent over an engine with DCP, but also recognized that it could go up as high as 5 percent
above and beyond DCP to account for the implementation of more complex CVVL systems
such as BMW's "Valvetronic" and Fiat "MultiAir" systems. Thus, the effectiveness range for
CVVL in this Joint TSD ranges from 1.5 to 7 percent depending on the complexity level of
the application
. For this rulemaking, NHTSA has increased the incremental effectiveness values for
this technology to a range of 3.6 to 4.9 percent from 1.5 to 3.5 percent in the MYs 2012-2016
final rule.
In the 2012-2016 rule and the 2010 TAR, the agencies estimated the DMC of CVVL
at $174 (2007$), $320 (2007$), $349 (2007$), $866 (2007$) and $947 (2007$) for an OHC-
14, OHC-V6, OHC-V8, OHV-V6 and OHV-V8 engine, respectively. Adjusted for 2010$,
these DMCs become $183 (2010$), $335 (2010$), $366 (2010$), $893 (2010$) and $977
(2010$) for this analysis all of which are considered applicable in MY 2015. As indicated in
this section, CVVL is considered only applicable to DOHC engine designs. The DMCs for
OHV engines are meant to reflect additional costs associated with moving to a DOHC engine
design.
This technology is considered to be on the flat-portion of the learning curve. The
agencies have applied a medium complexity ICM of 1.39 to this technology through 2018 and
1.29 thereafter. The resultant costs are shown in Table 3-29.
Table 3-29 Costs for Continuous Variable Valve Lift - CWL (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
Engine type
OHC-I4
OHC-V6
OHC-V8
OHV-V6
OHV-V8
OHC-I4
OHC-V6
OHC-V8
2017
$174
$319
$348
$857
$937
$70
$129
$141
2018
$170
$313
$341
$840
$919
$70
$129
$141
2019
$167
$306
$334
$823
$901
$53
$96
$105
2020
$164
$300
$327
$807
$883
$52
$96
$105
2021
$160
$294
$321
$791
$865
$52
$96
$105
2022
$157
$288
$314
$775
$847
$52
$96
$104
2023
$154
$283
$308
$760
$830
$52
$96
$104
2024
$151
$277
$302
$744
$814
$52
$95
$104
2025
$148
$271
$296
$729
$798
$52
$95
$104
3-87
-------
Technologies Considered in the Agencies' Analysis
1C
1C
TC
TC
TC
TC
TC
OHV-V6
OHV-V8
OHC-I4
OHC-V6
OHC-V8
OHV-V6
OHV-V8
$347
$380
$244
$448
$489
$1,205
$1,317
$346
$379
$241
$441
$482
$1,187
$1,298
$259
$283
$220
$403
$439
$1,083
$1,184
$259
$283
$216
$396
$432
$1,066
$1,166
$258
$282
$213
$390
$426
$1,048
$1,147
$258
$282
$209
$384
$419
$1,032
$1,129
$257
$281
$206
$378
$412
$1,016
$1,112
$257
$281
$203
$372
$406
$1,001
$1,095
$256
$280
$200
$367
$400
$986
$1,078
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; OHC=overhead cam; OHV=overhead
valve; all costs are incremental to the baseline.
3.4.1.6 Variable Valve Actuation (VVA)
For this final rule, consistent with the proposal, NHTSA has combined two valve
control technologies for OHV engines. Coupled cam phasing (CCPO) and discrete valve lift
(DVVLO) into one technology defined as variable valve actuation (VVA). The agency
estimates the incremental effectiveness for WA applied to and OHV engine as 2.71 to 3.59
percent. This effectiveness value is slightly lower than coupled cam phasing for overhead cam
applications (CCPS) based on the assumption that VVA would be applied to an OHV engine
after cylinder deactivation (DEAC). For more information on combining these technologies
please refer to NHTSA's FRIA.
3.4.1.7 Stoichiometric Gasoline Direct Injection (SGDI)
Stoichiometric gasoline direct injection (SGDI), or Spark Ignition Direct injection
(SIDI), engines inject fuel at high pressure directly into the combustion chamber (rather than
the intake port in port fuel injection). SGDI requires changes to the injector design, an
additional high pressure fuel pump, new fuel rails to handle the higher fuel pressures and
changes to the cylinder head and piston crown design. Direct injection of the fuel into the
cylinder improves cooling of the air/fuel charge within the cylinder, which allows for higher
compression ratios and increased thermodynamic efficiency without the onset of combustion
knock. Recent injector design advances, improved electronic engine management systems
and the introduction of multiple injection events per cylinder firing cycle promote better
mixing of the air and fuel, enhance combustion rates, increase residual exhaust gas tolerance
and improve cold start emissions. SGDI engines achieve higher power density and match
well with other technologies, such as boosting and variable valvetrain designs.
Several manufacturers are manufacturing vehicles with SGDI engines, including
VW/Audi, BMW, Toyota, Ford, and General Motors. Additionally, BMW, GM, Ford and
VW/Audi have announced plans to significantly increase the number of SGDI engines in their
portfolios.
NHTSA and EPA reviewed estimates from the MYs 2012-2016 final rule and 2010
TAR, which stated an effectiveness range of SGDI to be between 2 and 3 percent. NHTSA
and EPA reviewed estimates from the Alliance of Automobile Manufactures, which projects 3
percent gains in fuel efficiency and a 7 percent improvement in torque. The torque increase
provides the opportunity to downsize the engine allowing an increase in efficiency of up to a
5.8 percent. NHTSA and EPA also reviewed other published literature, reporting 3 percent
effectiveness for SGDI.41 Confidential manufacturer data reported an efficiency effectiveness
3-88
-------
Technologies Considered in the Agencies' Analysis
range of 1 to 2 percent. Based on data from the 2011 Ricardo study and reconfiguration of the
new lumped parameter model, EPA and NHTSA have revised this value to 1.5 percent1111.
Combined with other technologies (i.e.., boosting, downsizing, and in some cases, cooled
EGR), SGDI can achieve greater reductions in fuel consumption and CC>2 emissions
compared to engines of similar power output.
The NHTSA and EPA cost estimates for SGDI take into account the changes required
to the engine hardware, engine electronic controls, ancillary and Noise Vibration and
Harshness (NVH) mitigation systems. Through contacts with industry NVH suppliers, and
manufacturer press releases, the agencies believe that the NVH treatments will be limited to
the mitigation of fuel system noise, specifically from the injectors and the fuel lines and have
included corresponding cost estimates for these NVH controls. In the 2012-2016 FRM, the
agencies estimated the DMC for SGDI at $213 (2007$), $321 (2007$) and $386 (2007$) for
13/14, V6 and V8 engines, respectively. Adjusted for 2010$, these DMCs become $222
(2010$), $334 (2010$) and $402 (2010$) for this analysis all of which are considered
applicable in MY 2012. This technology is considered to be on the flat-portion of the learning
curve. The agencies have applied a medium complexity ICM of 1.39 to this technology
through 2018 and 1.29 thereafter. The resultant costs are shown in Table 3-30.
Table 3-30 Costs for Stoichiometric Gasoline Direct Injection (2010$)
Cost type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
Engine type
13/14
V6
V8
13/14
V6
V8
13/14
V6
V8
2017
$192
$290
$348
$84
$127
$153
$277
$417
$501
2018
$188
$284
$341
$84
$127
$153
$273
$411
$494
2019
$185
$278
$335
$63
$95
$114
$248
$373
$449
2020
$181
$273
$328
$63
$95
$114
$244
$367
$442
2021
$177
$267
$321
$63
$95
$114
$240
$362
$435
2022
$174
$262
$315
$63
$94
$114
$236
$356
$429
2023
$170
$257
$309
$63
$94
$113
$233
$351
$422
2024
$167
$251
$302
$62
$94
$113
$229
$346
$416
2025
$164
$246
$296
$62
$94
$113
$226
$340
$409
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are incremental to the baseline.
3.4.1.8 Turbocharging and Downsizing (TRBDS)
The specific power of a naturally aspirated engine is primarily limited by the rate at
which the engine is able to draw air into the combustion chambers. Turbocharging and
supercharging (grouped together here as boosting) are two methods to increase the intake
manifold pressure and cylinder charge-air mass above naturally aspirated levels. Boosting
increases the airflow into the engine, thus increasing the specific power level, and with it the
ability to reduce engine displacement while maintaining performance. This effectively
reduces the pumping losses at lighter loads in comparison to a larger, naturally aspirated
engine.
14 However, because GDI is a key enabler for modern, highly downsized turbocharged engines, this difference
will be overshadowed by the higher effectiveness for turbocharging and downsizing when they are combined
into packages.
3-89
-------
Technologies Considered in the Agencies' Analysis
Almost every major manufacturer currently markets a vehicle with some form of
boosting. While boosting has been a common practice for increasing performance for several
decades, turbocharging has considerable potential to improve fuel economy and reduce CO2
emissions when the engine displacement is also reduced. Specific power levels for a boosted
engine often exceed 100 hp/L, compared to average naturally aspirated engine power densities
of roughly 70 hp/L. As a result, engines can be downsized roughly 30 percent or higher while
maintaining similar peak output levels. In the last decade, improvements to turbocharger
turbine and compressor design have improved their reliability and performance across the
entire engine operating range. New variable geometry turbines and ball-bearing center
cartridges allow faster turbocharger spool-up (virtually eliminating the once-common "turbo
lag") while maintaining high flow rates for increased boost at high engine speeds. Low speed
torque output has been dramatically improved for modern turbocharged engines. However,
even with turbocharger improvements, maximum engine torque at very low engine speed
conditions, for example launch from standstill, is increased less than at mid and high engine
speed conditions. The potential to downsize engines may be less on vehicles with low
displacement to vehicle mass ratios for example a very small displacement engine in a vehicle
with significant curb weight, in order to provide adequate acceleration from standstill,
particularly up grades or at high altitudes.
Use of GDI systems with turbocharged engines and air-to-air charge air cooling also
reduces the fuel octane requirements for knock limited combustion and allows the use of
higher compression ratios. Ford's "Ecoboost" downsized, turbocharged GDI engines
introduced on MY 2010 vehicles allow the replacement of V8 engines with V6 engines with
improved in 0-60 mph acceleration and with fuel economy improvements of up to 12
percent.42
Recently published data with advanced spray-guided injection systems and more
aggressive engine downsizing targeted towards reduced fuel consumption and CC>2 emissions
reductions indicate that the potential for reducing CO2 emissions for turbocharged, downsized
GDI engines may be as much as 15 to 30 percent relative to port-fuel-injected
97 9R 9Q ^0 ^ 1
engines. ' ' ' ' Confidential manufacturer data suggests an incremental range of fuel
consumption and CC>2 emission reduction of 4.8 to 7.5 percent for turbocharging and
downsizing. Other publicly-available sources suggest a fuel consumption and CO2 emission
reduction of 8 to 13 percent compared to current-production naturally-aspirated engines
without friction reduction or other fuel economy technologies: a joint technical paper by
Bosch and Ricardo suggesting fuel economy gain of 8 to 10 percent for downsizing from a 5.7
liter port injection V8 to a 3.6 liter V6 with direct injection using a wall-guided direct
injection system;43 a Renault report suggesting a 11.9 percent NEDC fuel consumption gain
for downsizing from a 1.4 liter port injection in-line 4-cylinder engine to a 1.0 liter in-line 4-
cylinder engine, also with wall-guided direct injection;44 and a Robert Bosch paper suggesting
a 13 percent NEDC gain for downsizing to a turbocharged DI engine, again with wall-guided
injection.45 These reported fuel economy benefits show a wide range depending on the SGDI
technology employed.
NHTSA and EPA reviewed estimates from the 2012-2016 final rule, the TAR, and
existing public literature. The previous estimate from the MYs 2012-2016 suggested a 12 to
14 percent effectiveness improvement, which included low friction lubricant (level one),
3-90
-------
Technologies Considered in the Agencies' Analysis
engine friction reduction (level one), DCP, DVVL and SGDI, over baseline fixed-valve
engines, similar to the estimate for Ford's Ecoboost engine, which is already in production.
Additionally, the agencies analyzed Ricardo vehicle simulation data for various turbocharged
engine packages. Based on this data, and considering the widespread nature of the public
estimates, the effectiveness of turbocharging and downsizing is highly dependent upon
implementation and degree of downsizing.
In alignment with these variances, for this final rule, consistent with the proposal, the
agencies evaluated 4 different levels of downsized and turbocharged high Brake Mean
Effective Pressure (BMEP)11. engines; 18-bar, 24-bar, 24-bar with cooled exhaust gas
recirculation (EGR) and 27'-bar with cooled EGR All engines are assumed to include
gasoline direct injection (SGDI) and effectiveness values include the benefits of this
technology. In addition, the agencies believe to implement in production a 27 bar boost level,
it is necessary to incorporate cooled exhaust gas recirculation (EGR) and also require a 2-
stage turbocharger as well as engine changes to increase robustness. The cooled EGR
technology is discussed later in this section.
NHTSA and EPA have revised the effectiveness to reflect this new information and
assume that turbocharging and downsizing, alone, will provide a 12 to 24.6 percent
effectiveness improvement (dependent upon degree of downsizing and boost levels) over
naturally aspirated, fixed-valve engines. More specifically, 12.1 to 14.9 percent for 18-bar
engines, which is equal to the boost levels evaluated in the MYs 2012-2016 final rule,
assuming 33 percent downsizing, 16.4 to 20.1 percent for 24-bar engines, assuming 50
percent downsizing, 19.3 to 23.0 percent for 24-bar engines with cooled EGR, assuming 50
percent downsizing and 20.6 to 24.6 percent for 27-bar engines with cooled EGR, assuming
56 percent downsizing. For comparison purposes an 18-bar engine with low friction lubricant
(level one), engine friction reduction (level one), DCP, DVVL and SGDI, which is equivalent
to MYs 2012-2016 assumed turbocharging and downsizing technology, now results in a 16.8
to 20.9 percent effectiveness improvement. Coupling turbocharging and downsizing with low
friction lubricant (level one and two), engine friction reductions (level one and two), DCP,
DVVL and SGDI, for the MYs 2017-2025 timeframe, yields 18.0 to 22.4 percent for 18-bar
engines 20.4 to 25.2 percent for 24-bar engines, 23.2 to 27.9 percent for 24-bar engine with
cooled EGR and 24.0 to 28.8 percent for 27-bar with cooled EGR over naturally aspirated,
fixed-valve engines.
As noted above, the agencies relied on engine teardown analyses conducted by EPA,
FEV and Munro to develop costs for turbocharged GDI engines.46 In the 2012-2016 FRM,
the agencies estimated the DMC for turbocharging to 18 bar BMEP at $404 (2007$) and $681
(2007$) for 14 and V6/V8 engines, respectively, where the higher cost for the V-configuration
11 Brake Mean Effective Pressure is the average amount of pressure in pounds per square inch (psi) that must be
exerted on the piston to create the measured horsepower. This indicates how effective an engine is at filling the
combustion chamber with an air/fuel mixture, compressing it and achieving the most power from it. A higher
BMEP value contributes to higher overall efficiency.
3-91
-------
Technologies Considered in the Agencies' Analysis
engines represents twin turbochargers versus the single turbocharger in the I-configuration
engine. These DMCs become $420 (2010$) and $708 (2010$), respectively, for this analysis.
In the 2010 TAR, the agencies presented costs for 24 bar BMEP turbocharging at 1.5x the
cost of the 18 bar BMEP technology. This additional cost covered the incremental cost
increase of a variable geometry turbocharger (see 2010 TAR at page B-12). Thus, the DMC
for 24 bar BMEP would be $630 (2010$) and $1,062 (2010$) for I-configuration and V-
configuration engines, respectively. Note also for this final rule, the agencies are estimating
the DMC of the 27 bar BMEP technology at 2.5x the 18 bar BMEP technology, or $1,050
(2010$) and $1,771 (2010$) for I-configuration and V-configuration engines, respectively. All
of these turbocharger-related DMCs are considered applicable in the 2012MY. The agencies
consider each turbocharger technology to be on the flat portion of the learning curve and have
applied a medium complexity ICM of 1.39 through 2018 for 18 bar and through 2024 for 24
and 27 bar, then 1.29 to each thereafter. The resultant costs are shown in Table 3-31.
Table 3-31 Costs for Turbocharging (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
Technology
(BMEP)
18 bar
18 bar
24 bar
24 bar
27 bar
27 bar
18 bar
18 bar
24 bar
24 bar
27 bar
27 bar
18 bar
18 bar
24 bar
24 bar
27 bar
27 bar
Engine type
I-engine
V-engine
I-engine
V-engine
I-engine
V-engine
I-engine
V-engine
I-engine
V-engine
I-engine
V-engine
I-engine
V-engine
I-engine
V-engine
I-engine
V-engine
2017
$365
$614
$547
$922
$911
$1,536
$160
$270
$240
$405
$401
$675
$525
$885
$787
$1,327
$1,312
$2,211
2018
$357
$602
$536
$903
$893
$1,505
$160
$270
$240
$404
$400
$674
$517
$872
$776
$1,308
$1,293
$2,179
2019
$350
$590
$525
$885
$875
$1,475
$120
$202
$239
$403
$399
$672
$470
$792
$765
$1,289
$1,274
$2,148
2020
$343
$578
$515
$867
$858
$1,446
$119
$201
$239
$403
$398
$671
$462
$779
$754
$1,270
$1,256
$2,117
2021
$336
$567
$504
$850
$841
$1,417
$119
$201
$238
$402
$397
$670
$455
$768
$743
$1,252
$1,238
$2,087
2022
$330
$555
$494
$833
$824
$1,389
$119
$200
$238
$401
$397
$669
$448
$756
$732
$1,234
$1,220
$2,057
2023
$323
$544
$484
$816
$807
$1,361
$119
$200
$238
$400
$396
$667
$442
$744
$722
$1,217
$1,203
$2,028
2024
$316
$533
$475
$800
$791
$1,334
$118
$200
$237
$400
$395
$666
$435
$733
$712
$1,200
$1,186
$2,000
2025
$310
$523
$465
$784
$775
$1,307
$118
$199
$177
$299
$296
$499
$428
$722
$643
$1,083
$1,071
$1,805
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are incremental to the baseline.
The costs for the downsizing portion of the turbo/downsize technology is more
complex. The agencies have described those cost and how they were developed—based
primarily on FEV teardowns but some were scaled based on teardowns to generate costs for
downsizing situations that were not covered by teardowns—in both the 2012-2016 FRM and
the 2010 TAR. The DMCs used for this analysis are identical to those used in the 2010 TAR
except that they have been updated to 2010 dollars. Notable is the fact that many of the
downsizing costs are negative because they result in fewer parts and less material than the
engine from which they are "derived." For example a V8 engine could be replaced by a
turbocharged V6 engine having two fewer cylinders and as many as eight fewer valves (in the
case of a V8 DOHC downsized to a V6 DOHC). Importantly, the agencies have used an
approach to calculating indirect costs that results in positive indirect costs regardless of
whether the DMC is positive or negative. This is done by calculating indirect costs based on
the absolute value of the DMC, then adding the indirect cost to the DMC to arrive at the total
3-92
-------
Technologies Considered in the Agencies' Analysis
cost. This way, the agencies are never making a negative DMC "more negative" when
accounting for the indirect costs. This approach has been used in the 2012-2016 final rule and
the 2010 TAR. Given the history of the downsizing costs used by the agencies, many are
considered applicable in the 2012MY and many in the 2017MY.JJ All are considered to be on
the flat portion of the learning curve. The agencies have applied a medium complexity ICM
of 1.39 through 2018 and 1.29 thereafter. The resultant costs are shown in Table 3-32.
Table 3-32 Costs for Engine Downsizing (2010$)
Cost
type
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
TC
Technology
14 DOHC to 13
14 DOHC to 14
V6 DOHC to 14
V6SOHC2Vto
14
V6OHVtoI4
V8 DOHC to 14
V8DOHCtoV6
V8SOHC2Vto
14
V8SOHC3Vto
14
V8SOHC2Vto
V6
V8SOHC3Vto
V6
V8OHVtoI4
V8OHVtoV6
14 DOHC to 13
14 DOHC to 14
V6 DOHC to 14
V6SOHC2Vto
14
V6OHVtoI4
V8 DOHC to 14
V8DOHCtoV6
V8SOHC2Vto
14
V8SOHC3Vto
14
V8SOHC2Vto
V6
V8SOHC3Vto
V6
V8OHVtoI4
V8OHVtoV6
14 DOHC to 13
2017
-$174
-$77
-$494
-$345
$281
-$854
-$247
-$656
-$731
-$76
-$140
-$242
$328
$77
$34
$217
$152
$109
$331
$109
$254
$283
$33
$62
$94
$127
-$98
2018
-$171
-$75
-$484
-$338
$272
-$828
-$242
-$637
-$709
-$74
-$137
-$234
$318
$76
$34
$217
$151
$108
$330
$108
$253
$282
$33
$61
$93
$126
-$94
2019
-$167
-$74
-$474
-$331
$264
-$804
-$237
-$617
-$687
-$73
-$135
-$227
$308
$57
$25
$162
$113
$81
$246
$81
$189
$210
$25
$46
$70
$94
-$110
2020
-$164
-$72
-$465
-$325
$256
-$779
-$233
-$599
-$667
-$71
-$132
-$220
$299
$57
$25
$162
$113
$81
$245
$81
$188
$210
$25
$46
$69
$94
-$107
2021
-$161
-$71
-$455
-$318
$249
-$756
-$228
-$581
-$647
-$70
-$129
-$214
$290
$57
$25
$161
$113
$80
$244
$81
$188
$209
$25
$46
$69
$94
-$104
2022
-$157
-$69
-$446
-$312
$241
-$733
-$223
-$564
-$627
-$68
-$127
-$207
$281
$57
$25
$161
$113
$80
$244
$81
$187
$208
$25
$46
$69
$93
-$101
2023
-$154
-$68
-$437
-$306
$236
-$719
-$219
-$552
-$615
-$67
-$124
-$203
$276
$57
$25
$161
$112
$80
$243
$80
$187
$208
$25
$46
$69
$93
-$98
2024
-$151
-$67
-$429
-$300
$232
-$704
-$215
-$541
-$603
-$66
-$122
-$199
$270
$57
$25
$161
$112
$80
$243
$80
$187
$208
$25
$46
$69
$93
-$95
2025
-$148
-$65
-$420
-$294
$227
-$690
-$210
-$530
-$591
-$64
-$119
-$195
$265
$57
$25
$160
$112
$80
$242
$80
$186
$207
$25
$45
$69
$93
-$92
JJ The engine downsize costs based on actual FEV teardowns were considered applicable to the 2012MY, as was
explained for some downsize costs in the 2012-2016 final rule and others in the 2010 TAR. For other downsize
costs—the two changes from OHV engines to DOHC engines—the agencies did not use FEV teardowns or
extrapolations from FEV teardowns, and instead used the methodology employed in the 2008 EPA Staff Report,
a methodology determined by both agencies to result in cost estimates more appropriate for the 2017MY. The
new downsize costs—those for V8 engines downsized to 14 engines—use a combination of V8 to V6 then V6 to
14 downsize costs and are considered applicable to the 2017MY within the context of this analysis.
3-93
-------
Technologies Considered in the Agencies' Analysis
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
14 DOHC to 14
V6DOHCtoI4
V6SOHC2Vto
14
V6OHVtoI4
V8 DOHC to 14
V8DOHCtoV6
V8SOHC2Vto
14
V8SOHC3Vto
14
V8SOHC2Vto
V6
V8SOHC3Vto
V6
V8OHVtoI4
V8OHVtoV6
-$43
-$277
-$193
$390
-$523
-$139
-$402
-$448
-$42
-$79
-$148
$454
-$41
-$267
-$187
$381
-$499
-$134
-$383
-$427
-$41
-$76
-$141
$444
-$48
-$312
-$218
$345
-$558
-$156
-$429
-$477
-$48
-$89
-$158
$403
-$47
-$303
-$212
$337
-$534
-$152
-$411
-$457
-$46
-$86
-$151
$393
-$46
-$294
-$205
$329
-$512
-$147
-$393
-$438
-$45
-$83
-$145
$384
-$44
-$285
-$199
$321
-$490
-$143
-$376
-$419
-$44
-$81
-$139
$375
-$43
-$277
-$193
$316
-$476
-$138
-$365
-$407
-$42
-$78
-$134
$369
-$42
-$268
-$187
$311
-$462
-$134
-$355
-$395
-$41
-$76
-$131
$363
-$40
-$260
-$182
$307
-$448
-$130
-$344
-$383
-$40
-$74
-$127
$358
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are incremental to the baseline; all
resultant engines are DOHC.
Note that the V8 to 14 engine downsize is new for this final rule, consistent with the
proposal. This level of engine downsizing is considered for this analysis only if it also
includes 27 bar BMEP turbo boost which, in addition, requires the addition of cooled EGR
(discussed below). As a result, any 27 bar BMEP engine in this analysis will be 14
configuration and will include cooled EGR.
With the information shown in Table 3-31 and Table 3-32, the costs for any
turbo/downsize change can be determined. These costs are shown in Table 3-33.
Table 3-33 Total Costs for Turbo/Downsizing (2010$)
Downsize
Technology
14 DOHC to 13
14 DOHC to 13
14 DOHC to 13
14 DOHC to 14
14 DOHC to 14
14 DOHC to 14
V6 DOHC to 14
V6 DOHC to 14
V6 DOHC to 14
V6 SOHC 2V to
14
V6 SOHC 2V to
14
V6 SOHC 2V to
14
V6 OHV to 14
V6 OHV to 14
V6 OHV to 14
V8 DOHC to 14
V8 DOHC to 14
V8 DOHC to 14
V8 DOHC to V6
V8 DOHC to V6
V8 DOHC to V6
Turbo
Technology
(BMEP)
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
2017
$427
$690
$1,214
$482
$744
$1,269
$248
$510
$1,035
$331
$594
$1,119
$914
$1,177
$1,701
$1
$264
$789
$746
$1,188
$2,073
2018
$423
$681
$1,199
$476
$734
$1,251
$250
$508
$1,026
$330
$589
$1,106
$898
$1,156
$1,674
$18
$277
$794
$738
$1,174
$2,045
2019
$359
$654
$1,164
$421
$716
$1,226
$157
$452
$962
$251
$546
$1,056
$815
$1,110
$1,619
-$88
$207
$716
$635
$1,132
$1,991
2020
$356
$647
$1,149
$415
$707
$1,209
$159
$450
$953
$251
$542
$1,044
$799
$1,090
$1,593
-$72
$219
$722
$628
$1,118
$1,965
2021
$352
$639
$1,134
$410
$697
$1,192
$161
$449
$944
$250
$537
$1,032
$784
$1,072
$1,567
-$56
$231
$726
$620
$1,105
$1,940
2022
$348
$632
$1,120
$404
$688
$1,176
$163
$447
$935
$249
$533
$1,021
$770
$1,053
$1,542
-$41
$243
$731
$613
$1,092
$1,914
2023
$344
$624
$1,106
$399
$679
$1,160
$165
$445
$927
$248
$529
$1,010
$758
$1,038
$1,519
-$34
$246
$728
$606
$1,078
$1,890
2024
$340
$617
$1,092
$393
$670
$1,145
$167
$444
$918
$248
$524
$999
$746
$1,023
$1,498
-$27
$250
$725
$599
$1,066
$1,866
2025
$337
$551
$979
$388
$602
$1,031
$169
$383
$811
$247
$461
$890
$735
$949
$1,378
-$19
$195
$623
$592
$953
$1,675
3-94
-------
Technologies Considered in the Agencies' Analysis
V8 SOHC 2V to
14
V8 SOHC 2V to
14
V8 SOHC 2V to
14
V8 SOHC 3V to
14
V8 SOHC 3V to
14
V8 SOHC 3V to
14
V8 SOHC 2V to
V6
V8 SOHC 2V to
V6
V8 SOHC 2V to
V6
V8 SOHC 3V to
V6
V8 SOHC 3V to
V6
V8 SOHC 3V to
V6
V8 OHV to 14
V8 OHV to 14
V8 OHV to 14
V8 OHV to V6
V8 OHV to V6
V8 OHV to V6
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
18 bar
24 bar
27 bar
$123
$385
$910
$77
$339
$864
$842
$1,284
$2,169
$806
$1,248
$2,133
$377
$639
$1,164
$1,339
$1,781
$2,666
$134
$392
$910
$90
$349
$866
$831
$1,267
$2,138
$796
$1,232
$2,103
$376
$635
$1,152
$1,316
$1,752
$2,623
$41
$336
$846
-$8
$287
$797
$744
$1,241
$2,100
$703
$1,200
$2,059
$312
$607
$1,116
$1,194
$1,691
$2,550
$52
$343
$845
$5
$296
$799
$733
$1,224
$2,071
$693
$1,184
$2,031
$311
$602
$1,105
$1,172
$1,663
$2,510
$62
$350
$845
$18
$305
$800
$723
$1,207
$2,042
$684
$1,169
$2,003
$311
$598
$1,093
$1,151
$1,636
$2,471
$72
$356
$844
$29
$313
$801
$712
$1,191
$2,014
$675
$1,153
$1,976
$310
$594
$1,082
$1,131
$1,609
$2,432
$76
$357
$838
$35
$315
$796
$702
$1,175
$1,986
$666
$1,138
$1,950
$307
$587
$1,069
$1,113
$1,586
$2,397
$80
$357
$832
$40
$317
$791
$692
$1,159
$1,959
$657
$1,124
$1,924
$304
$581
$1,056
$1,096
$1,563
$2,363
$84
$298
$727
$45
$259
$688
$682
$1,043
$1,766
$648
$1,010
$1,732
$302
$516
$944
$1,080
$1,441
$2,163
All costs are total costs (Direct manufacturing costs + Indirect costs); all costs are incremental to the
baseline; all resultant engines are DOHC; note that costs are shown for 27 bar BMEP engines with V6
engines. In fact, the agencies do not believe that manufacturers will employ 27 bar BMEP technology on V6
engines to comply with the final standards, instead using the additional boost to allow for downsizing V6
engines to smaller 14 engines than would be used for 18 bar BMEP or 24 bar BMEP 14 engines and/or
downsizing V8 engines to 14 engines. As a result, whenever a 27 bar BMEP engine is chosen by either
agency's model, the engine configuration will be an 14 and will include cooled EGR, as discussed in section
3.4.1.8.
3.4.1.9 Cooled Exhaust-Gas Recirculation (EGR)
While not considered in the technology packages used for assessing potential
compliance pathways in the 2012-2016 light-duty rule, the agencies have considered an
emerging technology referred to as cooled exhaust gas recirculation (cooled-EGR) as applied
to downsized, turbocharged GDI engines. In the 2010 TAR, the agencies considered this
technology as an advanced gasoline technology since it was considered an emerging and not
yet available technology in the light-duty gasoline market. While a cooled or "boosted" EGR
technology was discussed in the 2012-2016 light-duty rule record, the technology considered
here is comparatively more advanced as described in the 2010 TAR. As such, the agencies
have considered new costs and new effectiveness values for it. The effectiveness values used
for vehicle packages with cooled EGR within this analysis reflect a conservative estimate of
system performance at approximately 24-bar BMEP. Vehicle simulation modeling of
technology packages using the more highly boosted and downsized cooled EGR engines (up
3-95
-------
Technologies Considered in the Agencies' Analysis
to 27-bar BMEP, and utilizing EGR rates of 20-25%) with dual-stage turbocharging has been
completed as part of EPA's contract with Ricardo Engineering as described in 3.3.1.2. For
this FRM, consistent with the proposal, the agencies have updated the effectiveness of vehicle
packages with cooled EGR using the new Ricardo vehicle simulation modeling runs.
Cooled exhaust gas recirculation or Boosted EGR is a combustion concept that
involves utilizing EGR as a charge dilutent for controlling combustion temperatures and
cooling the EGR prior to its introduction to the combustion system. Higher exhaust gas
residual levels at part load conditions reduce pumping losses for increased fuel economy. The
additional charge dilution enabled by cooled EGR reduces the incidence of knocking
combustion and obviates the need for fuel enrichment at high engine power. This allows for
higher boost pressure and/or compression ratio and further reduction in engine displacement
and both pumping and friction losses while maintaining performance. Engines of this type
use GDI and both dual cam phasing and discrete variable valve lift. The EGR systems
considered in this final rule, consistent with the proposal, would use a dual-loop system with
both high and low pressure EGR loops and dual EGR coolers. The engines would also use
single-stage, variable geometry turbocharging with higher intake boost pressure available
across a broader range of engine operation than conventional turbocharged SI engines. Such a
system is estimated to be capable of an additional 3 to 5 percent effectiveness relative to a
turbocharged, downsized GDI engine without cooled-EGR.47'48 The agencies have also
considered a more advanced version of such a cooled EGR system that employs very high
combustion pressures by using dual stage turbocharging. This modeling work has been
completed by Ricardo Engineering. The simulation modeling is similar to work that Ricardo
conducted for EPA for its 2008 staff report on GHG effectiveness of light-duty vehicle
technologies.49 The agencies have considered this more advanced cooled EGR approach for
this final rule, consistent with the proposal.
For the MYs 2012-2016 final rule and TAR, NHTSA and EPA assumed a 5
percent fuel consumption effectiveness for cooled EGR compared to a conventional
downsized DI turbocharged engine.50 Based on the data from the Ricardo and Lotus
reports, NHTSA and EPA estimate the incremental reduction in fuel consumption for
EGR Boost to be 5 percent over a turbocharged and downsized DI engine. Thus, if
cooled EGR is applied to 24-bar engine, adding the 19.3 percent from the turbocharging
and downsizing to the 5 percent gain from cooled EGR results in total fuel consumption
reduction of 22.1 percent. This is in agreement with the range suggested in the Lotus
and Ricardo reports.
In the 2010 TAR, the agencies estimated the DMC of the cooled EGR system at $240
(2007$, see 2010 TAR at page B-12)). This DMC becomes $244 (2010$) for this analysis.
This DMC is considered applicable in the 2012MY. The agencies consider cooled EGR
technology to be on the flat portion of the learning curve and have applied a medium
complexity ICM of 1.39 through 2024 then 1.29 thereafter. The resultant costs are shown in
Table 3-34.
Table 3-34 Costs for Cooled EGR (2010$)
Cost
type
Engine type
2017
2018
2019
2020
2021
2022
2023
2024
2025
3-96
-------
Technologies Considered in the Agencies' Analysis
DMC
1C
TC
All
All
All
$212
$93
$305
$208
$93
$301
$204
$93
$296
$199
$93
$292
$195
$92
$288
$192
$92
$284
$188
$92
$280
$184
$92
$276
$180
$69
$249
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are
incremental to the baseline.
Note that, in the 2010 TAR, the agencies presented the cooled EGR system costs
inclusive of turbo charging costs (see 2010 TAR, Table B2.2-1 at page B-12). For this
analysis, the agencies are presenting the cooled EGR costs as a stand-alone technology that
can be added to any turbo/downsized engine provided sufficient boost is provided and
sufficient engine robustness is accounted for. As such, the cooled EGR system is considered
applicable only the 24 bar BMEP and 27 bar BMEP engines. Further, the agencies believe
that 24 bar BMEP engines are capable of maintaining NOx control without cooled EGR, so
each agency's respective models may choose 24 bar BMEP engines with and/or without
cooled EGR. However, as noted above, 27 bar BMEP engines are considered to require
cooled EGR to maintain NOx emission control. As such, neither agency's model is allowed
to choose 27 bar BMEP technology without also adding cooled EGR.
3.4.1.10 Diesel Engine Technology (DSL)
Diesel engines have several characteristics that give them superior fuel efficiency
compared to conventional gasoline, spark-ignited engines. Pumping losses are much lower
due to lack of (or greatly reduced) throttling in a diesel engine. The diesel combustion cycle
operates at a higher compression ratio than does a gasoline engine. As a result, turbocharged
light-duty diesels typically achieve much higher torque levels at lower engine speeds than
equivalent-displacement naturally-aspirated gasoline engines. Future high BMEP
turbocharged and downsized engines, mentioned above, are projected to improve torque
levels at lower engine speeds thus reducing the diesel advantage in this area. Diesels also
operate with a very lean air/fuel mixture. These attributes - reduced pumping losses, higher
compression ratio and lean/air fuel mixture — allow the engine to extract more energy from a
given mass of fuel than a gasoline engine, and thus make it more efficient. Additionally,
diesel fuel has higher energy content per gallon than does gasoline. While diesel fuel has a
higher energy content than gasoline, it also contains more carbon per gallon than does
gasoline: diesel produces 22.2 pounds of CC>2 per gallon when burned, while gasoline
produces 19.4 pounds of CO2 per gallon. This higher carbon content slightly offsets the GHG
emissions benefit of diesel fuel relative to gasoline, however, the disbenefit is more than
compensated by the greater efficiency of the diesel engine. Since diesel engines are more fuel
efficient than current naturally aspirated PFI gasoline engines, the agencies anticipate that
manufacturers will evaluate and potentially invest in diesel engine production as a way to
comply with more stringent CAFE standards. However, there are two primary reasons why
manufacturers might not choose to invest significantly in diesel engine technologies as a way
to comply with the CAFE and GHG standards for MYs 2017-2025.
As discussed above, even though diesel has higher energy content than gasoline it also
has a higher carbon density that results in higher amounts of CC>2 emitted per gallon,
approximately 15 percent more than a gallon of gasoline. This is commonly referred to as the
"carbon penalty" associated with using diesel fuel - a diesel vehicle yields greater fuel
3-97
-------
Technologies Considered in the Agencies' Analysis
economy improvements compared to its CC>2 emissions reduction improvements, so a
manufacturer that invests in diesel technology to meet CAFE standards may have more
trouble meeting the GHG standards than if it used a different and more cost effective (from a
GHG perspective) technology.
And second, diesel engines also have emissions characteristics that present challenges
to meeting federal Tier 2 NOx emissions standards. By way of comparison for readers
familiar with the European on-road fleet, which contains many more diesel vehicles than the
U.S. on-road fleet, U.S. Tier 2 emissions fleet average requirement of bin 5 require roughly
45 to 65 percent more NOX reduction compared to the Euro VI standards.
Despite considerable advances by manufacturers in developing Tier 2-compliant diesel
engines, it remains somewhat of a systems-engineering challenge to maintain the full fuel
consumption advantage of the diesel engine while meeting Tier 2 emissions regulations
because some of the emissions reduction strategies can increase fuel consumption (relative to
a Tier 1 compliant diesel engine), depending on the combination of strategies employed. A
combination of combustion improvements (that reduce NOx emissions leaving the engine)
and aftertreatment (capturing and reducing NOx emissions via a NOx adsorption catalyst, or
via selective catalytic reduction (SCR) using a reductant such as urea) that have left the
engine before they leave the vehicle tailpipe) are being introduced on Tier 2 compliant light-
duty diesel vehicles today. However, recently there have been a small number of
announcements that diesel engines will be added to some passenger cars, in some cases a
segment first for a manufacturer51, or that new passenger car diesel engines are being
designed to meet all global emissions regulations.52 This suggests to the agencies that some
manufacturers may be planning to use diesel engines in their plans to meet the tighter CAFE
standards in the mid-term, which may be enabled by advances in diesel engine and emission
control technology. Manufacturers that focus on diesel engines have also stated to the
agencies their expectation that diesel engines will continue to be a viable technology for
improving fuel economy and GHG emissions in the future.
We spend time here discussing available emissions reduction technologies for diesel
engines as part of this rulemaking because of the potential they have to impact fuel economy
and GHG emissions for the vehicles that have them. With respect to combustion
improvements, we note that several key advances in diesel engine combustion technology
have made it possible to reduce emissions coming from the engine prior to aftertreatment,
which reduces the need for aftertreatment. These technologies include improved fuel systems
(higher injection pressure and multiple-injection capability), advanced controls and sensors to
optimize combustion and emissions performance, higher EGR levels and EGR cooling to
reduce NOx, and advanced turbocharging systems. These systems are available today and
they do not adversely impact fuel efficiency. However, additional improvements in these
technologies will be needed to reduce engine emissions further, should future emissions
standards become more stringent. Further development may also be needed to reduce the fuel
efficiency penalty associated with EGR.
With respect to catalytic exhaust emission control systems, typical 3-way exhaust
catalysts without NOx storage capability are not able to reduce NOx emissions from engines
operated lean of stoichiometry (diesel or lean-burn gasoline). To reduce NOx, hydrocarbons,
3-98
-------
Technologies Considered in the Agencies' Analysis
and particulate emissions, all diesels will require a catalyzed diesel particulate filter (CDPF)
and sometimes a separate diesel oxidation catalyst (DOC), and either a lean NOx trap (LNT)
^ or the use of a selective catalytic reduction system, typically base-metal zeolite urea-SCR11.
The increased cost of diesel emissions control technologies relative to powertrains
with stoichiometric gasoline engines that are approaching comparable efficiency may also
make diesels less attractive to manufacturers as a technology solution for more stringent
CAFE and GHG standards. However, recognizing that some manufacturers may still employ
diesel technology to meet the future standards, the agencies have included diesels in our
analysis as follows:
The agencies sought to ensure that diesel engines would have equivalent performance
to comparable gasoline engine vehicles. For the Subcompact, Compact, and Midsize
Passenger Car, Performance Subcompact Car, and Small Light Truck vehicle subclasses, the
agencies assumed that an 14 gasoline base engine would be replaced by an in-line 4-cylinder
diesel engine with displacement varying around 2.0 liters. For the Performance Compact,
Performance Midsize, Large Passenger Car, Minivan, and Midsize Truck vehicle subclasses
for the CAFE model, the agencies assumed that a V6 gasoline base engine would be replaced
by an in-line 4-cylinder diesel engine with displacement varying around 2.8 liters. For the
Large Truck and Performance Large Car vehicle subclasses for the CAFE model, the agencies
assumed that a V8 gasoline base engine would be replaced with a V6 diesel engine with
displacement varying around 4.0 liters to meet vehicle performance requirements. It was also
assumed that diesel engines for all of these classes would utilize SCR aftertreatment systems
given recent improvements in zeolite-based SCR systems and system efficiency. These
assumptions impacted our estimates of the costs of implementing diesel engines as compared
to the base gasoline engines.
^ A lean NOX trap operates by oxidizing NO to NO2 in the exhaust and storing NO2 on alkali sorbent material,
most often BaO. When the control system determines (via mathematical model and typically a NOX sensor) that
the trap is saturated with NOX, it switches the engine into a operating mode just rich of stoichiometry that allow
NOx to be released from the alkali storage and temporarily allow three-way function of the catalyst similar to
three-way catalysts used in stoichiometric gasoline applications. LNTs preferentially store sulfate compounds
from the fuel, which reduces NOx storage capacity over time, thus the system must undergo periodic
desulfurization by operating at a net-fuel-rich condition at high temperatures in order to retain NOX trapping
efficiency.
11 An SCR aftertreatment system uses a reductant (typically, ammonia derived from urea) that is injected into the
exhaust stream ahead of the SCR catalyst. Ammonia is a strong reductant even under net lean conditions. It
combines with NOX in the SCR catalyst to form N2 and water. The hardware configuration for an SCR system is
sometimes more complicated than that of an LNT, due to the onboard urea storage and delivery system (which
requires a urea pump and injector to inject urea into the exhaust stream), which generally makes an SCR system
cost more than an LNT system. While a rich engine-operating mode is not required for NOX reduction, the urea
is typically injected at a rate of approximately 3 percent of the fuel consumed. The agencies understand that
manufacturers designing SCR systems intend to align urea tank refills with standard maintenance practices such
as oil changes as more diesel vehicles are introduced into the market. For diesel vehicles currently on the
market, this is generally already the practice, and represents an ongoing maintenance cost for vehicles with this
technology.
3-99
-------
Technologies Considered in the Agencies' Analysis
Diesel engines are more costly than port-injected spark-ignition gasoline engines.
These higher costs result from more costly components, more complex systems for emissions
control, and other factors. The vehicle systems that are impacted include:
• Fuel systems (higher pressures and more responsive injectors);
• Controls and sensors to optimize combustion and emissions performance;
• Engine design (higher cylinder pressures require a more robust engine, but higher
torque output means diesel engines can have reduced displacement);
• Turbocharger(s);
• Aftertreatment systems, which tend to be more costly for diesels;
In the MYs 2012-2016 final rule, the agencies estimated the DMC for converting a
gasoline PFI engine with 3-way catalyst aftertreatment to a diesel engine with diesel
aftertreatment at $1,697 (2007$), $2,399 (2007$), $1,956 (2007$) and $2,676 (2007$) for a
small car, large car, medium/large MPV & small truck, and large truck, respectively (see final
Joint TSD, Table 3-12 at page 3-44). All of these costs were for SCR-based diesel systems,
with the exception of the small car, which was a LNT-based system. For this final rule,
consistent with the proposal, we are using the same methodology as used in the MYs 2012-
2016 final rule, but have made four primary changes to the cost estimates as was also done in
the proposal for this rule. First, the agencies have not estimated costs for a LNT-based
system, and instead have estimated costs for all vehicle types assuming they will employ
SCR-based systems. Second, the agencies assumed that manufacturers would meet a Tier 2
bin 2 average rather than a Tier 2 bin 5 average, assuming that more stringent levels of
compliance will be required in the future. In order to estimate costs for Tier 2 bin 2 compliant
vehicles, catalyst volume costs were estimated based on an assumed increase in volume of 20
percent. This was the estimated necessary increase needed to meet Tier 2, bin 2 emission
level of 0.02 grams of NOx per mile. Increased catalyst volume resulted in a higher cost
estimate for diesel aftertreatment than was estimated for the MYs 2012-2016 final rule. The
third is to update all platinum group metal costs from the March 2009 values used in the
2012-2016 final rule to February 2011 values.111"1 The February 2011 values were used for
purposes of the NPRM analysis, at which time they represented the most recent monthly
average prices available at the time the agencies "locked-down" all cost estimates for the
purposes of moving into the modeling phase of analysis."11 For the final rule analysis, the
mm As reported by Johnson-Matthey, the March 2009 monthly average costs were $1,085 per Troy ounce and
$1,169 per Troy ounce for platinum (Pt) and rhodium (Rh), respectively. As also reported by Johnson-Matthey,
the February 2011 monthly average costs were $1,829 per Troy ounce and $2,476 per Troy ounce for Pt and Rh,
respectively. See www.platinum.matthey.com.
1111 Note that there is no good way of determining what PGM prices to use when conducting cost analyses. Spot
prices are inherently dangerous to use because spot prices, like stock prices on the stock market, can vary
considerably from day to day. One could argue that an average price is best, but average prices can vary
considerably depending on the length of time included in the average. And if too much time is included in the
average, then average prices from a time prior to PGM use in diesel engines may be included which would lead
some to conclude that we had cherry picked our values. Given no good option, it seems most transparent and
3-100
-------
Technologies Considered in the Agencies' Analysis
agencies did not update the cost for platinum group metals. The fourth is to include an
additional $50 DMC for all costs to cover costs associated with improvements to fuel and urea
controls. All of the diesel costs are considered applicable to MY 2012. The agencies
consider diesel technology to be on the flat portion of the learning curve and have applied a
medium complexity ICM of 1.39 through 2018, and then an ICM of 1.29 thereafter. The
resultant costs are shown in Table 3-35.
Table 3-35 Costs for Conversion to Advanced Diesel (2010$)
Cost type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
Vehicle class
Small car
Standard car
Large car
Small MPV
Large MPV
Large truck
Small car
Standard car
Large car
Small MPV
Large MPV
Large truck
Small car
Standard car
Large car
Small MPV
Large MPV
Large truck
2017
$2,059
$2,059
$2,522
$2,064
$2,082
$2,886
$905
$905
$1,109
$907
$915
$1,268
$2,965
$2,965
$3,631
$2,971
$2,996
$4,154
2018
$2,018
$2,018
$2,472
$2,023
$2,040
$2,828
$903
$903
$1,106
$905
$913
$1,266
$2,922
$2,922
$3,578
$2,928
$2,953
$4,094
2019
$1,978
$1,978
$2,422
$1,982
$1,999
$2,772
$675
$675
$827
$677
$683
$946
$2,653
$2,653
$3,249
$2,659
$2,682
$3,718
2020
$1,938
$1,938
$2,374
$1,943
$1,959
$2,716
$674
$674
$826
$676
$681
$945
$2,612
$2,612
$3,200
$2,618
$2,641
$3,661
2021
$1,900
$1,900
$2,326
$1,904
$1,920
$2,662
$673
$673
$824
$674
$680
$943
$2,572
$2,572
$3,151
$2,578
$2,600
$3,605
2022
$1,862
$1,862
$2,280
$1,866
$1,882
$2,609
$672
$672
$823
$673
$679
$941
$2,533
$2,533
$3,103
$2,539
$2,561
$3,550
2023
$1,824
$1,824
$2,234
$1,828
$1,844
$2,556
$671
$671
$821
$672
$678
$940
$2,495
$2,495
$3,056
$2,501
$2,522
$3,496
2024
$1,788
$1,788
$2,190
$1,792
$1,807
$2,505
$669
$669
$820
$671
$677
$938
$2,457
$2,457
$3,010
$2,463
$2,484
$3,443
2025
$1,752
$1,752
$2,146
$1,756
$1,771
$2,455
$668
$668
$819
$670
$676
$937
$2,420
$2,420
$2,964
$2,426
$2,446
$3,392
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are incremental to the baseline.
For the MYs 2012-16 final rule and TAR, NHTSA and EPA estimated the fuel
consumption reduction of a SCR-based diesel engine to be between 20 to 25 percent over a
baseline gasoline engine. NHTSA and EPA have revisited these values and have now
estimated, based on the Ricardo 2011 study, the effectiveness of a SCR-based diesel engine to
be 28.4 to 30.5 percent. For purposes of CC>2 reduction, EPA estimates a 7 to 20 percent for
light-duty diesels equipped with SCR.
3.4.2 Transmission Technologies
NHTSA and EPA have also reviewed the transmission technology estimates used in
the 2012-2016 final rule and the 2010 TAR. In doing so, NHTSA and EPA considered or
reconsidered all available sources and updated the estimates as appropriate. The section
below describes each of the transmission technologies considered for this rulemaking. As
discussed above, for the final rule NHTSA has updated the effectiveness values for advanced
transmissions when coupled to naturally-aspirated engines based on the ANL simulation
least self-serving to simply choose a price and report its basis. In the end, the PGM costs represent 16-23
percent of the diesel DMC in this analysis. Further, diesels play very little to no role in enabling compliance
with the final standards.
3-101
-------
Technologies Considered in the Agencies' Analysis
modeling. These changes are documented in detail in NHTSA's RIA. These changes are not
included in this joint TSD because they are specific to NHTSA's analysis only.
3.4.2.1 Improved Automatic Transmission Control (Aggressive Shift Logic and Early
Torque Converter Lockup)
Calibrating the transmission shift schedule to upshift earlier and quicker, and to lock-
up or partially lock-up the torque converter under a broader range of operating conditions can
reduce fuel consumption and CC>2 emissions. However, this operation can result in a
perceptible degradation in noise, vibration, and harshness (NVH). The degree to which NVH
can be degraded before it becomes noticeable to the driver is strongly influenced by
characteristics of the vehicle, and although it is somewhat subjective, it always places a limit
on how much fuel consumption can be improved by transmission control changes.
Aggressive Shift Logic and Early Torque Converter Lockup are best optimized
simultaneously when added to an automatic transmission due to the fact that adding both of
them requires only minor modifications to the transmission mechanical components or
calibration software. As a result, these two technologies are combined in the modeling when
added to an automatic transmission. Since a dual clutch transmission (DCT) has no torque
converter, the early torque converter lockup technology is not included when adding ASL to
the DCT.
3.4.2.2 Aggressive Shift Logic
During operation, a transmission's controller manages the operation of the
transmission by scheduling the upshift or downshift, and, in automatic transmissions, locking
or allowing the torque converter to slip based on a preprogrammed shift schedule. The shift
schedule contains a number of lookup table functions, which define the shift points and torque
converter lockup based on vehicle speed and throttle position, and other parameters such as
temperature. Aggressive shift logic (ASL) can be employed in such a way as to maximize
fuel efficiency by modifying the shift schedule to upshift earlier and inhibit downshifts under
some conditions, which reduces engine pumping losses and engine friction. The application
of this technology does require a manufacturer to confirm that drivability, durability, and
NVH are not significantly degraded.
For this final rule, consistent with the proposal, the agencies considered two levels of
ASL. The first level is that discussed in the 2012-2016 final rule and the 2010 TAR. ASL-
level 1 is an early upshift strategy whereby the transmission shifts to the next higher gear
"earlier" (or at lower RPM during a gradual acceleration) than would occur in a traditional
automatic transmission. This early upshift reduces fuel consumption by allowing the engine
to operate at a lower RPM and higher load, which typically moves the engine into a more
efficient operating region.
ASL-level 2 is a shift optimization strategy whereby the engine and/or transmission
controller(s) continuously evaluate all possible gear options that would provide the necessary
tractive power (while limiting the adverse effects on driveline NVH) and select the gear that
lets the engine run in the most efficient operating zone. Ricardo acknowledged in its report
that the ASL-level 2 ("shift optimization") strategy currently causes significant implications
3-102
-------
Technologies Considered in the Agencies' Analysis
for drivability and hence affects consumer acceptability. However, Ricardo recommended the
inclusion of this technology for the 2020-2025 timeframe with the assumption that
manufacturers will develop a means of yielding the fuel economy benefit without adversely
affecting driver acceptability. The agencies believe these drivability challenges could include
shift busyness - that is, a high level of shifting compared to current vehicles as perceived by
the customers. The agencies note that in confidential discussions with two major transmission
suppliers, the suppliers described transmission advances which reduce shifting time and
provide smoother torque transitions than today's designs, making the shifting event less
apparent to the driver, however these improvements will not influence the customer's
perception of shift business related to the changes in engine speed.
In addition, the agencies note that several auto companies and transmission firms have
announced future introduction of transmissions into the U.S. market with even a higher
number of gears than were included in the Ricardo simulation and in the agencies' feasibility
assessment for this final rule, consistent with the proposal (which is 8 forward speeds). These
announcements include both 9 and 10 speed transmissions which may present further
challenges with shift busyness, given the availability of one or two additional gears. At the
same time, the associated closer gear spacing will generally result in smaller engine speed
changes during shifting that may be less noticeable to the driver.
The agencies are including shift optimization in the analysis under the premise that
manufacturers and suppliers are developing means to mitigate these drivability issues by MY
2017, as assumed in the 2011 Ricardo study (more information on Ricardo's treatment of the
optimized shift strategy is described in Section 6.4 of the 2011 Ricardo report). If
manufacturers are not able to solve these drivability issues, the assumed effectiveness could
be lower and the cost could be higher or both. The agencies sought comment on the feasibility
of ASL-level 2 and the likelihood that manufacturers will be able to overcome the drivability
issues, however no comments were submitted on this issue.
In MYs 2012-2016 final rule, the agencies estimated an effectiveness improvement of
1 to 2 percent for aggressive shift logic which was supported by the 2002 NAS and
NESCCAF reports as well as confidential manufacturer data. The agencies updated the
effectiveness of ASL-level 1 ranging from 1.9 to 2.7 based on 2010 Ricardo study. In CAFE
model an incremental effectiveness ranging for both ASL and early torque converter lockup
ranging from 2.3 to 3.1 percent is applied (Early torque converter has effectiveness of 0.5
percent).
ASL-level 2 is new to this analysis which is based on the shift optimization algorithm
in 2011 Ricardo study. The effectiveness for ASL-level 2 ranges from 5.1 to 7.0 percent
improvement over transmission with unimproved shift logic or roughly 4 to 5 percent over a
transmission that already incorporates aggressive shift logic. In the CAFE model, an
incremental effectiveness ranging from 3.27 to 4.31 percent is applied.
In the 2012-2016 rule, the agencies estimated the DMC at $26 (2007$) which was
considered applicable to the 2015MY. This DMC becomes $27 (2010$) for this analysis.
The agencies consider ASL-level 1 technology to be on the flat portion of the learning curve
and have applied a medium complexity ICM of 1.39 through 2018 then 1.29 thereafter. For
3-103
-------
Technologies Considered in the Agencies' Analysis
ASL-level 2, the agencies are estimating the DMC at an equivalent $27 (2010$) except that
this cost is considered applicable to the 2017MY. Essentially this yields a nearly negligible
incremental cost for ASL-level 2 over ASL-level 1. The agencies consider ASL-level 2
technology to be on the flat portion of the learning curve and have applied a medium
complexity ICM of 1.39 through 2024 then 1.29 thereafter. The timing of the ASL-level 2
ICMs is different than that for the level 1 technology because the level 2 technology is newer
and not yet being implemented in the fleet. The resultant costs are shown in Table 3-36.
Note that both levels of ASL technology are incremental to the baseline system, so ASL-level
2 is not incremental to ASL-level 1.
Table 3-36 Costs for Aggressive Shift Logic Levels 1 & 2 (2010$)
Cost type
DMC
DMC
1C
1C
TC
TC
Technology
ASL-level 1
ASL-level 2
ASL-level 1
ASL-level 2
ASL-level 1
ASL-level 2
Transmission
type
All
All
All
All
All
All
2017
$26
$27
$7
$7
$33
$34
2018
$26
$27
$7
$7
$32
$33
2019
$25
$26
$5
$7
$30
$32
2020
$24
$25
$5
$7
$30
$32
2021
$24
$24
$5
$7
$29
$31
2022
$24
$24
$5
$7
$29
$30
2023
$23
$23
$5
$7
$28
$30
2024
$23
$23
$5
$7
$28
$29
2025
$22
$22
$5
$5
$27
$27
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are incremental to the baseline.
3.4.2.3 Early Torque Converter Lockup
A torque converter is a fluid coupling located between the engine and transmission in
vehicles with automatic transmissions and continuously-variable transmissions (CVT). This
fluid coupling allows for slip so the engine can run while the vehicle is idling in gear (as at a
stop light), provides for smoothness of the powertrain, and also provides for torque
multiplication during acceleration, and especially launch. During light acceleration and
cruising, the inherent slip in a torque converter causes increased fuel consumption, so modern
automatic transmissions utilize a clutch in the torque converter to lock it and prevent this
slippage. Fuel consumption can be further reduced by locking up the torque converter at
lower vehicle speeds, provided there is sufficient power to propel the vehicle, and noise and
vibration are not excessive.00 If the torque converter cannot be fully locked up for maximum
efficiency, a partial lockup strategy can be employed to reduce slippage. Early torque
converter lockup is applicable to all vehicle types with automatic transmissions. Some torque
converters will require upgraded clutch materials to withstand additional loading and the
slipping conditions during partial lock-up. As with aggressive shift logic, confirmation of
acceptable drivability, performance, durability and NVH characteristics is required to
successfully implement this technology.
00 Although only modifications to the transmission calibration software are considered as part of this technology,
very aggressive early torque converter lock up may require an adjustment to damper stiffness and hysteresis
inside the torque converter.
3-104
-------
Technologies Considered in the Agencies' Analysis
Regarding the effectiveness of Early Torque Converter Lockup, the 2012-2016 final
rule, TAR, and the 2010 Ricardo study estimated an effectiveness improvement of 0.4 to 0.5
percent.
In the 2012-2016 rule, the agencies estimated the DMC at $24 (2007$) which was
considered applicable to the 2015MY. This DMC remains $25 (2010$) for this analysis.pp
The agencies consider early torque converter lockup 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. The resultant costs are shown in Table 3-37.
Table 3-37 Costs for Early Torque Converter Lockup (2010$)
Cost type
DMC
1C
TC
Transmission
type
Automatic
Automatic
Automatic
2017
$24
$6
$30
2018
$23
$6
$29
2019
$23
$5
$27
2020
$22
$5
$27
2021
$22
$5
$27
2022
$21
$5
$26
2023
$21
$5
$26
2024
$20
$5
$25
2025
$20
$5
$25
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are incremental to
the baseline.
3.4.2.4 High Efficiency Gearbox
For this rule, a high efficiency gearbox refers to some or all of a suite of incremental
gearbox improvement technologies that should be available within the 2017 to 2025
timeframe. The majority of these improvements address mechanical friction within the
gearbox. These improvements include but are not limited to: shifting clutch technology
improvements (especially for smaller vehicle classes), improved kinematic design, dry sump
lubrication systems, more efficient seals, bearings and clutches (reducing drag), component
superfinishing and improved transmission lubricants. More detailed description can be found
in the 2011 Ricardo report53. Note that the high efficiency gearbox technology is applicable
to any type of transmission.
EPA analyzed detailed transmission efficiency input data provided by Ricardo and
implemented it directly into the lumped parameter model. Based on the LP effectiveness
resulting from these inputs, EPA and NHTSA estimate that a high efficiency gearbox can
provide a GHG or fuel consumption reduction in the range of 3.8 to 5.7 percent (3.8% for
4WD trucks with an unimproved rear axle) over a baseline automatic transmission in
MY2017 and beyond.
The agencies estimate the DMC of the high efficiency gearbox at $200 (2009$). We
have based this on the DMC for engine friction reduction in a V8 engine which, as presented
in Table 3-24 is $197 (2010$). In the proposal, we rounded this value up to $200 (2009$)
pp As is true throughout this presentation of cost estimates, the agencies round costs to the nearest dollar. In the
actual model input files, the cost in 2007$ would have been $23.68 and the cost in 2009$ is $24.42. So an
impact of the dollar-year conversion is reflected in the analysis even when it does not appear so in this
presentation.
3-105
-------
Technologies Considered in the Agencies' Analysis
which becomes $202 (2010$) for the final analysis. This DMC is considered applicable for
the 2017MY. The agencies consider high efficiency gearbox technology to be on the flat
portion of the learning curve and have applied a low complexity ICM of 1.24 through 2024
then 1.19 thereafter. The resultant costs are shown in Table 3-38.
Table 3-38 Costs for High Efficiency Gearbox (2010$)
Cost type
DMC
1C
TC
Transmission
type
Automatic/Dual
clutch
Automatic/Dual
clutch
Automatic/Dual
clutch
2017
$202
$49
$251
2018
$196
$49
$245
2019
$190
$49
$239
2020
$184
$49
$233
2021
$179
$49
$227
2022
$173
$49
$222
2023
$170
$48
$218
2024
$167
$48
$215
2025
$163
$39
$202
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; all costs are incremental to the
baseline.
3.4.2.5 Automatic 6-, 7- and 8-Speed Transmissions (NAUTO and 8SPD)
Manufacturers can also choose to replace 4- and 5-speed transmission with 6-, 7-, or
8-speed automatic transmissions. Additional ratios allow for further optimization of engine
operation over a wider range of conditions, but this is subject to diminishing returns as the
number of speeds increases. As additional planetary gear sets are added (which may be
necessary in some cases to achieve the higher number of ratios), additional weight and friction
are introduced. Also, the additional shifting of such a transmission can be perceived as
bothersome to some consumers, so manufacturers need to develop strategies for smooth
shifts. Some manufacturers are replacing 4- and 5-speed automatics with 6-speed automatics,
and 7- and 8-speed automatics have also entered production. While a six speed transmission
application was most prevalent for the 2012-2016 final rule, eight speed transmissions are
expected to be readily available and applied in the 2017 through 2025 timeframe.
As discussed in the MY 2011 CAFE final rule, confidential manufacturer data
projected that 6-speed transmissions could incrementally reduce fuel consumption by 0 to 5
percent from a baseline 4-speed automatic transmission, while an 8-speed transmission could
incrementally reduce fuel consumption by up to 6 percent from a baseline 4-speed automatic
transmission. GM has publicly claimed a fuel economy improvement of up to 4 percent for
its new 6-speed automatic transmissions.54 The 2008 EPA Staff Technical Report found a 4.5
to 6.5 percent fuel consumption improvement for a 6-speed over a 4-speed automatic
transmission.55 Based on this information, NHTSA estimated in the MY 2011 rule, that the
conversion to a 6-,7- and 8-speed transmission (NAUTO) from a 4 or 5-speed automatic
transmission with IATC would have an incremental fuel consumption benefit of 1.4 percent to
3.4 percent, for all vehicle classes. From a baseline 4 or 5 speed transmission without IATC,
the incremental fuel consumption benefit would be approximately 3 to 6 percent, which is
consistent with the EPA Staff Report estimate. In MYs 2012-2016 final rule, NHTSA and
EPA reviewed these effectiveness estimates and concluded that they remain accurate. While
the CAFE model follows the incremental approach discussed above, the GHG model
estimates the packaged effectiveness of 4.5 to 6.5 percent
3-106
-------
Technologies Considered in the Agencies' Analysis
In this FRM analysis, consistent with the proposal, the agencies divided the
improvement for this technology into two steps, first from 4 or 5 speed transmission to 6 or 7
speed transmission (NAUTO), then from 6 or 7 speed transmission to 8 speed transmission
(8SPD). The effectiveness estimates for NAUTO and 8SPD are based on 2011 Ricardo study.
In this FRM analysis, consistent with the proposal, the effectiveness for a 6-speed
transmission relative to a 4-speed base transmission ranges from 3.1 to 3.9 percent (2.1
percent for large truck with unimproved rear axle) including 7 percent of transmission
gearbox efficiency improvement that the agencies assumed accompanying the new 6 speed
transmission after MY 2010. NHTSA incorporated this effectiveness estimate into the CAFE
model as incremental improvement over IATC ranging from 1.89 to 2.13 percent. In this
FRM analysis, consistent with the proposal, the agencies assumed that 8-speed transmission
will not start to phase in until MY2017. NHTSA applied 8-speed automatic transmission
succeeding 6-speed automatic transmission to vehicles with towing requirement, such as
Minivan, Midsize light truck and large light truck. All other vehicle subclasses use 8-speed
DCT to succeed 6-speed DCT. The effectiveness for an 8-speed DCT relative to a 4-speed
DCT transmission ranges from 11.1 to 13.1 percent for subcompact car, small car and small
light truck. The effectiveness for an 8-speed automatic transmission relative to 4-speed
automatic transmission ranges for large CUV and large truck ranges from 8.7 to 9.2 percent in
the lumped parameter model. This translates into effectiveness in the range of 3.85 to 4.57
percent for an 8-speed DCT relative to a 6-speed DCT and 4.9 to 5.34 percent for 8-speed
automatic transmission relative to 6-speed automatic transmission in CAFE model.
In the 2010 TAR, the agencies estimated the DMC at -$13 (2008$) for a 6 speed
automatic transmission relative to a 4 speed auto transmission, applicable in the 2017MY (see
2010 TAR, Table B2.1-1 at page B-10). For the 2012MY, that DMC was -$15 (2008$),
although that value was not presented in the TAR. The latter DMC remains -$15 (2010$) for
this analysis which is considered to be applicable in the 2012MY. The agencies consider 6
speed automatic transmission 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. The resultant
costs are shown in Table 3-39.
New for the proposal was the cost of an 8 speed automatic transmission. For the cost
of this technology, the agencies have relied on a tear-down study completed by FEV since
publication of the TAR. 6 In that study, the 8 speed auto transmission was found to be $62
(2007$) more costly than the 6 speed auto transmission. This DMC becomes $64 (2010$) for
this analysis. Adding the $64 (2010$) to the -$15 (2010$) DMC for a 6 speed relative to a 4
speed, the 8 speed auto transmission relative to a 4 speed auto transmission would be $50
(2010$). The agencies consider this DMC to be applicable to the 2012MY. The agencies
consider the 8 speed auto transmission technology to be on the flat portion of the learning
curve and have applied a medium complexity ICM of 1.39 through the 2018MY then 1.29
3-107
-------
Technologies Considered in the Agencies' Analysis
thereafter.qq The resultant costs for both 6 speed and 8 speed auto transmissions are shown in
Table 3-39.
Table 3-39 Costs for 6 and 8 Speed Automatic Transmissions (2010$)
Cost
type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
Transmission type
6spAT from 4spAT
SspAT from 6spAT
SspAT from 4spAT
6spAT from 4spAT
SspAT from 6spAT
SspAT from 4spAT
6spAT from 4spAT
SspAT from 6spAT
SspAT from 4spAT
2017
-$13
$56
$43
$4
$25
$19
-$9
$80
$62
2018
-$13
$55
$42
$4
$24
$19
-$9
$79
$61
2019
-$12
$54
$41
$3
$18
$14
-$10
$72
$55
2020
-$12
$53
$40
$3
$18
$14
-$9
$71
$54
2021
-$12
$51
$40
$3
$18
$14
-$9
$70
$54
2022
-$12
$50
$39
$3
$18
$14
-$9
$69
$53
2023
-$11
$49
$38
$3
$18
$14
-$9
$68
$52
2024
-$11
$48
$37
$3
$18
$14
-$8
$67
$51
2025
-$11
$47
$37
$3
$18
$14
-$8
$66
$50
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; sp=speed; AT=automatic transmission
Note that the cost for the 8 speed automatic transmission relative to the 6 speed
automatic transmission is lower here than that used in the recent heavy-duty GHG rule. In
that rule, we remained consistent with the proposal for that rule which carried an estimated
DMC of $210 (2008$). That DMC was based on an estimate derived by NAS (see NAS
2010, Table 7-10). For this final rule, consistent with the proposal, we have chosen to use the
more recent DMC shown in Table 3-39 which is based on a tear-down analysis done by FEV.
3.4.2.6 Dual Clutch Transmissions / Automated Manual Transmissions (DCTAM)
An Automated Manual Transmission (AMT) is mechanically similar to a conventional
manual transmission, but shifting and launch functions are automatically controlled by the
electronics. There are two basic types of AMTs, single-clutch and dual-clutch (DCT). A
single-clutch AMT is essentially a manual transmission with automated clutch and shifting.
Because of shift quality issues with single-clutch designs, DCTs are far more common in the
U.S. and are the basis of the estimates that follow. A DCT uses separate clutches (and
separate gear shafts) for the even-numbered gears and odd-numbered gears. In this way, the
next expected gear is pre-selected, which allows for faster and smoother shifting. For
example, if the vehicle is accelerating in third gear, the shaft with gears one, three and five
has gear three engaged and is transmitting power. The shaft with gears two, four, and six is
idle, but has gear four engaged. When a shift is required, the controller disengages the odd-
gear clutch while simultaneously engaging the even-gear clutch, thus making a smooth shift.
If, on the other hand, the driver slows down instead of continuing to accelerate, the
transmission will have to change to second gear on the idling shaft to anticipate a downshift.
This shift can be made quickly on the idling shaft since there is no torque being transferred on
it.
qq This ICM would be applied to the 6 speed to 8 speed increment of $64 (2010$) applicable in 2012. The 4
speed to 6 speed increment would carry the low complexity ICM.
3-108
-------
Technologies Considered in the Agencies' Analysis
In addition to single-clutch and dual-clutch AMTs, there are also wet clutch and dry
clutch designs which are used for different types of vehicle applications. Wet clutch AMTs
offer a higher torque capacity that comes from the use of a hydraulic system that cools the
clutches. Wet clutch systems are less efficient than the dry clutch systems due to the losses
associated with hydraulic pumping. Additionally, wet AMTs have a higher cost due to the
additional hydraulic hardware required.
Overall, DCTs likely offer the greatest potential for effectiveness improvements
among the various transmission options presented in this report because they offer the
inherently lower losses of a manual transmission with the efficiency and shift quality
advantages of electronic controls. The lower losses stem from the elimination of the
conventional lock-up torque converter, and a greatly reduced need for high pressure hydraulic
circuits to hold clutches or bands to maintain gear ratios (in automatic transmissions) or hold
pulleys in position to maintain gear ratio (in Continuously Variable Transmissions).
However, the lack of a torque converter will affect how the vehicle launches from rest, so a
DCT will most likely be paired with an engine that offers sufficient torque at low engine
speeds to allow for adequate launch performance or provide lower launch gears to
approximate the torque multiplication of the torque converter to provide equivalent
performance.
In MYs 2012-2016 final rule, EPA and NHTSA estimated a 5.5 to 9.5 percent
improvement in fuel consumption over a baseline 4/5-speed automatic transmission for a wet
clutch DCT, which was assumed for all but the smallest of vehicle subclasses, Subcompact
and Compact cars and small LT. This results in an incremental effectiveness estimate of 2.7
to 4.1 percent over a 6-speed automatic transmission with IATC. For Subcompact and
Compact Cars and small LT, which were assumed to use a dry clutch DCT, NHTSA
estimated an 8 to 13 percent fuel consumption improvement over a baseline 4/5-speed
automatic transmission, which equates to a 5.5 to 7.5 percent incremental improvement over
the 6-speed transmission.
Based on the 2011 Ricardo study, EPA and NHTSA have concluded that 8 to 13
percent effectiveness is appropriate for 6-speed DCTs and 11 to 16 percent is appropriate for
8-speed DCTs for this final rule, consistent with the proposal. These values include not only
the DCT but also the increase in stepped gears and also a high efficiency gearbox (mentioned
later). Independent of other technologies, this translates to an effectiveness for the DCT,
alone, of 4 to 5% (for wet-clutch designs) and 5 to 6% (for dry-clutch designs) compared to a
baseline automatic transmission of similar vintage and number of fixed gears.
In this FRM analysis, consistent with the proposal, NHTSA applied an incremental
effectiveness of 4 percent for a 6-speed dry DCT and 3.4 to 3.8 percent for a wet DCT
compared to a 6-speed automatic transmission based on the lumped parameter model which
includes the accompanied transmission efficiency improvement for MY 2010 and after
transmissions. This translates to an effectiveness range of 7.4 to 8.6 percent compared to a 4
speed automatic transmission for dry clutch design and 7.4 to 7.9 percent for a wet clutch
design. NHTSA did not apply DCTs to vehicles with towing requirements, such as Minivan,
Midsize light truck and large pickup truck. EPA did not apply DCTs to vehicle types
classified as towing as described in Chapter 1 of EPA's RIA.
3-109
-------
Technologies Considered in the Agencies' Analysis
In the 2010 TAR, the agencies estimated the DMC at -$234 (2008$) for a 6 speed dry-
clutch DCT and -$165 for a 6 speed wet-clutch DCT with both DMCs applicable in the
2017MY (see 2010 TAR, Table B2.1-1 at page B-10) and both incremental to a 4 speed
automatic transmission. In the 2010 TAR, we pointed to Chapter 3 of the 2012-2016 final
joint TSD where we noted that the DCT costs of-$147 (2007$ and incremental to a 6-speed
automatic transmission) were based on a FEV tear-down study that assumed 450,000 units of
production. We went on to state that we did not consider there to be sufficient US capacity in
the 2012-2016 timeframe to produce 450,000 units and for that reason we were adjusting the
tear-down values accordingly. The TAR timeframe for consideration was 2017-2025, and in
the TAR we argued that production capacity would exist and that the FEV tear-down results
be valid without adjustment. As noted in the proposal to this rule, we continue to believe that
to be the case. In the final joint TSD supporting the 2012-2016 rule we also noted that the
negative tear-down estimates found by FEV were not surprising when considering the relative
simplicity of a dual-clutch transmission compared to an automatic transmission. Again, we
continue to consider this to be true.
For this analysis, we consider the 2010 TAR DMCs to be applicable to the 2012MY,
thus the DMCs become -$238 (2010$) and -$168 (2010$) for 6 speed dry- and wet-clutch
DCTs, respectively, both applicable in the 2012MY and incremental to a 4 speed auto
transmission. The agencies consider the 6 speed DCT technology to be on the flat portion of
the learning curve and have applied a medium complexity ICM of 1.39 through 2018 then
1.29 thereafter. The resultant costs are shown in Table 3-40.
New for this rulemaking is costing for an 8 speed DCT. For the cost of this
technology, the agencies have relied on a tear-down study completed by FEV since
publication of the TAR.57 In that study, the 8 speed DCT was found to be $198 (2007$) more
costly than the 6 speed DCT. This DMC increment becomes $206 (2010$) for this analysis.
Adding the $206 (2010$) to the -$238 (2010$) DMC and the -$168 (2010$) DMC for a 6
speed dry- and wet-clutch DCT, the 8 speed dry- and wet-clutch DCTs relative to a 4 speed
auto transmission would be -$32 (2010$) and $38 (2010$), respectively. The agencies
consider this DMC to be applicable to the 2012MY. The agencies consider the 8 speed DCT
technology to be on the flat portion of the learning curve and have applied a medium
complexity ICM of 1.39 through the 2024MY then 1.29 thereafter. The 8 speed DCT has a
later switch to long term ICMs because it is a newer technology that is not currently
implemented in the fleet. The resultant costs for both 6 speed and 8 speed DCTs are shown in
Table 3-40.
Table 3-40 Costs for 6 & 8 Speed Dual Clutch Transmissions (2010$)
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
Transmission type
6spDCT-dry
6sp DCT-wet
8sp DCT-dry
8sp DCT-wet
6spDCT-dry
6sp DCT-wet
8sp DCT-dry
8sp DCT-wet
2017
-$207
-$146
-$28
$33
$91
$64
$12
$14
2018
-$203
-$143
-$27
$32
$91
$64
$12
$14
2019
-$199
-$140
-$27
$32
$68
$48
$12
$14
2020
-$195
-$137
-$26
$31
$68
$48
$12
$14
2021
-$191
-$134
-$26
$30
$68
$48
$12
$14
2022
-$187
-$132
-$25
$30
$67
$48
$12
$14
2023
-$183
-$129
-$25
$29
$67
$47
$12
$14
2024
-$179
-$127
-$24
$29
$67
$47
$12
$14
2025
-$176
-$124
-$24
$28
$67
$47
$9
$11
3-110
-------
Technologies Considered in the Agencies' Analysis
TC
TC
TC
TC
6spDCT-dry
6sp DCT-wet
8sp DCT-dry
8sp DCT-wet
-$116
-$82
-$16
$47
-$112
-$79
-$15
$47
-$131
-$92
-$15
$46
-$127
-$89
-$14
$45
-$123
-$87
-$14
$45
-$119
-$84
-$13
$44
-$116
-$82
-$13
$44
-$112
-$79
-$12
$43
-$109
-$77
-$15
$39
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; sp=speed; dry=dry clutch; wet=wet-clutch
Note that all costs are relative to a 4 speed automatic transmission.
3.4.2.7 6-Speed Manual Transmissions (6MAN)
Manual transmissions are entirely dependent upon driver input to shift gears: the
driver selects when to perform the shift and which gear to select. This is the most efficient
transfer of energy of all transmission layouts, because it has the lowest internal gear losses,
with a minimal hydraulic system, and the driver provides the energy to actuate the clutch.
From a systems viewpoint, however, vehicles with manual transmissions have the drawback
that the driver may not always select the optimum gear ratio for fuel economy. Nonetheless,
increasing the number of available ratios in a manual transmission can improve fuel economy
by allowing the driver to select a ratio that optimizes engine operation more often. Typically,
this is achieved through adding overdrive ratios to reduce engine speed at cruising velocities
(which saves fuel through reduced engine pumping losses) and pushing the torque required of
the engine towards the optimum level. However, if the gear ratio steps are not properly
designed, this may require the driver to change gears more often in city driving, resulting in
customer dissatisfaction. Additionally, if gear ratios are selected to achieve improved launch
performance instead of to improve fuel economy, then no fuel saving effectiveness is realized.
The 2012-2016 final rule estimated an effectiveness increase of 0.5 percent for
replacing a 5-speed manual with a 6-speed manual transmission, which was derived from
confidential manufacturer data Based on the updated LPM for this 2017-2025 rule, NHTSA
has found that an effectiveness increase of 2.0 to 2.5 percent is possible when moving from a
5-speed to a 6-speed manual transmission with improved internals. NHTSA updated costs
from the 2012-2016 final rule to reflect the ICM low complexity markup of 1.11 which
resulted in an incremental compliance cost of $250 as compared to $338 for MY 2012. This
represents a DMC of $225 (2007$) which becomes $234 (2010$) for this analysis, applicable
in the 2012MY. NHTSA continues to consider a 6 speed manual transmission to be on the
flat portion of the learning curve and has applied a low complexity ICM of 1.24 through 2018
then 1.19 thereafter. NHTSA's resultant costs for a 6 speed manual transmission are shown in
Table 3-41.
Table 3-41 Costs for 6 Speed Manual Transmission (2010$)
Cost
type
DMC
1C
TC
Transmission type
6sp manual
6sp manual
6sp manual
2017
$204
$57
$260
2018
$199
$57
$256
2019
$196
$45
$240
2020
$192
$44
$236
2021
$188
$44
$232
2022
$184
$44
$229
2023
$181
$44
$225
2024
$177
$44
$221
2025
$173
$44
$218
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost; sp=speed; dry=dry clutch; wet=wet-clutch
Note that all costs are relative to a 5 speed manual transmission.
3-111
-------
Technologies Considered in the Agencies' Analysis
3.4.3 Vehicle electrification and hybrid electric vehicle technologies
For the costs presented in this electrification and hybrid vehicle section, we have
estimated costs for vehicle classes since the technologies are closely linked to the size of the
vehicle as opposed to the number of cylinders on the engine or its valvetrain configuration.
The vehicle classes for which we have estimated costs are consistent with the six vehicle
classes developed for the lumped parameter model. Each agency has used the vehicle class
specific costs and mapped those into their respective model-specific vehicle classes or types
as shown in Table 3-42. This table simply presents the mapping of lumped parameter model
vehicle classes (or cost vehicle classes) into model-specific vehicle classes (or vehicle types in
the case of EPA's OMEGA model, please refer to Chapter 1 of EPA's final RIA for more
details) to help the reader understand how the vehicle classes used for costing relate to the
vehicle classes used for modeling. Note that there have been changes in the EPA data since
the proposal. EPA now characterizes cost vehicle classes more consistently with the way they
are classified in the lumped parameter model to avoid any confusion that the proposed cost
vehicle classes may have generated. EPA has also reconfigured its 19 vehicle types in an
effort to more closely align the vehicle types with the actual vehicles contained in each. Both
of these changes are detailed in Chapter 1 of EPA's final RIA.
Table 3-42 Mapping of Vehicle Class into each Agency's Model-Specific Vehicle Classes or Types
EPA Vehicle
Class for Cost
Purpose
Subcompact/
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Lumped
Parameter
Classification
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
Example
Fiesta
Focus
Yaris
Fusion
Taurus
Camry
Crown
Victoria
Mustang
Escape
Rav4
Tacoma
Edge
Explorer
4Runner
Sienna
F150
Tundra
OMEGA Model
Vehicle Type*
1
2,3,4
5,6
7,13
8, 9, 10, 14, 15
11, 12, 16, 17, 18, 19
NHTSA/CAFE
Model
Classification
Subcompact
Subcompact Perf PC
Compact
Compact Perf PC
Mid-size PC
Mid-size Perf PC
Large PC
Large Perf PC
Small LT
Midsize LT
Minivan LT
Large LT
* OMEGA uses 19 vehicle types as shown here and described in detail in Chapter 1 of EPA's final RIA.
3-112
-------
Technologies Considered in the Agencies' Analysis
3.4.3.1 Electrical Power Steering (EPS) / Electrohydraulic Power Steering (EHPS)
Electric power steering (EPS) and Electrohydraulic power steering (EHPS) provide a
potential reduction in CC>2 emissions and fuel consumption over hydraulic power steering
because of reduced overall accessory loads. This 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. EPS is an enabler for all vehicle hybridization technologies since it provides power
steering when the engine is off. EPS may be implemented on most vehicles with a standard
12V system. Some heavier vehicles may require a higher voltage system or EHPS which may
add cost and complexity.
The 2012-2016 final rule, EPA and NHTSA estimated a 1 to 2 percent effectiveness
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. NHTSA and EPA
reviewed these effectiveness estimates and found them to be accurate, thus they have been
retained for this final rule, consistent with the proposal. For large pickup truck the agencies
used EHPS due to the utility requirement of these vehicles. The effectiveness of EHPS is
estimated to be 0.8 percent.
In the MY 2012-2016 final rule, the agencies estimated the DMC at $88 (2007$).
Converting to 2010$, this DMC becomes $92 for this analysis, consistent with the recent
heavy-duty GHG rule, which is considered applicable in the 2015MY. The agencies use the
same DMC for EPS as for EHPS. Technically, EHPS is less costly than EPS. However, we
believe that EHPS is likely to be used, if at all, on the largest trucks and utility vehicles. As
such, it would probably need to be heavier-duty than typical EPS systems and the agencies
consider the net effect to place EHPS on par with EPS in terms of costs. The agencies
consider EPS/EHPS technology to be on the flat portion of the learning curve and have
applied a low complexity ICM of 1.24 through 2018 then 1.19 thereafter. The resultant costs
are shown in Table 3-43.
Table 3-43 Costs of Electrical/Electro-hydraulic Power Steering (2010$)
Cost type
DMC
1C
TC
2017
$87
$22
$109
2018
$86
$22
$108
2019
$84
$18
$101
2020
$82
$18
$100
2021
$80
$18
$98
2022
$79
$18
$96
2023
$77
$18
$95
2024
$76
$18
$93
2025
$74
$18
$92
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3.4.3.2 Improved Accessories
The accessories on an engine, including the alternator, coolant and oil pumps are
traditionally mechanically-driven. A reduction in CC>2 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
3-113
-------
Technologies Considered in the Agencies' Analysis
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.
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 2012-2016 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.
For this final rule, consistent with the proposal, the agencies considered two levels of
improved accessories. For level one of this technology (IACC1) NHTSA now 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 2012-2016 final rule and TAR for level
IACC1. More precisely, the agencies used effectiveness value in 1.2 to 1.8 percent range
varying based on different vehicle subclasses. The incremental effectiveness for this
technology in relative to EPS in the CAFE model is 0.91 to 1.61 percent. The combined
effectiveness for IACC1 and IACC2 ranges from 3.1 to 3.9 percent and NHTSA applied
incremental effectiveness of IACC2 in relative to IACC1 ranging from 1.74 to 2.55 percent.
In the 2012-2016 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
3-114
-------
Technologies Considered in the Agencies' Analysis
complexity ICM of 1.24 through 2018 then 1.19 thereafter. The resultant costs are shown in
Table 3-44.
Table 3-44 Costs for Improved Accessory Technology - Levels 1 & 2 (2010$)
Cost type
DMC
DMC
1C
1C
TC
TC
IACC
Technology
IACC1
IACC2
IACC1
IACC2
IACC1
IACC2
2017
$71
$114
$18
$29
$89
$143
2018
$70
$112
$18
$29
$88
$141
2019
$68
$110
$14
$23
$82
$133
2020
$67
$107
$14
$23
$81
$131
2021
$65
$105
$14
$23
$80
$128
2022
$64
$103
$14
$23
$78
$126
2023
$63
$101
$14
$23
$77
$124
2024
$62
$99
$14
$23
$76
$122
2025
$60
$97
$14
$23
$75
$120
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Note that both levels of IACC technology are incremental to the baseline case.
3.4.3.3 Air Conditioner Systems
We have a detailed description of the A/C program in Chapter 5 of this joint TSD.
The reader is directed to that chapter to learn the specifics of the program, the credits
involved, and details behind the costs we have estimated. Table 3-45 is a copy of Table 5-18
showing the total costs for A/C controls used in this final rule.
Table 3-45 Total Costs for A/C Control Used in This Final rule (2010$)
Car/
Truck
Car
Truck
Fleet
Cost type
TC
TC
TC
TC
TC
TC
TC
Rule
Reference
Control
Both
Reference
Control
Both
Both
2017
$76
$25
$101
$58
$2
$60
$86
2018
$75
$40
$115
$57
$46
$103
$111
2019
$70
$57
$127
$54
$73
$127
$127
2020
$69
$65
$134
$53
$82
$134
$134
2021
$68
$79
$147
$52
$95
$147
$147
2022
$67
$77
$144
$51
$93
$144
$144
2023
$66
$72
$138
$50
$88
$138
$138
2024
$65
$71
$135
$49
$86
$135
$135
2025
$64
$69
$133
$49
$84
$133
$133
TC=Total cost
3.4.3.4 Stop-start (12V Micro Hybrid)
The stop-start technology we consider for this final rule, consistent with the
proposal—also known as idle-stop or 12-volt micro-hybrid—is the most basic hybrid system
that facilitates idle-stop capability. When vehicle comes to a stop, the system will
automatically shut down the internal combustion engine and restarts the engine when vehicle
starts to move again. This is especially beneficial to reduce emission and fuel consumption
when vehicle spends significant amount of time stopping in inner city driving or a traffic jam.
Along with other enablers, this system typically replaces the standard 12-volt starter with an
improved unit capable of higher power and increased cycle life. These systems typically
incorporate an improved battery to prevent voltage-droop on restart. Different from MY
2012-2016 rule, this technology is applied to all vehicle classes, including large pickup truck.
3-115
-------
Technologies Considered in the Agencies' Analysis
In MYs 2012-2016 final rule, even though EPA did not use 12 volt stop-start technology,
NHTSA and EPA jointly reviewed the assumption. The effectiveness NHTSA used in the
CAFE model for MYs 2012-2016 final rule ranged from 2 to 4 percent, depending on whether
the vehicle is equipped with a 4-, 6- or 8-cylinder engine, with the 4-cylinder engine having
the lowest range and the 8-cylinder having the highest. In this FRM analysis, consistent with
the proposal, when combining IACC1, IACC2 and 12V stop-start system, the estimated
effectiveness based on 2010 Ricardo study ranges from 4.8 percent to 5.9 percent. The
agencies applied this effectiveness in the FRM analysis, consistent with the proposal. For
CAFE modeling, the incremental effectiveness for 12V stop-start relative to IACC2 is 1.68 to
2.2 percent. Importantly, the effectiveness values presented here represent two-cycle
effectiveness. Because stop-start technology provides considerable off-cycle benefits, both
agencies apply a credit value to the technology. Off-cycle credits are discussed in Chapter 5
of this Joint TSD.
In the 2012-2016 rule, the agencies estimated the DMC at $282 (2007$) to $350
(2007$) for small cars through large trucks, respectively. Converting to 2010$, these DMCs
become $295 (2010$) through $367 (2010$) for this analysis which are considered applicable
in the 2015MY. The agencies consider 12V stop-start technology to be on the steep portion
of the learning curve in the 2012-2016 timeframe and flat thereafter and have applied a
medium complexity ICM of 1.39 through 2018 then 1.29 thereafter. The resultant costs are
shown in Table 3-46.
Table 3-46 EPA and NHTSA Costs for 12V Micro Hybrid or 12V Stop-Start (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
Vehicle
Class
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
2017
$287
$287
$325
$1 1 C
325
$325
$356
$114
$114
$129
$129
$129
$142
$401
$401
$454
$454
$454
$498
2018
$278
$278
$315
$315
$315
$346
$114
$114
$129
$129
$129
$141
$392
$392
$444
$444
$444
$487
2019
$270
$270
$306
$306
$306
$335
$85
$85
$96
$96
$96
$105
$O e A
354
$354
$402
$402
$402
$441
2020
$261
$261
$296
$296
$296
$325
$85
$85
$96
$96
$96
$105
$346
$346
$392
$392
$392
$430
2021
$254
$254
$288
$288
$288
$315
$84
$84
$96
$96
$96
$105
$338
$338
$383
$383
$383
$420
2022
$246
$246
$279
$279
$279
$306
$84
$84
$95
$95
$95
$105
$330
$o o f\
330
$O T /I
374
$1 *7 A
374
$374
$410
2023
$239
$239
$271
$271
$271
$297
$84
$84
$95
$95
$95
$104
$322
$322
$366
$366
$366
$401
2024
$232
$232
$262
$262
$262
$288
$84
$84
$95
$95
$95
$104
$315
$315
$357
$357
$357
$392
2025
$225
$225
$255
$255
$255
$279
$83
$83
$94
$94
$94
$104
$308
$308
$349
$349
$349
$OOO
JoJ
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3-116
-------
Technologies Considered in the Agencies' Analysis
3.4.3.5 Mild Hybrid
Mild hybrid systems, also called Higher Voltage Stop-Start and Belt Mounted
Integrated Starter Generator (BISG) systems are similar to a micro-hybrid system, offering
idle-stop functionality, except that they utilize larger electric machine and a higher capacity
battery, typically 42 volts or above, thus enabling a limited level of regenerative braking
unavailable for a MHEV. The larger electric machine and battery also enables a limited
degree of power assist, which MHEV cannot provide. However, because of the limited torque
capacity of the belt-driven design, these systems have a smaller electric machine, and thus
less capability than crank-integrated or stronger hybrid systems. These systems replace the
conventional alternator with a belt-driven starter/alternator and may add high voltage
electrical accessories (which may include electric power steering and an auxiliary automatic
transmission pump). The limited electrical requirements of these systems allow the use of
lead-acid batteries or supercapacitors for energy storage, or the use of a small lithium-ion
battery pack, as is modeled in this analysis. While the mild hybrid system was not applied in
the NPRM analysis because the agencies did not have solid information regarding its likely
architecture, effectiveness or cost, the agencies are including the technology in the final rule
because we now have good information about it. Further, the agencies are making available
credits for mild hybrid pickup trucks in an effort to encourage such technologies. Lastly, the
simulation modeling and cost estimation results show that the mild hybrid system could be a
cost effective technology.
For the BISG technology the agencies sized the system using a 15 kW
starter/generator and 0.25 kWh Li-ion battery pack, which is similar to General Motors'
eAssist BISG, which is available in MY 2012 Buick LaCrosse, Buick Regal, and Chevrolet
Malibu vehicles. The agencies made this size system available to all vehicle subclasses,
believing that manufacturers might use a similar strategy to control component complexity
across the subclasses. As mentioned above, estimates were developed by ANL using
Autonomie full vehicle simulation software. The absolute effectiveness for the CAFE analysis
ranged from 8.5 to 11.6 percent depending on vehicle subclass. The effectiveness values
include technologies that would be expected to incorporated with BISG which are stop/start
(MHEV) and improved accessories (IACC1 and IACC2), however the effectiveness values do
not include electric power steering (EPS).
The costs for the mild hybrid technology are all new for this final rule and were
developed in a manner consistent with costs generated for strong hybrids. These costs are
presented in sections 3.4.3.7 through 3.4.3.10 of this Joint TSD. The same cost and
effectiveness results were applied by both NHTSA and EPA.
3.4.3.5.1 Integrated Motor Assist (IMA)/Crank Integrated Starter Generator (CISC)
CO
IMA is a system developed and marketed by Honda and is similar to CISG. They
both utilize a thin axial electric motor bolted to the engine's crankshaft and connected to the
transmission through a torque converter or clutch. The axial motor is motor/generator that
typically operates above 100 volts (but lower than the stronger hybrid systems discussed
below, which typically operate at around 300 volts) and can provide sufficient torque for
launch as well as generate sufficient current to provide significant levels of brake energy
3-117
-------
Technologies Considered in the Agencies' Analysis
recovery. The motor/generator also acts as the starter for the engine and can replace a typical
accessory-driven alternator. Current EVIA/CISG systems typically do not launch the vehicle
on electric power alone, although some commercially available systems can cruise on electric
power and dual-clutch IMA and CISG could be applied to all classes of vehicles. This
technology is not used as an enabling technology in this FRM analysis, consistent with the
proposal, by either EPA or NHTSA due to our expectation that manufacturers will be moving
to more cost effective technologies.
EPA relied on a combination of certification data (comparing vehicles available with
and without a hybrid system and backing out other components where appropriate) and
manufacturer-supplied information to determine that the effectiveness of these systems in
terms of CC>2 reduction is 30 percent for small cars, 25 percent for large cars, and 20 percent
for minivans and small trucks similar to the range estimated by NHTSA for the respective
vehicle classes. The effectiveness for small cars assumes engine downsizing to maintain
approximately equivalent performance. The large car, minivan, and small truck effectiveness
values assume less engine downsizing in order to improve vehicle performance and/or
maintain towing and hauling performance.
In the 2012-2016 final rule, the agencies estimated the DMC at $1,973, $2,497,
$2,508, $2,366 and $3,063 (all values in 2007$) for a small car, large car, minivan, small
truck and large truck, respectively. For this final rule, the DMCs are $2,070, $2,620, $2,631
and $3,214 (all values in 2010$) for small car/standard car, large car, small MPV and large
MPV/truck. All of these DMCs are considered applicable in the 2015MY. The agencies
consider the IMA technology to be on the steep portion of the learning curve and have applied
a highl complexity ICM of 1.56 through 2018 then 1.35 thereafter. The resultant costs are as
shown in Table 3-47. As noted earlier, the IMA technology is not included as an enabling
technology in this analysis, although it is included as a baseline technology because it exists
in the baseline fleet. The agencies moved away from this technology and applied P2 hybrids
instead because P2 is more cost effective than IMA.
Table 3-47 Costs for IMA Hybrids (2010$)
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
Vehicle Class
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
2017
$2,008
$2,541
$2,552
$3,118
$1,162
$1,471
$1,478
$1,805
$3,170
$4,013
$4,029
$4,923
2018
$1,947
$2,465
$2,475
$3,024
$1,159
$1,467
$1,473
$1,799
$3,106
$3,932
$3,948
$4,823
2019
$1,889
$2,391
$2,401
$2,933
$709
$898
$901
$1,101
$2,598
$3,289
$3,302
$4,034
2020
$1,832
$2,319
$2,329
$2,845
$707
$895
$899
$1,098
$2,540
$3,215
$3,228
$3,944
2021
$1,777
$2,250
$2,259
$2,760
$706
$893
$897
$1,096
$2,483
$3,143
$3,156
$3,856
2022
$1,724
$2,182
$2,191
$2,677
$704
$891
$895
$1,093
$2,428
$3,073
$3,086
$3,770
2023
$1,672
$2,117
$2,126
$2,597
$702
$889
$893
$1,090
$2,375
$3,006
$3,018
$3,687
2024
$1,622
$2,053
$2,062
$2,519
$701
$887
$891
$1,088
$2,323
$2,940
$2,952
$3,607
2025
$1,573
$1,992
$2,000
$2,443
$699
$885
$889
$1,086
$2,273
$2,877
$2,889
$3,529
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3-118
-------
Technologies Considered in the Agencies' Analysis
3.4.3.6 HEV, PHEV, EV and Fuel Cell Vehicle Technologies
A hybrid vehicle is a vehicle that combines two or more sources of propulsion energy,
where one uses a consumable fuel (like gasoline), and one is rechargeable (during operation,
or by another energy source). Hybrid technology is well established in the U.S. market and
more manufacturers are adding hybrid models to their lineups. Hybrids reduce fuel
consumption through three major mechanisms:
• The internal combustion engine can be optimized (through downsizing, modifying
the operating cycle, or other control techniques) to operate at or near its most
efficient point more of the time. Power loss from engine downsizing can be
mitigated by employing power assist from the secondary power source.
• Some of the energy normally lost as heat while braking can be captured and stored
in the energy storage system for later use.
• The engine is turned off when it is not needed, such as when the vehicle is coasting
or when stopped.
Hybrid vehicles utilize some combination of the three above mechanisms to reduce
fuel consumption and CC>2 emissions. A fourth mechanism to reduce petroleum fuel
consumption, available only to plug-in hybrids, is by substituting the petroleum fuel energy
with energy from another source, such as the electric grid. The effectiveness of fuel
consumption and CC>2 reduction depends on the utilization of the above mechanisms and how
aggressively they are pursued. One area where this variation is particularly prevalent is in the
choice of engine size and its effect on balancing fuel economy and performance. Some
manufacturers choose not to downsize the engine when applying hybrid technologies. In
these cases, performance is vastly improved, while fuel efficiency improves significantly less
than if the engine was downsized to maintain the same performance as the conventional
version. While this approach has been used in cars such as the Lexus 600h luxury vehicle, it
is more likely to be used in the future for vehicles like trucks where towing and/or hauling are
an integral part of their performance requirements. In these cases, if the engine is downsized,
the battery can be quickly drained during a long hill climb with a heavy load, leaving only a
downsized engine to carry the entire load. Because towing capability is currently a heavily-
marketed truck attribute, manufacturers are hesitant to offer a truck with downsized engine
which can lead to a significantly diminished towing performance when the battery state of
charge level is low, and therefore engines are traditionally not downsized for these vehicles.
Although hybrid vehicles using other energy storage concepts (flywheel, hydraulic)
have been developed, the automotive systems in production for passenger cars and light
trucks are all hybrid electric vehicles (HEV) that use battery storage and electric drive
systems. This appears likely to be the case for the foreseeable future. HEVs are part of a
continuum of vehicles using systems with differing levels of electric drive and electric energy
storage. This range of vehicles includes relatively basic system without electric energy
storage such as engine start/stop systems; HEV systems with varying degrees of electric
storage and electric drive system capability including mild-hybrid electric vehicles (MHEV)
3-119
-------
Technologies Considered in the Agencies' Analysis
with limited capability but lower cost; strong hybrid electric vehicles (SHEV) with full
hybridization capability such as the P2 hybrid technology which the agencies evaluate as a
compliance option in this FRM; plug-in hybrid electric vehicles (PHEV) with differing
degrees of all electric range and battery electric vehicles (EV) that rely entirely on electric
drive and battery electric energy storage.
Different HEV, PHEV and EV concepts utilize these mechanisms differently, so they
are treated separately for the purposes of this analysis. In many applications, particularly with
PHEV and EV, the battery represents the most costly and system-limiting sub-component of
the hybrid system. Currently, there are many battery chemistries being developed and refined
for hybrid applications that are expected to enhance the performance of future hybrid
vehicles. Section 3.4.3.6.4 contains a discussion of battery energy storage and the major
hybrid concepts that were determined to be available during the MY 2017-2015 timeframe.
Fuel cell vehicles are a separate category of electric vehicle that rely entirely on
electric propulsion with electricity produced on-board the vehicle using a proton-exchange-
membrane fuel cell (PEMFC) fueled with hydrogen. Fuel cell vehicles under development
are typically configured as a hybrid with battery storage used to provide brake energy
recovery and improved response to fast transients in vehicle energy demand.
3.4.3.6.1 Power-split hybrid
Power-split hybrid (PSHEV) - a hybrid electric drive system that replaces the
traditional transmission with a single planetary gear set and two motor/generators. The
smaller motor/generator uses the engine to either charge the battery or to 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. The planetary gear
splits engine power between the first motor/generator and the drive motor to either charge the
battery or supply power to the wheels. Power-split hybrids are not used as an enabling
technology in this final rule, consistent with the proposal.
In MYs 2012-2016 final rule, EPA and NHTSA used a combination of manufacturer-
supplied information and a comparison of vehicles available with and without a hybrid system
from EPA's fuel economy test data to determine that the effectiveness is 19 to 36 percent for
the classes to which it is applied. The estimate would depend on whether engine downsizing
is also assumed. In the CAFE incremental model, the range of effectiveness used was 23 to
33 percent as engine downsizing is not assumed (and accounted for elsewhere).
For this analysis, in order to estimate baseline costs, the agencies are using power-split
HEV costs generated by FEV as part of a tear-down study. In that study, FEV found the
DMC of the entire power-split system (battery-pack and non-battery components) to be
$2,853 (2007$), $3,175 (2007$), $3,435 (2007$), $4,168 (2007$) for vehicle sized, for
example, like a Ford Fiesta, Ford Focus, Ford Fusion and Ford Flex, respectively. For this
analysis, these values become $2,967, $3,302, $3,572 and $4,335, respectively, all in 2010
dollars. In the 2012-2016 final rule, the agencies estimated the DMC of a large truck power-
split system at $5,137 (2007$) which becomes $5,391 for this analysis (2010$) and we are
using this value for the large MPV vehicle class. All of these DMCs are considered
3-120
-------
Technologies Considered in the Agencies' Analysis
applicable in the 2015MY. The agencies consider the power-split technology to be on the flat
portion of the learning curve and have applied a highl complexity ICM of 1.56 through 2018
then 1.35 thereafter. The resultant costs are as shown in Table 3-48. As noted earlier, the
Power-split technology is not included as an enabling technology in this analysis, although it
is included as a baseline technology because it exists in the baseline fleet.
Table 3-48 Costs for Power-Split Hybrids (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
Vehicle Class
Small car
Standard car
Large car
Small MPV
Large MP
Small car
Standard car
Large car
Small MPV
Large MP
Small car
Standard car
Large car
Small MPV
Large MP
2017
$2,820
$3,139
$3,396
$4,120
$5,125
$1,663
$1,851
$2,002
$2,429
$3,021
$4,483
$4,990
$5,398
$6,549
$8,146
2018
$2,764
$3,076
$3,328
$4,038
$5,023
$1,659
$1,846
$1,998
$2,424
$3,015
$4,423
$4,923
$5,326
$6,462
$8,037
2019
$2,709
$3,015
$3,261
$3,957
$4,922
$1,017
$1,131
$1,224
$1,485
$1,847
$3,725
$4,146
$4,485
$5,442
$6,769
2020
$2,655
$2,954
$3,196
$3,878
$4,824
$1,015
$1,129
$1,222
$1,483
$1,844
$3,669
$4,084
$4,418
$5,361
$6,668
2021
$2,602
$2,895
$3,132
$3,801
$4,727
$1,013
$1,128
$1,220
$1,480
$1,841
$3,615
$4,023
$4,352
$5,281
$6,568
2022
$2,549
$2,837
$3,070
$3,725
$4,633
$1,012
$1,126
$1,218
$1,478
$1,838
$3,561
$3,963
$4,288
$5,202
$6,471
2023
$2,498
$2,781
$3,008
$3,650
$4,540
$1,010
$1,124
$1,216
$1,475
$1,835
$3,508
$3,905
$4,224
$5,125
$6,375
2024
$2,449
$2,725
$2,948
$3,577
$4,449
$1,008
$1,122
$1,214
$1,473
$1,832
$3,457
$3,847
$4,162
$5,050
$6,281
2025
$2,400
$2,671
$2,889
$3,505
$4,360
$1,007
$1,120
$1,212
$1,471
$1,829
$3,406
$3,791
$4,101
$4,976
$6,190
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3.4.3.6.2 2-mode hybrid
2-mode hybrid (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 CO2 emissions at
highway speeds relative to other types of hybrid electric drive systems. 2-mode hybrids were
not been considered in the proposal and the agencies sought comments on whether or not 2-
mode hybrids should be considered for vehicles with towing requirements, such as pickup
trucks. However, no comments were received on their applicability in the future and thus
consistent with the proposal, 2-mode hybrids were not included in the final rule analysis.
For MYs 2012-2016 final rule, the CAFE model considered a range of 23 to 33
percent with a midpoint of 28 percent, assuming no engine downsizing to preserve the utility
nature of medium and large trucks (e.g., maintaining full towing capability even in situations
with low battery charge) and EPA estimates CC>2 emissions reduction effectiveness to be 25
percent for large trucks (LDT3 and LDT4 categories) based on vehicle certification data.
EPA estimates an effectiveness of 40 percent for smaller vehicles.
The agencies have estimated the costs for 2-mode hybrids using costs used in the 2010
TAR. For this analysis, the 2-mode battery pack DMC is estimated at $1,100 (2010$) and the
3-121
-------
Technologies Considered in the Agencies' Analysis
DMC of non-battery components is estimated at $2,997 (2010$). The battery pack DMC is
considered to be applicable for the 2025MY while the non-battery pack DMC would be
applicable for the 2012MY. The agencies consider the 2-mode battery packs to be on the
steep portion of the learning curve during the 2017-2025 timeframe. The agencies have
applied a highl complexity ICM of 1.56 through 2018 then 1.35 thereafter. For 2-mode non-
battery components, the agencies consider them to be on the flat portion of the learning curve
in the 2017-2025 timeframe and have applied a highl complexity ICM of 1.56 through 2018
then 1.35 thereafter. The resultant 2-mode hybrid costs are presented in Table 3-49.
Table 3-49 Costs for 2-Mode Hybrids (2010$)
Cost
type
Vehicle Class
2017
2018
2019
2020
2021
2022
2023
2024
2025
Battery-pack
DMC
1C
TC
Small MPV/Large
MPV/Truck
$2,148
$688
$2,835
$1,718
$660
$2,378
$1,718
$399
$2,118
$1,374
$389
$1,763
$1,374
$389
$1,763
$1,374
$389
$1,763
$1,374
$389
$1,763
$1,374
$389
$1,763
$1,100
$380
$1,479
Non-battery pack components
DMC
1C
TC
Small MPV/Large
MPV/Truck
$2,600
$1,664
$4,264
$2,548
$1,660
$4,208
$2,497
$1,019
$3,517
$2,447
$1,018
$3,465
$2,398
$1,016
$3,415
$2,350
$1,015
$3,365
$2,303
$1,013
$3,317
$2,257
$1,012
$3,269
$2,212
$1,010
$3,222
Battery -pack and non-battery pack components
TC
Small MPV/Large
MPV/Truck
$7,099
$6,586
$5,634
$5,228
$5,178
$5,128
$5,080
$5,032
$4,702
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3.4.3.6.3 P2 Hybrid
A P2 hybrid is hybrid technology that uses a transmission-integrated electric motor
placed between the engine and a gearbox or CVT and coupled to the engine crankshaft via a
clutch. The engine and the drive motor are mechanically independent of each other, allowing
the engine or motor to power the vehicle separately or combined. Disengaging the engine
clutch allows all-electric operation and more efficient brake-energy recovery. The P2 HEV
system is similar to the Honda IMA HEV architecture with the exception of the added clutch,
and larger batteries and motors. Examples of this include the Hyundai Sonata HEV and
Infmiti M35h. The agencies believe that the P2 is an example of a "strong" hybrid technology
that is typical of what will be prevalent in the timeframe of this rule. The agencies could have
equally chosen the power-split architecture as the representative HEV architecture. These two
HEV's have similar average effectiveness values (combined city and highway fuel economy),
though the P2 systems may have lower cost due to having only a single, smaller
motor/generator.
For purposes of this rulemaking analysis, the agencies are assuming that P2 hybrids
will become the dominant technology in the MYs 2017-2025 timeframe, replacing costlier
power-split or 2-mode architectures while providing substantially similar efficiency
improvement. At the present time, P2 hybrids are relatively new to the market and the
agencies have not attempted to quantify any measurable performance differential between
these technologies. As mentioned, the 2011 Hyundai Sonata, 2011 Volkswagen Touareg
3-122
-------
Technologies Considered in the Agencies' Analysis
Hybrid, the 2011 Porsche S Hybrid, and the 2012 Infiniti M35 Hybrid are examples of P2
hybrids currently in production and available to consumers. The agencies are aware of some
articles in trade journals, newspapers and other reviews that some first generation P2 hybrid
vehicles with automatic transmissions have trade-offs in NVH and drivability - though these
reviews do not cover all of the P2 systems available today, and a number of reviews are very
positive with respect to NVH and drivability. The agencies recognize that manufacturers will
have several years to test, develop and improve P2 technology in the years before 2017. We
expect that manufacturers will address any perceived integration issues in early production
models. However, we believe it is important to continue to monitor development of P2
hybrids and market acceptance of this technology. We will continue to gather information on
these issues and consider them as part of the mid-term evaluation.
The agencies requested comment regarding the potential of P2 hybrids to overcome
these issues or others and we specifically sought comment from automakers developing and
considering P2 technology on whether they believe these to be significant impediments to
deployment and how they may be addressed. There were no comments submitted.
The effectiveness used for vehicle packages with the P2-hybrid configuration within
this analysis reflects a conservative estimate of system performance. Vehicle simulation
modeling of technology packages using the P-2 hybrid has recently been completed under a
contract with Ricardo Engineering. The agencies have updated the effectiveness of hybrid
electric vehicle packages using the new Ricardo vehicle simulation modeling runs for this
analysis.
Due to the lower cost and comparative effectiveness of P2 hybrid in relative to other
strong hybrid technologies, such as power-split hybrid and 2-mode hybrid, the agencies
assume P2 hybrid application for all vehicle sub-classes in this FRM analysis, consistent with
the proposal, and increased HEV effectiveness by approximately 2% comparing to 2012-2016
light duty GHG/CAFE final rule based on published data for new HEVs that have entered into
production, such as 2011 Hyundai Sonata hybrid, 2010 Hyundai Elantra LPI HEV (Korean
market only), 2011 Infiniti G35 Hybrid and 2011 Volkswagen Touareg Hybrid). In addition,
for the Large Car, Minivan and Small Truck subclasses, the agencies further increased HEV
effectiveness by assuming that towing capacity could be reduced from their current rating^ to
approximately 1,500 pounds for some vehicles in these subclasses without significantly
impacting consumers' need for utility in these vehicles.ss The agencies believe that
consumers for these vehicles who require higher towing capacity could acquire it by
purchasing a vehicle with a more capable non-hybrid powertrain (as they do today).tt
" Current small SUVs and Minivans have an approximate average towing capacity of 2000 pounds (without a
towing package), but range from no towing capacity to 3500 pounds.
ss We note that there are some gasoline vehicles in the large car/minivan/small truck segments sold today which
do not have any towing rating.
tt The agencies recognize that assuming that certain consumers will choose to purchase non-hybrid vehicles in
order to obtain their desired towing capacity could lead to some increase in fuel consumption and CO2 emissions
as compared to assuming that towing capacity is maintained for hybrid vehicles across the board. However, the
3-123
-------
Technologies Considered in the Agencies' Analysis
Moreover, it is likely that some fraction of consumers who purchase the larger engine option
do so for purposes of hauling and acceleration performance, not just maximum towing.
A reduction in towing capacity allows greater engine downsizing, which increases
estimated overall HEV system incremental effectiveness by 5 to 10 percent for Large Cars,
Minivans, and Small Trucks, similar to the HEV effectiveness value assumed for Small Cars
and Compact Cars."11
Based on the recent Ricardo study, the effectiveness for P2 hybrid used in this FRM,
consistent with the proposal, is 46.2 percent for subcompact and compact passenger cars, 48.6
percent for midsize passenger car, 49.4 percent for large passenger car, 46.1 percent for small
light truck, 45.7 percent for midsize SUV, truck and minivan and 45.1 percent for large
pickup truck.
The process for battery sizing for the P2 hybrids is explained in Section 3.4.3.8. The
battery sizing is different for the 2008 and 2010 baseline vehicle fleets, because vehicle mass
for each subclass is slightly different between the two baseline fleets, thus requiring a slightly
different battery size to maintain equivalent performance. The battery sizes with no applied
mass reduction are listed in Table 3-50.
Table 3-50 NHTSA Battery Sizes for P2 Hybrid Applied in Volpe Model without Mass Reduction (kWh)
Baseline
Fleet
2008
2010
Subcompact
PC/PerfPC
Compact PC/
PerfPC
0.81
0.84
Midsize
PC/PerfPC
1.00
1.02
Large PC/Perf
PC
1.16
1.20
Midsize LT
Minivan
1.28
1.27
Small LT
1.04
1.06
Large
LT
1.49
1.56
The agencies have applied a high complexity ICM to both the battery and non-battery
component costs for P2 hybrid. But for battery for P2 hybrid, the ICM switches from short
term value of 1.56 to long term value of 1.35 at 2024 while for the non-battery component the
switch happens at 2018.
The costs for P2 hybrids without mass reduction as used in the Volpe model are listed
in Table 3-51. The battery costs are calculated using the battery sizes for both the 2008 and
2010 baseline fleets. NHTSA accounts the cost impact from the interaction between mass
reduction and sizing of the electrification system (battery and non-battery system) as a cost
agencies think it likely that the net improvement in fuel consumption and CO2 emissions due to the increased
numbers of hybrids available for consumers to choose will offset any potential increase in fuel consumption and
CO2 emissions resulting from consumers selecting the higher-performance non-hybrid powertrain vehicles.
uu The effectiveness of HEVs for heavier vehicles which require conventional towing capabilities is markedly
less because the rated power of the 1C engine must be similar to its non-hybrid brethren. As such, there is less
opportunity for downsizing with these vehicles.
3-124
-------
Technologies Considered in the Agencies' Analysis
synergy as described in section 3.4.3.9. Estimated costs for P2 HEVs with mass reduction as
used in the OMEGA model are presented in Sections 3.4.3.9 and 3.4.3.10 below.
Table 3-51 NHTSA Costs for P2 Hybrid Applied in Volpe Model without Mass Reduction (2010$)
Tech.
Battery
Battery
Battery
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Non-
battery
Non-
battery
Non-
battery
Battery
Battery
Cost
Type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
NHTSA
Vehicle
Class
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Midsize LT
Minivan
Small LT
Large LT
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Midsize LT
Minivan
Small LT
Large LT
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
Baseline
Fleet
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2017
$733
$726
$818
$809
$959
$946
$887
$885
$796
$787
$1,029
$1,020
$1,474
$1,468
$1,645
$1,627
$1,949
$1,906
$1,817
$1,798
$1,587
$1,557
$1,918
$1,901
$413
$409
$461
2018
$711
$704
$793
$784
$931
$918
$860
$858
$773
$763
$998
$989
$1,445
$1,438
$1,612
$1,595
$1,910
$1,868
$1,780
$1,762
$1,555
$1,526
$1,879
$1,863
$411
$408
$459
2019
$689
$683
$769
$761
$903
$890
$834
$832
$749
$740
$968
$960
$1,416
$1,410
$1,580
$1,563
$1,872
$1,830
$1,745
$1,727
$1,524
$1,496
$1,842
$1,825
$410
$406
$458
2020
$669
$662
$746
$738
$876
$864
$809
$807
$727
$718
$939
$931
$1,388
$1,381
$1,549
$1,531
$1,834
$1,794
$1,710
$1,693
$1,493
$1,466
$1,805
$1,789
$409
$405
$456
2021
$648
$642
$724
$716
$849
$838
$785
$783
$705
$697
$911
$903
$1,360
$1,354
$1,518
$1,501
$1,798
$1,758
$1,676
$1,659
$1,464
$1,436
$1,769
$1,753
$407
$404
$455
2022
$629
$623
$702
$694
$824
$813
$761
$760
$684
$676
$884
$876
$1,333
$1,327
$1,487
$1,471
$1,762
$1,723
$1,642
$1,626
$1,434
$1,408
$1,733
$1,718
$406
$402
$453
2023
$610
$604
$681
$674
$799
$788
$739
$737
$663
$655
$857
$850
$1,306
$1,300
$1,457
$1,441
$1,727
$1,688
$1,609
$1,593
$1,406
$1,380
$1,699
$1,684
$405
$401
$452
2024
$592
$586
$661
$653
$775
$765
$716
$715
$643
$636
$831
$824
$1,280
$1,274
$1,428
$1,413
$1,692
$1,655
$1,577
$1,561
$1,378
$1,352
$1,665
$1,650
$404
$400
$451
2025
$574
$569
$641
$634
$752
$742
$695
$693
$624
$617
$807
$799
$1,254
$1,249
$1,400
$1,384
$1,658
$1,621
$1,546
$1,530
$1,350
$1,325
$1,631
$1,617
$248
$246
$277
3-125
-------
Technologies Considered in the Agencies' Analysis
Battery
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Non-
battery
Non-
battery
Non-
battery
Battery
Battery
Battery
Battery
Battery
Battery
Non-
battery
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
PC/PerfPC
Large
PC/PerfPC
Midsize LT
Minivan
Small LT
Large LT
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Midsize LT
Minivan
Small LT
Large LT
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Midsize LT
Minivan
Small LT
Large LT
Subcompact
PC/PerfPC
Compact
PC/PerfPC
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
$456
$541
$533
$500
$499
$449
$443
$580
$575
$943
$939
$1,053
$1,041
$1,247
$1,219
$1,162
$1,150
$1,015
$996
$1,227
$1,216
$1,145
$1,135
$1,278
$1,264
$1,500
$1,480
$1,386
$1,383
$1,245
$1,230
$1,609
$1,595
$2,418
$2,407
$454
$539
$532
$498
$497
$447
$442
$578
$573
$941
$937
$1,050
$1,039
$1,244
$1,217
$1,160
$1,148
$1,013
$994
$1,224
$1,213
$1,122
$1,111
$1,252
$1,239
$1,469
$1,449
$1,358
$1,355
$1,220
$1,205
$1,576
$1,562
$2,386
$2,375
$453
$537
$530
$496
$495
$446
$440
$576
$571
$578
$575
$645
$638
$764
$747
$712
$705
$622
$610
$752
$745
$1,099
$1,089
$1,227
$1,213
$1,440
$1,420
$1,331
$1,327
$1,195
$1,181
$1,544
$1,531
$1,994
$1,985
$451
$535
$528
$495
$494
$444
$439
$574
$569
$577
$574
$644
$637
$763
$746
$711
$704
$621
$610
$751
$744
$1,077
$1,067
$1,202
$1,189
$1,411
$1,392
$1,304
$1,301
$1,171
$1,157
$1,513
$1,500
$1,965
$1,956
$450
$534
$526
$493
$492
$443
$438
$572
$567
$576
$574
$643
$636
$762
$745
$710
$703
$620
$609
$749
$743
$1,056
$1,046
$1,179
$1,166
$1,383
$1,364
$1,278
$1,275
$1,148
$1,134
$1,483
$1,470
$1,936
$1,927
$448
$532
$525
$492
$490
$442
$436
$571
$566
$575
$573
$642
$635
$761
$744
$709
$702
$619
$608
$748
$742
$1,035
$1,025
$1,155
$1,143
$1,356
$1,337
$1,253
$1,250
$1,125
$1,112
$1,454
$1,442
$1,908
$1,899
$447
$530
$523
$490
$489
$440
$435
$569
$564
$574
$572
$641
$634
$759
$743
$708
$701
$618
$607
$747
$741
$1,015
$1,006
$1,133
$1,121
$1,330
$1,311
$1,229
$1,226
$1,104
$1,090
$1,426
$1,414
$1,881
$1,872
$446
$529
$522
$489
$488
$439
$434
$567
$562
$574
$571
$640
$633
$758
$741
$707
$700
$617
$606
$746
$739
$996
$986
$1,111
$1,099
$1,304
$1,286
$1,205
$1,202
$1,082
$1,069
$1,399
$1,386
$1,854
$1,845
$274
$325
$320
$300
$299
$270
$266
$348
$345
$573
$570
$639
$632
$757
$740
$706
$699
$616
$605
$745
$738
$822
$814
$918
$907
$1,077
$1,062
$995
$993
$894
$883
$1,155
$1,145
$1,827
$1,819
3-126
-------
Technologies Considered in the Agencies' Analysis
Non-
battery
Non-
battery
Non-
battery
Non-
battery
Non-
battery
TC
TC
TC
TC
TC
Midsize
PC/PerfPC
Large
PC/PerfPC
Midsize LT
Minivan
Small LT
Large LT
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
$2,698
$2,668
$3,196
$3,125
$2,979
$2,949
$2,602
$2,554
$3,144
$3,116
$2,663
$2,633
$3,155
$3,085
$2,940
$2,910
$2,568
$2,520
$3,103
$3,076
$2,225
$2,201
$2,636
$2,577
$2,457
$2,432
$2,146
$2,106
$2,593
$2,570
$2,192
$2,168
$2,597
$2,540
$2,421
$2,396
$2,115
$2,075
$2,555
$2,533
$2,160
$2,137
$2,559
$2,503
$2,386
$2,361
$2,084
$2,045
$2,518
$2,496
$2,129
$2,106
$2,522
$2,466
$2,351
$2,327
$2,054
$2,015
$2,482
$2,460
$2,098
$2,075
$2,486
$2,431
$2,317
$2,294
$2,024
$1,986
$2,446
$2,424
$2,068
$2,046
$2,450
$2,396
$2,284
$2,261
$1,995
$1,958
$2,411
$2,389
$2,039
$2,016
$2,415
$2,362
$2,252
$2,229
$1,966
$1,930
$2,376
$2,355
3.4.3.6.4 Plug-In Hybrid
Plug-In Hybrid Electric Vehicles (PHEVs) are very similar to Hybrid Electric
Vehicles, but with three significant functional differences. The first is the addition of a means
to charge the battery pack from an outside source of electricity (e.g., the electric grid).
Second, a PHEV would have a larger battery pack with more energy storage, and a greater
capability to be discharged. Finally, a PHEV would have a control system that allows the
battery pack to be significantly depleted during normal operation.
Table 3-52 below, illustrates how PHEVs compare functionally to both hybrid electric
vehicles (HEV) and electric vehicles (EV). These characteristics can change significantly
within each vehicle class/subclass, so this is simply meant as an illustration of the general
characteristics. In reality, the design options are so varied that all these vehicles exist on a
continuum with HEVs on one end and EVs on the other.
Table 3-52 Conventional, HEVs, PHEVs, and EVs Compared
Attribute
Drive Power
Engine Size
Electric Range
Battery Charging
Increasing Electrification
Conventional
Engine
Full Size
None
None
HEV
Blended
Engine/Electric
Full Size or Smaller
None to Very Short
On-Board
PHEV
Blended
Engine/Electric
Smaller or Much
Smaller
Short to Medium
Grid/On-Board
EV
Electric
No Engine
Medium to Long
Grid Only
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 course, depend on the amount of electric drive the vehicle is capable of 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
3-127
-------
Technologies Considered in the Agencies' Analysis
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.
In analyzing the impacts of grid-connected vehicles like PHEVs and EVs, the
emissions from the electrical generation can be accounted for if a full upstream and
downstream analysis is desired. While this issue is being studied on an on-going basis,
upstream CC>2 emissions are not unique to grid-connected technologies and so are not
included in this analysis. The respective agencies' RIAs and NHTSA's EIS have more
information on upstream emissions.
PHEVs will be considerably more costly than conventional vehicles and some other
advanced technologies due to the fact that PHEVs require both conventional internal
combustion engine and electrical driving system and the larger expensive battery pack. To
take full advantage of their capability, consumers would have to be willing to charge the
vehicles during electricity off-peak hours during the night, and would need access to electric
power where they park their vehicles. For many urban dwellers who may park on the street,
or in private or public lots or garages, charging may not be practical. Charging may be
possible at an owner's place of work, but that would increase grid loading during peak hours
which would eliminate some of the benefits to utilities of off-peak charging versus on-peak.
Oil savings will still be the same in this case assuming the vehicle can be charged fully.
The effectiveness potential of PHEVs depends on many factors, the most important
being the energy storage capacity designed into the battery pack. To estimate the fuel
consumption and tailpipe CC>2 reduction potential of PHEVs, EPA has developed an in-house
vehicle energy model (PEREGRIN) to estimate the fuel consumption/CO2 emissions
reductions of PHEVs. This model is based on the PERE (Physical Emission Rate Estimator)
physics-based model used as a fuel consumption input for EPA's MOVES mobile source
emissions model.
How EPA Estimates PHEV Effectiveness
The PHEV small car, large car, minivan and small trucks were modeled using
parameters from a midsize car similar to today's hybrids and scaled to each vehicle's weight.
The large truck PHEV was modeled separately assuming no engine downsizing. PHEVs can
have a wide variation in the All Electric Range (AER) that they offer. Some PHEVs are of
the "blended" type where the engine is on during most of the vehicle operation, but the
proportion of electric energy that is used to propel the vehicle is significantly higher than that
used in a PSHEV or 2MHEV. Each PHEV was modeled with enough battery capacity for a
20-mile-equivalent AER and a power requirement to provide similar performance to a hybrid
vehicle. 20 miles was selected because it offers a good compromise for vehicle performance,
3-128
-------
Technologies Considered in the Agencies' Analysis
weight, battery packaging and cost. Given expected near-term battery capability, a 20 mile
range represents the likely capability that will be seen in PHEVs in the near-to-mid term.
To calculate the total energy use of a PHEV, the PHEV can be thought of as operating
in two distinct modes, electric (EV) mode, and hybrid (HEV) mode. At the tailpipe, the CC>2
emissions during EV operation are zero. The EV mode fuel economy can then be combined
with the HEV mode fuel economy using the Utility Factor calculation in SAE J1711 to
determine a total MPG value for the vehicle. (See Table 3-53)
Table 3-53 Sample Calculation of PHEV Gasoline-Equivalent CO2 Reduction
EV energy comb (0.55 city / 0.45 hwy)
EV range (from PEREGRIN)
SAE Jl 711 utility factor
HEV mode comb FE (0.55 city / 0.45 hwy)
Total UF-adjusted FE (UF*FCEV + (1-UF)*FCHEV)
Baseline FE
Percent FE gain
Percent CO2 reduction
Midsize Car
0.252 kwh/mi
20 miles
0.30
49.1 mpg
70.1 mpg
29.3 mpg
139%
-58%
Large Truck
0.429 kwh/mi
20 miles
0.30
25. 6 mpg
36.6 mpg
19.2 mpg
90%
-47%
Calculating a total fuel consumption and tailpipe CO2 reduction based on model
outputs and the Utility Factor calculations results in a 58 percent reduction for small cars,
large cars, minivans, and small trucks. For large trucks, the result is a 47 percent reduction.
The lower improvement is due to less engine downsizing in the large truck class.
How NHTSA Estimates PHEV Effectiveness
For purposes of CAFE analysis, we assume that all future PHEVs during the
rulemaking timeframe will meet the range requirements to qualify as a dual fuel vehicle.
When calculating the fuel economy of a dual-fuel PHEV, NHTSA uses a petroleum
equivalency factor for electricity consumption as stated in 49 U.S.C. 32904 and 32905.
When deciding PHEV and EV effectiveness, NHTSA referenced the fuel economy of
3 pairs of vehicles for which NHTSA has fuel economy data in the CAFE database. These
three vehicles pairs are MiniE electric vehicle versus gasoline powered Mini with automatic
transmission, Tesla Roadster electric vehicle versus gasoline powered rear-wheel-drive Lotus
Elise Sedan with a 6-speed manual transmission, and Nissan Leaf electric vehicle versus
gasoline powered Nissan Sentra with automatic transmission. The fuel economy and fuel
consumption for the first two pairs are shown in Table 3-54. Nissan Leaf information is used
but not shown in the table because it is confidential information. Because technologies are
applied in the CAFE model in an incremental manner, the effectiveness for each technology is
incremental to the previous technology. In the electrification decision tree of the CAFE
model, the order of technology selection starts from gasoline only powertrain, then moves to
strong hybrid, to plug-in hybrid electric vehicle, and finally to electric vehicle. So the
incremental effectiveness for each step has to be defined.
3-129
-------
Technologies Considered in the Agencies' Analysis
Table 3-54 EV Fuel Economy and Fuel Consumption
104 mile range (Mini website)
MiniE (mpg)
Mini Gas ATX (mpg)
227 mile range (EPA)
Tesla Roadster
Lotus Elise sedan M6 RWD
Fuel economy (mpg)
342.4
38.6
346.8
30.6
Fuel consumption
(gPm)
0.0029206
0.0259067
0.0028835
0.0326797
In order to calculate the effectiveness of PHEV for purposes of a CAFE standard, fuel
economy for strong hybrid electric vehicle (SHEV) is calculated first using the incremental
effectiveness of strong hybrid from LPM model which is around 46 percent. For example, the
derived fuel economy for SHEV based on Mini Gas ATX is 71.7 mpg. Then the fuel economy
from gasoline source for PHEV is assumed to be the same as SHEV fuel economy, i.e. 71.7
mpg in the case of Mini E. The petroleum equivalent fuel economy from the electricity source
is set to be the same as the EV fuel economy, e.g., 342.4 mpg in the case of Mini E. The
combined fuel economy for PHEV is calculated using the 50-50 weighting factor as follows.
Consistent with 49 U.S.C. 32904 and 32905, NHTSA is using a 50-50 weighting
factor in the calculation above for CAFE model analysis of PHEV through 2019. After 2019,
NHTSA will use the utility factor method defined by SAE standard J1711 for calculating
CAFE fuel economy of PHEV. NHTSA expects that a PHEV with a 30 mile charge depleting
range may reasonably represent the PHEVs that manufacturers may produce in MYs 2017 to
2025. According to SAE standard J2841, a vehicle with 30 mile charge depleting range has a
0.668 city specific utility factor and a 0.337 highway specific utility factor, which together
give a 0.52 combined utility factor (55% city/45% highway split). Therefore NHTSA selected
a PHEV with a 30 mile range for the CAFE model analysis, and the selection of a PHEV with
a 30 mile range maintains continuity between pre-2020 and post-2020 PHEV fuel economy
calculations. NHTSA assumes a 0.50 utility factor for MY2020 and beyond. In the FRM
analysis, consistent with the proposal, EPA models a 20-mile range and a 40-mile range
PHEV.
The incremental fuel consumption reduction for PHEV is then calculated in relative to
strong HEV. Using the example of Mini E, the incremental fuel consumption reduction for
PHEV relative to SHEV is 39.5 percent as shown below.
3-130
-------
Technologies Considered in the Agencies' Analysis
Table 3-55 lists the incremental effectiveness calculation for two pairs of vehicles,
MiniE and Tesla Roaster. Incremental fuel consumption calculation for PHEV based on
Nissan Leaf is not shown in Table 3-55 due to confidentiality of the fuel economy rating. The
derived incremental effectiveness for Nissan Leaf is 40.6%. The average incremental
effectiveness of these three pairs of vehicles is 40.65 percent which is used in CAFE
modeling.
Table 3-55 Incremental Effectiveness Calculation for purposes of CAFE modeling
Mini E
Combined Fuel Economy [mpg]
Gasoline Fuel Economy [mpg]
Electric Petroleum Equivalent Fuel
Economy [mpg]
Combined Fuel Consumption[gpm]
Gasoline Fuel Consumption [gpm]
Incremental Combined Fuel Consumption
[%]
Gasoline Weighing Factor[%]
Electricity Weighing Factor [%]
Gasoline
38.6
SHEV2
71.7
71.7
0.0139414
0.0139414
PHEV1
118.6
71.7
342.4
0.0084310
0.0139414
39.5%
50%
50%
EV1
342.4
0.0029206
65.4%
0%
100%
Tesla
Combined Fuel Economy [mpg]
Gasoline Fuel Economy [mpg]
Electric Petroleum Equivalent Fuel
Economy [mpg]
Combined Fuel Consumption[gpm]
Gasoline Fuel Consumption [gpm]
Incremental Combined Fuel Consumption
[%]
Gasoline Weighing Factor[%]
Gasoline
30.6
SHEV2
56.7
56.7
0.017647
0.017647
PHEV1
97.4
56.7
346.8
0.0102653
0.0176471
41.8%
50%
EV1
346.8
0.0028835
71.9%
0%
3-131
-------
Technologies Considered in the Agencies' Analysis
Electricity Weighing Factor [%]
50%
100%
Once the fuel economy of the PHEV is calculated, the effectiveness of PHEV
incremental to EV can be calculated similarly using the formula below.
The average effectiveness for the three pairs of vehicles of 68.54% is used in CAFE
modeling.
The cost of PHEV consists of three parts, the cost for battery, the cost for non-battery
systems and the cost for charger and the labor to install it. The battery sizing is calculated as
in Section 3.4.3.8 and listed in Table 3-56. Costs for PHEVs without mass reduction as used
in the Volpe model are listed in Table 3-57 to Table 3-61. NHTSA accounts the cost impact
from the interaction between mass reduction and sizing of the electrification system (battery
and non-battery system) as a cost synergy as described in section 3.4.3.9. Sections 3.4.3.9
and 3.4.3.10 contain the cost for PHEVs with mass reduction as used in EPA's OMEGA
model. PHEV20 and PHEV40 are sized by EPA with the methodologies discussed in section
3.4.3.8.
Table 3-56 NHTSA Battery Sizes for PHEV30 Hybrid Applied in Volpe Model without Mass Reduction
(kWh)
Baseline
Fleet
2008
2010
Subcompact
PC/PerfPC
Compact PC/
PerfPC
10.42
10.81
Midsize
PC/PerfPC
12.82
13.13
Large PC/Perf
PC
15.21
15.79
Midsize LT
Minivan
17.09
16.94
Small LT
13.48
13.69
Large
LT
19.73
20.27
The battery pack DMCs for PHEV20 and PHEV40 are calculated using ANL's
BatPaC model. NHTSA modeled a PHEV 30 for this final rule, for which NHTSA averaged
the costs of PHEV20s and PHEV40s.
The agencies have applied a high complexity ICM to non-battery component cost for
PHEV and PHEV charger, which switch from short term value of 1.56 to long term value of
1.35 at 2018. The agencies applied a higher ICM factor to the battery of PHEV due to the fact
that it a more complex technology. The ICM for PHEV battery switches from short term
value of 1.77 to long term value of 1.50 at 2024.
Table 3-57 NHTSA Costs Applied in Volpe Model for PHEV30 with No Mass Reduction (2010$)
3-132
-------
Technologies Considered in the Agencies' Analysis
Tech.
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
Cost
Type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
DMC
DMC
DMC
1C
1C
NHTSA
Vehicle
Class
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
Baseline
Fleet
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
2017
$6,208
$6,095
$7,415
$7,251
$9,835
$9,610
$2,586
$2,522
$3,252
$3,132
$4,685
$4,494
$210
$1,010
$2,671
$2,622
$3,190
$3,119
$4,231
$4,134
$1,655
$1,614
$2,081
$2,004
$2,997
$2,875
$67
$0
2018
$8,259
$8,097
$5,932
$5,801
$7,868
$7,688
$2,535
$2,472
$3,187
$3,070
$4,591
$4,405
$168
$1,010
$2,579
$2,532
$3,081
$3,012
$4,086
$3,993
$1,651
$1,610
$2,077
$2,000
$2,991
$2,869
$65
$0
2019
$8,259
$8,097
$5,932
$5,801
$7,868
$7,688
$2,484
$2,422
$3,124
$3,008
$4,499
$4,316
$168
$1,010
$2,579
$2,532
$3,081
$3,012
$4,086
$3,993
$1,014
$989
$1,275
$1,228
$1,836
$1,762
$65
$0
2020
$6,607
$6,477
$4,746
$4,640
$6,294
$6,150
$2,434
$2,374
$3,061
$2,948
$4,409
$4,230
$134
$1,010
$2,506
$2,460
$2,993
$2,927
$3,970
$3,879
$1,012
$987
$1,273
$1,226
$1,834
$1,759
$62
$0
2021
$6,607
$6,477
$4,746
$4,640
$6,294
$6,150
$2,386
$2,326
$3,000
$2,889
$4,321
$4,145
$134
$1,010
$2,506
$2,460
$2,993
$2,927
$3,970
$3,879
$1,011
$986
$1,271
$1,224
$1,831
$1,756
$62
$0
2022
$6,607
$6,477
$4,746
$4,640
$6,294
$6,150
$2,338
$2,280
$2,940
$2,831
$4,235
$4,063
$134
$1,010
$2,506
$2,460
$2,993
$2,927
$3,970
$3,879
$1,009
$984
$1,269
$1,222
$1,828
$1,754
$62
$0
2023
$6,607
$6,477
$4,746
$4,640
$6,294
$6,150
$2,291
$2,234
$2,881
$2,775
$4,150
$3,981
$134
$1,010
$2,506
$2,460
$2,993
$2,927
$3,970
$3,879
$1,008
$983
$1,267
$1,220
$1,825
$1,751
$62
$0
2024
$6,607
$6,477
$4,746
$4,640
$6,294
$6,150
$2,245
$2,190
$2,824
$2,719
$4,067
$3,902
$134
$1,010
$2,506
$2,460
$2,993
$2,927
$3,970
$3,879
$1,006
$981
$1,265
$1,219
$1,823
$1,749
$62
$0
2025
$5,286
$5,182
$3,797
$3,712
$5,035
$4,920
$2,200
$2,146
$2,767
$2,665
$3,986
$3,824
$108
$1,010
$1,578
$1,550
$1,885
$1,844
$2,501
$2,444
$1,005
$980
$1,264
$1,217
$1,820
$1,746
$37
$0
3-133
-------
Technologies Considered in the Agencies' Analysis
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
TC
TC
TC
TC
TC
TC
TC
TC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
$8,878
$8,717
$10,605
$10,370
$14,066
$13,744
$4,241
$4,136
$5,333
$5,136
$7,682
$277
$277
$1,010
$7,545
$7,408
$9,013
$8,813
$11,954
$11,681
$4,186
$4,082
$5,264
$5,069
$7,582
$233
$233
$1,010
$7,545
$7,408
$9,013
$8,813
$11,954
$11,681
$3,498
$3,411
$4,399
$4,236
$6,336
$233
$233
$1,010
$6,479
$6,361
$7,739
$7,567
$10,264
$10,030
$3,446
$3,361
$4,334
$4,174
$6,243
$197
$197
$1,010
$6,479
$6,361
$7,739
$7,567
$10,264
$10,030
$3,396
$3,312
$4,271
$4,113
$6,152
$197
$197
$1,010
$6,479
$6,361
$7,739
$7,567
$10,264
$10,030
$3,347
$3,264
$4,209
$4,054
$6,063
$197
$197
$1,010
$6,479
$6,361
$7,739
$7,567
$10,264
$10,030
$3,299
$3,217
$4,148
$3,995
$5,975
$197
$197
$1,010
$6,479
$6,361
$7,739
$7,567
$10,264
$10,030
$3,251
$3,171
$4,089
$3,938
$5,890
$197
$197
$1,010
$4,757
$4,670
$5,682
$5,556
$7,536
$7,364
$3,205
$3,126
$4,031
$3,882
$5,806
$145
$145
$1,010
3.4.3.6.5 Electric vehicles
Electric vehicles (EV) - are vehicles with all-electric drive and with vehicle systems
powered by energy-optimized batteries charged primarily from grid electricity. While the
2016 FRM did not anticipate a significant penetration of EVs, in this analysis, EVs with
several ranges have been included. The GHG effectiveness is unchanged from estimates used
for 2016 model year vehicles in the 2012-2016 final rule which is 100 percent GHG
reduction. Per 49 U.S.C. 32904, NHTSA uses the Petroleum Equivalency Factor (PEF) in
calculating the effectiveness for EVs as stated in the section above for PHEV. The PEF is
determined by the U.S. Department of Energy as specified in 10 CFR Part 474. The PEF
accounts for U.S. average fossil-fuel electricity generation and transmission efficiencies,
petroleum refining and distribution efficiency, the energy content of gasoline, and includes a
0.15 divisor to incentivize the use of electricity in vehicles. The current PEF for electricity is
82.049 kWh per gallon of gasoline.
Once the fuel economy of the PHEV is calculated as shown in the previous section,
the effectiveness of PHEV incremental to EV can be calculated similarly using the formula
below.
3-134
-------
Technologies Considered in the Agencies' Analysis
The average effectiveness for the three pairs of vehicles of 68.54% is used in CAFE
modeling.
Battery costs assume that battery packs for EV applications will be designed to last for
the full useful life of the vehicle at a useable state of charge equivalent to 80% of the nominal
battery pack capacity. NHTSA included two levels of EVs, a 75-mile range EV and a 150-
mile range EV in this FRM analysis, consistent with the proposal. As this technology is
entering the market, it is expected that the OEMs will try to keep the cost low at the beginning
so that there will be more penetration. Due to the high cost of the battery packs at this early
stage of EVs, OEM will try to limit the battery pack size to reduce cost. Also the early
adopters for this technology are normally urban drivers and range anxiety is not believed to be
a big concern to them. Therefore NHTSA applied a 75-mile range EV for early adoption of
this technology in the market, up to 5% penetration. As the technology develops and as the
market penetration increases beyond 5%, NHTSA expects that OEMs would provide longer
driving range to help the consumers overcome range anxiety. NHTSA applied 150-mile EV
for this broad market adoption of this technology.
The cost of an EV consists of three parts, cost of battery pack, cost of non-battery systems,
and cost of charger and charger installation labor. An algorithm was used to select battery
sizes. The algorithm is described in Section 3.4.3.8 and the battery sizes applied in the Volpe
model for each type of EV and vehicle subclass are listed in Table 3-63.
Table 3-58 NHTSA Battery Sizes for EVs Applied in Volpe Model with No Mass Reduction (kWh)
EV75
EV100
EV150
Baseline
Fleet
2008
2010
2008
2010
2008
2010
Subcompact
PC/PerfPC
Compact PC/
PerfPC
22.79
23.65
30.39
31.54
45.58
47.31
Midsize
PC/PerfPC
28.03
28.72
37.38
38.30
56.07
57.45
Large
PC/PerfPC
33.28
34.54
44.37
46.05
66.55
69.08
Midsize LT
Minivan
n/a
n/a
n/a
n/a
n/a
n/a
Small LT
29.48
29.95
39.30
39.94
58.96
59.90
Large
LT
n/a
n/a
n/a
n/a
n/a
n/a
3-135
-------
Technologies Considered in the Agencies' Analysis
The agencies have applied a high complexity ICM to non-battery component cost for
EVs and EV chargers, which switch from short term value of 1.56 to long term value of 1.35
at 2018. The agencies applied a higher ICM factor to the battery of EVs due to the fact that it
a more complex technology. The ICM for EV battery switches from short term value of 1.77
to long term value of 1.50 at 2024. The agencies present costs of EVs in Sections 3.4.3.9 and
3.4.3.10. The costs of EVs without mass reduction as applied in Volpe model are listed in
Table 3-58 to Table 3-60. NHTSA accounts the cost impact from the interaction between
mass reduction and sizing of electrification system (battery and non-battery system) as cost
synergy as described in section 3.4.3.9.
Table 3-59 NHTSA Costs for EV75 Applied in Volpe Model with No Mass Reduction (2010$)
Tech.
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
Battery
Battery
Battery
Cost
Type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
NHTSA
Vehicle
Class
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Baseline
Fleet
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
2010
2008
2010
2008
2010
2008
2017
$10,324
$10,121
$12,140
$11,881
$15,634
$15,238
$410
$354
$1,267
$1,156
$2,236
$2,080
$395
$1,010
$4,441
$4,354
$5,222
$5,111
$6,725
$6,555
2018
$8,259
$8,097
$9,712
$9,505
$12,507
$12,190
$398
$343
$1,229
$1,122
$2,169
$2,018
$316
$1,010
$4,289
$4,205
$5,044
$4,936
$1,717
$6,331
2019
$8,259
$8,097
$9,712
$9,505
$12,507
$12,190
$386
$333
$1,193
$1,088
$2,104
$1,957
$316
$1,010
$4,289
$4,205
$5,044
$4,936
$1,712
$6,331
2020
$6,607
$6,477
$7,769
$7,604
$10,006
$9,752
$375
$323
$1,157
$1,055
$2,041
$1,899
$253
$1,010
$4,167
$4,086
$4,901
$4,796
$1,708
$6,151
2021
$6,607
$6,477
$7,769
$7,604
$10,006
$9,752
$363
$313
$1,122
$1,024
$1,980
$1,842
$253
$1,010
$4,167
$4,086
$4,901
$4,796
$1,703
$6,151
2022
$6,607
$6,477
$7,769
$7,604
$10,006
$9,752
$352
$304
$1,088
$993
$1,920
$1,786
$253
$1,010
$4,167
$4,086
$4,901
$4,796
$1,699
$6,151
2023
$6,607
$6,477
$7,769
$7,604
$10,006
$9,752
$345
$298
$1,067
$973
$1,882
$1,751
$253
$1,010
$4,167
$4,086
$4,901
$4,796
$1,696
$6,151
2024
$6,607
$6,477
$7,769
$7,604
$10,006
$9,752
$338
$292
$1,045
$954
$1,844
$1,716
$253
$1,010
$4,167
$4,086
$4,901
$4,796
$1,693
$6,151
2025
$5,286
$5,182
$6,215
$6,083
$8,005
$7,802
$332
$286
$1,024
$935
$1,808
$1,681
$202
$1,010
$2,625
$2,573
$3,087
$3,021
$1,090
$3,874
3-136
-------
Technologies Considered in the Agencies' Analysis
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
DMC
DMC
DMC
1C
1C
TC
TC
TC
TC
TC
TC
1C
1C
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
$316
$272
$976
$890
$1,722
$1,602
$126
$0
$14,765
$14,475
$17,362
$16,992
$22,359
$21,793
$726
$626
$2,243
$2,047
$3,959
$3,682
$521
$1,010
$315
$272
$973
$888
$18
$1,597
$121
$0
$12,548
$12,302
$14,755
$14,441
$19,002
$18,521
$713
$615
$2,203
$2,010
$3,887
$3,615
$437
$1,010
$314
$271
$970
$885
$18
$1,593
$121
$0
$12,548
$12,302
$14,755
$14,441
$19,002
$18,521
$700
$603
$2,163
$1,973
$3,817
$3,550
$437
$1,010
$313
$270
$968
$883
$18
$1,588
$117
$0
$10,775
$10,563
$12,670
$12,400
$16,317
$15,903
$688
$593
$2,125
$1,938
$3,749
$3,487
$370
$1,010
$313
$269
$965
$881
$18
$1,584
$117
$0
$10,775
$10,563
$12,670
$12,400
$16,317
$15,903
$676
$582
$2,087
$1,904
$3,683
$3,426
$370
$1,010
$312
$269
$963
$878
$18
$1,580
$117
$0
$10,775
$10,563
$12,670
$12,400
$16,317
$15,903
$664
$572
$2,051
$1,871
$3,619
$3,367
$370
$1,010
$311
$268
$961
$877
$18
$1,578
$117
$0
$10,775
$10,563
$12,670
$12,400
$16,317
$15,903
$657
$566
$2,028
$1,850
$3,578
$3,328
$370
$1,010
$311
$268
$960
$876
$18
$1,575
$117
$0
$10,775
$10,563
$12,670
$12,400
$16,317
$15,903
$649
$559
$2,005
$1,829
$3,538
$3,291
$370
$1,010
$200
$172
$618
$563
$10
$1,014
$70
$0
$7,911
$7,755
$9,302
$9,104
$11,980
$11,676
$532
$458
$1,642
$1,498
$2,897
$2,695
$272
$1,010
Table 3-60 NHTSA Costs for EV100 Applied in Volpe Model with No Mass Reduction (2010$)
Tech.
Battery
Battery
Cost
Type
DMC
DMC
NHTSA
Vehicle
Class
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Baseline
Fleet
2010
2008
2010
2008
2017
$12,341
$12,063
$14,159
$13,919
2018
$9,873
$9,651
$11,327
$11,135
2019
$9,873
$9,651
$11,327
$11,135
2020
$7,898
$7,720
$9,062
$8,908
2021
$7,898
$7,720
$9,062
$8,908
2022
$7,898
$7,720
$9,062
$8,908
2023
$7,898
$7,720
$9,062
$8,908
2024
$7,898
$7,720
$9,062
$8,908
2025
$6,319
$6,176
$7,250
$7,127
3-137
-------
Technologies Considered in the Agencies' Analysis
Battery
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
Battery
Battery
Battery
Non-
battery
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
DMC
DMC
DMC
1C
1C
TC
TC
TC
TC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
2010
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
2010
2008
2010
2008
2010
2008
2010
$17,482
$17,025
$410
$354
$1,267
$1,156
$2,236
$2,080
$395
$1,010
$5,309
$5,189
$6,091
$5,988
$7,520
$7,324
$316
$272
$976
$890
$1,722
$1,602
$126
$0
$17,650
$17,253
$20,250
$19,907
$25,002
$24,349
$726
$13,985
$13,620
$398
$343
$1,229
$1,122
$2,169
$2,018
$316
$1,010
$5,127
$5,012
$5,883
$5,783
$7,263
$7,073
$315
$272
$973
$888
$1,717
$1,597
$121
$0
$15,000
$14,662
$17,210
$16,918
$21,248
$20,693
$713
$13,985
$13,620
$386
$333
$1,193
$1,088
$2,104
$1,957
$316
$1,010
$5,127
$5,012
$5,883
$5,783
$7,263
$7,073
$314
$271
$970
$885
$1,712
$1,593
$121
$0
$15,000
$14,662
$17,210
$16,918
$21,248
$20,693
$700
$11,188
$10,896
$375
$323
$1,157
$1,055
$2,041
$1,899
$253
$1,010
$4,982
$4,870
$5,716
$5,619
$7,057
$6,873
$313
$270
$968
$883
$1,708
$1,588
$117
$0
$12,880
$12,590
$14,778
$14,527
$18,245
$17,769
$688
$11,188
$10,896
$363
$313
$1,122
$1,024
$1,980
$1,842
$253
$1,010
$4,982
$4,870
$5,716
$5,619
$7,057
$6,873
$313
$269
$965
$881
$1,703
$1,584
$117
$0
$12,880
$12,590
$14,778
$14,527
$18,245
$17,769
$676
$11,188
$10,896
$352
$304
$1,088
$993
$1,920
$1,786
$253
$1,010
$4,982
$4,870
$5,716
$5,619
$7,057
$6,873
$312
$269
$963
$878
$1,699
$1,580
$117
$0
$12,880
$12,590
$14,778
$14,527
$18,245
$17,769
$664
$11,188
$10,896
$345
$298
$1,067
$973
$1,882
$1,751
$253
$1,010
$4,982
$4,870
$5,716
$5,619
$7,057
$6,873
$311
$268
$961
$877
$1,696
$1,578
$117
$0
$12,880
$12,590
$14,778
$14,527
$18,245
$17,769
$657
$11,188
$10,896
$338
$292
$1,045
$954
$1,844
$1,716
$253
$1,010
$4,982
$4,870
$5,716
$5,619
$7,057
$6,873
$311
$268
$960
$876
$1,693
$1,575
$117
$0
$12,880
$12,590
$14,778
$14,527
$18,245
$17,769
$649
$8,951
$8,717
$332
$286
$1,024
$935
$1,808
$1,681
$202
$1,010
$3,138
$3,067
$3,600
$3,539
$4,445
$4,329
$200
$172
$618
$563
$1,090
$1,014
$70
$0
$9,457
$9,244
$10,850
$10,666
$13,396
$13,046
$532
3-138
-------
Technologies Considered in the Agencies' Analysis
Non-
battery
Non-
battery
Charger
Charger
Labor
TC
TC
1C
1C
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
2008
2010
2008
2010
2008
2008/2010
2008/2010
$626
$2,243
$2,047
$3,959
$3,682
$521
$1,010
$615
$2,203
$2,010
$3,887
$3,615
$437
$1,010
$603
$2,163
$1,973
$3,817
$3,550
$437
$1,010
$593
$2,125
$1,938
$3,749
$3,487
$370
$1,010
$582
$2,087
$1,904
$3,683
$3,426
$370
$1,010
$572
$2,051
$1,871
$3,619
$3,367
$370
$1,010
$566
$2,028
$1,850
$3,578
$3,328
$370
$1,010
$559
$2,005
$1,829
$3,538
$3,291
$370
$1,010
$458
$1,642
$1,498
$2,897
$2,695
$272
$1,010
Table 3-61 NHTSA Costs for EV150 Applied in Volpe Model with No Mass Reduction (2010$)
Tech.
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
Battery
Battery
Battery
Cost
Type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
NHTSA
Vehicle
Class
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
Baseline
Fleet
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
2010
2008
2010
2008
2010
2017
$16,369
$15,939
$19,585
$19,240
$22,552
$21,936
$410
$355
$1,267
$1,157
$2,236
$2,082
$395
$1,010
$7,042
$6,857
$8,425
$8,277
$9,702
2018
$13,095
$12,751
$15,668
$15,392
$18,042
$17,549
$398
$344
$1,229
$1,123
$2,169
$2,019
$316
$1,010
$6,801
$6,622
$8,137
$7,993
$9,370
2019
$13,095
$12,751
$15,668
$15,392
$18,042
$17,549
$386
$334
$1,193
$1,089
$2,104
$1,959
$316
$1,010
$6,801
$6,622
$8,137
$7,993
$9,370
2020
$10,476
$10,201
$12,534
$12,313
$14,433
$14,039
$375
$324
$1,157
$1,056
$2,041
$1,900
$253
$1,010
$6,608
$6,434
$7,906
$7,767
$9,104
2021
$10,476
$10,201
$12,534
$12,313
$14,433
$14,039
$363
$314
$1,122
$1,025
$1,980
$1,843
$253
$1,010
$6,608
$6,434
$7,906
$7,767
$9,104
2022
$10,476
$10,201
$12,534
$12,313
$14,433
$14,039
$352
$305
$1,088
$994
$1,920
$1,788
$253
$1,010
$6,608
$6,434
$7,906
$7,767
$9,104
2023
$10,476
$10,201
$12,534
$12,313
$14,433
$14,039
$345
$299
$1,067
$974
$1,882
$1,752
$253
$1,010
$6,608
$6,434
$7,906
$7,767
$9,104
2024
$10,476
$10,201
$12,534
$12,313
$14,433
$14,039
$338
$293
$1,045
$954
$1,844
$1,717
$253
$1,010
$6,608
$6,434
$7,906
$7,767
$9,104
2025
$8,381
$8,161
$10,028
$9,851
$11,547
$11,231
$332
$287
$1,024
$935
$1,808
$1,682
$202
$1,010
$4,162
$4,053
$4,980
$4,892
$5,734
3-139
-------
Technologies Considered in the Agencies' Analysis
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
Battery
Battery
Battery
Non-
battery
Non-
battery
Non-
battery
Charger
Charger
Labor
DMC
DMC
DMC
1C
1C
TC
TC
TC
TC
TC
TC
1C
1C
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
Subcompact
PC/PerfPC
Compact
PC/PerfPC
Midsize
PC/PerfPC
Large
PC/PerfPC
All
All
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2010
2008
2008/2010
2008/2010
$9,437
$316
$273
$976
$891
$1,722
$1,603
$126
$0
$23,411
$22,796
$28,010
$27,517
$32,254
$31,372
$726
$628
$2,243
$2,048
$3,959
$3,685
$521
$1,010
$9,114
$315
$273
$973
$889
$1,717
$1,598
$121
$0
$19,896
$19,373
$23,805
$23,385
$27,411
$26,662
$713
$617
$2,203
$2,011
$3,887
$3,618
$437
$1,010
$9,114
$314
$272
$970
$886
$1,712
$1,594
$121
$0
$19,896
$19,373
$23,805
$23,385
$27,411
$26,662
$700
$606
$2,163
$1,975
$3,817
$3,552
$437
$1,010
$8,855
$313
$271
$968
$884
$1,708
$1,590
$117
$0
$17,084
$16,635
$20,441
$20,080
$23,537
$22,894
$688
$595
$2,125
$1,940
$3,749
$3,489
$370
$1,010
$8,855
$313
$270
$965
$881
$1,703
$1,585
$117
$0
$17,084
$16,635
$20,441
$20,080
$23,537
$22,894
$676
$585
$2,087
$1,906
$3,683
$3,428
$370
$1,010
$8,855
$312
$270
$963
$879
$1,699
$1,581
$117
$0
$17,084
$16,635
$20,441
$20,080
$23,537
$22,894
$664
$574
$2,051
$1,873
$3,619
$3,369
$370
$1,010
$8,855
$311
$269
$961
$878
$1,696
$1,579
$117
$0
$17,084
$16,635
$20,441
$20,080
$23,537
$22,894
$657
$568
$2,028
$1,852
$3,578
$3,330
$370
$1,010
$8,855
$311
$269
$960
$876
$1,693
$1,576
$117
$0
$17,084
$16,635
$20,441
$20,080
$23,537
$22,894
$649
$561
$2,005
$1,831
$3,538
$3,293
$370
$1,010
$5,578
$200
$173
$618
$564
$1,090
$1,014
$70
$0
$12,543
$12,214
$15,007
$14,743
$17,281
$16,809
$532
$460
$1,642
$1,499
$2,897
$2,697
$272
$1,010
3.4.3.6.6 Fuel cell electric vehicles
Fuel cell electric vehicles (FCEVs) - utilize a full electric drive platform but consume
electricity generated by an on-board fuel cell and hydrogen fuel. Fuel cells are electro-
chemical devices that directly convert reactants (hydrogen and oxygen via air) into electricity,
with the potential of achieving more than twice the efficiency of conventional internal
combustion engines. High pressure gaseous hydrogen storage tanks are used by most
automakers for FCEVs that are currently under development. The high pressure tanks are
similar to those used for compressed gas storage in more than 10 million CNG vehicles
3-140
-------
Technologies Considered in the Agencies' Analysis
worldwide, except that they are designed to operate at a higher pressure (350 bar or 700 bar
vs. 250 bar for CNG). Due to the uncertainty of the future availability for this technology,
FCEVs were not included in any OMEGA or Volpe model runs.
3.4.3.7 Batteries for Mild HEV, HEV, PHEV and EV Applications
The design of battery secondary cells can vary considerably between Stop/Start, Mild
HEV (ISO), HEV, PHEV and EV applications.
MHEV batteries: Due to their lower voltage (12-42 VDC) and reduced power and
energy requirements, MHEV systems may continue to use lead-acid batteries even long term
(2017 model year and later). MHEV battery designs differ from those of current starved-
electrolyte (typical maintenance free batteries) or flooded-electrolyte (the older style lead-acid
batteries requiring water "top-off) batteries used for starting, lighting and ignition (SLI) in
automotive applications. Standard SLI batteries are primarily designed to provide high-
current for engine start-up and then recharge immediately after startup via the vehicle's
charging system. Deeply discharging a standard SLI battery will greatly shorten its life.
MHEV applications are expected to use:
• Extended-cycle-life flooded (ELF) lead-acid batteries
• Absorptive glass matt, valve-regulated lead-acid (AGM/VRLA) batteries -or -
• Asymmetric lead-acid battery/capacitor hybrids (e.g., flooded ultrabatteries)
MHEV systems using electrolytic double-layer capacitors are also under development and
may provide improved performance and reduced cost in the post-2017 timeframe.
Mild HEV and HEV batteries: Mild HEV and HEV applications operate in a narrow,
short-cycling, charge-sustaining state of charge (SOC). Energy capacity in Mild HEV and
HEV applications is somewhat limited by the ability of the battery and power electronics to
accept charge and by space and weight constraints within the vehicle design. Mild HEV and
HEV battery designs tend to be optimized for high power density rather than high energy
density, with thinner cathode and anode layers and more numerous current collectors and
separators (Figure 3-20).
EV batteries: EV batteries tend to be optimized for high energy density and are
considerably larger and heavier than HEV batteries in order to provide sufficient energy
capacity. EV battery cells tend to have thicker cathode and anode layers and fewer collectors
and separators than HEV cells. This reduces the specific cost on a per-kWh basis for EV
battery cells relative to HEV battery cells.
PHEV batteries: PHEV battery designs are intermediate between power-optimized
HEV and energy-optimized EV battery cell designs. PHEV batteries must provide both
charge depleting operation similar to an EV and charge sustaining operation similar to an
HEV. Unlike HEV applications, charge-sustaining operation with PHEVs occurs at a
relatively low battery state of charge (SOC) which can pose a significant challenge with
3-141
-------
Technologies Considered in the Agencies' Analysis
respect to attaining acceptable battery cycle life. In the case of the GM Volt, this limits
charge depleting operation to a minimum SOC of approximately 30 percent.59 An alternative
approach for PHEV applications that has potential to allow extension of charge depletion to a
lower battery SOC is using energy-optimized lithium-ion batteries for charge depleting
operation in combination with the use of supercapacitors or power-optimized batteries for
charge sustaining operation.60
Figure 3-20: Schematic representation of power and energy optimized
prismatic-layered battery cells
Collector (-)
Cathode (-)
Separator
Anode (+)
Collector (+)
HEV Power-optimized Battery Cell
EV Energy-optimized Battery Cell
Power-split hybrid vehicles from Toyota, Ford and Nissan, integrated motor assist
hybrid vehicles from Honda and the GM 2-mode hybrid vehicles currently use nickel-metal
hydride (NiMH) batteries. Lithium-ion (Li-ion) batteries offer the potential to approximately
double both the energy and power density relative to current NiMH batteries, enabling much
more electrification of automotive drive applications such as PHEVs and EVs.
Li-ion batteries for high-volume automotive applications differ substantially from
those used in consumer electronics applications with respect to cathode chemistry,
construction and cell size. Li-ion battery designs currently in production by CPI (LG-Chem)
for the GM Volt PHEV and by AESC and GS-Yuasa (respectively) for the Nissan Leaf and
Mitsubishi iMiEV use large-format, layered-prismatic cells assembled into battery modules.
The modules are then combined into battery packs.
Two families of cathode chemistries are used in large-format, automotive Li-ion
batteries currently in production - LiMn2O4-spinel (CPI, GS-Yuasa, AESC) and LiFePO4
(A123 Systems). Current production batteries typically use graphite anodes. Automotive Li-
3-142
-------
Technologies Considered in the Agencies' Analysis
ion batteries using lithium nickel manganese cobalt (NMC) oxide cathodes with graphite
anodes are in advanced stages of development for PHEV and EV applications. The agencies
expect large-format Li-ion batteries to completely replace NiMH batteries for post-2017 HEV
applications. We also expect that stacked and/or folded prismatic Li-ion cell designs will
continue to be used for PHEV and EV applications and that NMC/graphite Li-ion batteries
will be a mature technology for 2017-2025 light-duty vehicle applications.
3.4.3.8 HEV, PHEV and EV System Sizing Methodology
Battery packs are (and will continue to be) one of the most expensive components for
EVs, PHEVs and HEVs. To obtain reasonable cost estimates for electrified vehicles, it was
therefore important to establish a reliable approach for determining battery attributes for each
vehicle and class. Both battery energy content ("size") and power rating are key inputs used
to establish costs per ANL's battery costing model. For EVs and PHEVs in particular, battery
size and weight are closely related, and so battery weight must be known as well. The
following section details the steps taken to size a battery for
a) EVs and PHEVs (at various all-electric ranges),
b) a more simplified separate approach for MHEVs and HEVs.
3.4.3.8.1 Battery Pack Sizing for EVs and PHEVs
Calculation of required battery pack energy requirements for EVs and PHEVs is not
straightforward. Because vehicle energy consumption is strongly dependent on weight, and
battery packs are very heavy, the weight of the battery pack itself can change the energy
required to move the vehicle. As vehicle energy consumption increases, the battery size must
increase for a given range (in the case of EVs and PHEVs) - as a result, vehicle weight
increases, and per-mile energy consumption increases as well, increasing the battery size, and
soon.
EPA built spreadsheets to estimate the required battery size for each vehicle and class.
Listed below are the steps EPA has taken in these spreadsheets to estimate not only battery
size, but associated weight for EVs and PHEVs of varying ranges and designs.
1. Establish baseline FE/energy consumption
2. Assume nominal weight of electrified vehicle (based on weight reduction
target)
3. Calculate vehicle energy demand at this target weight
4. Calculate required battery energy
5. Calculate actual battery and vehicle weight
6. Do vehicle weight and battery size match estimated values?
3-143
-------
Technologies Considered in the Agencies' Analysis
Steps 2-6 were iterated until each assumed weight reduction target (and nominal
vehicle weight) reconciled with required battery size and the calculated weight of each
vehicle.
Baseline vehicle energy consumption is estimated based on a fitted trendline for FE
vs. inertia weight, or ETW (from FE Trends data for 2008 MY vehicles, table M-80) and
converting to Wh/mi. This is shown in Figure 3-21.
2008 Fuel Economy vs. Inertia Weight
(source: Fuel Economy Trends Report, Table M-80)
si
£
O 20
y - O.OOOOOlSOx2 - 0.02194637X + 85.81284974
Rz= 0.99565332
Inertia wt (Ibs)
Figure 3-21: Average fuel economy based on inertia weight (ETW) from FE Trends data
Then, fuel economy was converted into energy consumption (assuming 33700 Wh
energy in 1 gallon of gasoline) and used to populate a range of test weights between 2000 and
6000 Ibs. A linear trendline was used to fit this curve and then applied to estimate generic
energy consumption for baseline vehicles of a given ETW (shown below in Figure 3-22).
3-144
-------
Technologies Considered in the Agencies' Analysis
2008 Energy Consumption vs. Inertia Weight
IS
Inertia wt (Ibs)
Figure 3-22: Equivalent energy consumption (in Wh/mi) for baseline vehicles
To calculate battery pack size, the electrified vehicle weight must first be known; to
calculate vehicle weight, the battery pack size must first be known. This circular reference
required an iterative solution. EPA assumed a target vehicle glider (a rolling chassis with no
powertrain) weight reduction and applied that to the baseline curb weight. The resulting
nominal vehicle weight was then used to calculate the vehicle energy demand. To calculate
the energy demand (efficiency) of an electric vehicle in Wh/mi, the following information
was needed:
• Baseline energy consumption / mpg
• Efficiency (r|) improvement of electric vehicle
• Change in road loads
In Table 3-62 below, the following definitions apply:
• Brake eff (brake efficiency) - the % amount of chemical fuel energy converted to
energy at the engine crankshaft (or, for batteries, the amount of stored electrical
energy converted to shaft energy entering the transmission)
• D/L eff (driveline efficiency) - the % of the brake energy entering the transmission
delivered through the driveline to the wheels
• Wheel eff (wheel efficiency) - the product of brake and driveline efficiency
• Cycle eff (cycle efficiency) - the % of energy delivered to the wheels used to
overcome road loads and power the vehicle (it does not include energy lost as
braking heat)
• Vehicle efficiency - the product of wheel and cycle efficiency
3-145
-------
Technologies Considered in the Agencies' Analysis
• Road loads - the amount of resistant energy the vehicle must overcome during a
city/highway test. Composed of vehicle weight (inertia), aerodynamic drag and
rolling resistance
Table 3-62: EV100 efficiency and energy demand calculations, 20% applied weight reduction
Class
Baseline
gas ICE
Small
car
Std car
Large
car
Small
MPV
Large
MPV
Truck
Brake
eff
24%
85%
85%
85%
85%
85%
85%
D/L
eff
81%
93%
93%
93%
93%
93%
93%
Wheel
eff
20%
79%
79%
79%
79%
79%
79%
Cycle
eff
77%
97%
97%
97%
97%
97%
97%
Vehicle
eff
15%
77%
77%
77%
77%
77%
77%
Road
loads
100%
88%
88%
88%
89%
89%
88%
Energy
reduction
83%
83%
83%
83%
83%
83%
Energy
eff
increase
478%
478%
478%
475%
475%
482%
IW-
based,
base ICE
nominal
mpgge
37
30
25
29
23
20
Base
fuel
energy
req'd
Wh/mi
912
1122
1332
1180
1497
1727
FTP
fuel
energy
req'd
Wh/mi
158
194
230
205
260
297
Onroad
fuel
energy
req'd
Wh/mi
225
277
329
293
372
424
The energy efficiency of a baseline vehicle (around 15%), as indicated in the table
above, was estimated using efficiency terms derived from EPA's lumped parameter model
(engine/battery brake efficiency, driveline efficiency, cycle efficiency and road load ratio to
baseline). To calculate the energy consumption of an EV (or PHEV in charge-depleting
mode), the following assumptions were made:
• "Brake" efficiency (for an EV, the efficiency of converting battery energy to
tractive energy at the transmission input shaft) was estimated at 85% - assuming,
roughly a 95% efficiency for the battery, motor, and power electronics,
respectively.
• The driveline efficiency (including the transmission) was comparable to the value
calculated by the lumped parameter model for an advanced 6-speed dual-clutch
transmission at 93%.
• The cycle efficiency assumes regenerative braking where 97% recoverable braking
energy is recaptured. As a result, most of the energy delivered to the wheels is
used to overcome road loads.
• The road loads were based on the weight reduction of the vehicle. In the case of a
100 mile EV with a 20% weight reduction, road loads (as calculated by the LP
model) are reduced to 88-89% of the baseline vehiclew.
' Included in this example road load calculation is a 10% reduction in rolling resistance and aerodynamic drag.
3-146
-------
Technologies Considered in the Agencies' Analysis
The energy consumption of the EV includes ratio of the roadloads of the EV to the
baseline vehicle, and the ratio of the efficiency of the EV compared to the baseline
vehicle. It is expressed mathematically as shown below in Equation 3-1: EV
energy consumption:
Equation 3-1: EV energy consumption
In Table 3-63, the baseline energy required (in Wh/mi) is in the column labeled "Base
fuel energy reqd". The energy required for each vehicle class EV over the FTP is in the
column "FTP fuel energy reqd Wh/mi" and incorporates the equation above. This energy rate
refers to the laboratory or unadjusted test cycle value, as opposed to a real-world "onroad"
value. EPA assumes a 30% fuel economy shortfall, based loosely on the 5-cycle Fuel
Economy Labeling Rule (year) which is directionally correct for electrified vehicles. This
corresponds to an increase in fuel consumption of 43%. Applying this 43% increase gives the
onroad energy consumption values for EVs as shown in the far right column of the previous
table. From this value, one can determine an appropriate battery pack size for the vehicle.
The required battery energy for EVs equals the onroad energy consumption,
multiplied by the desired range, divided by the useful state-of-charge window of the battery.
It is calculated as follows in Equation 3-2
Equation 3-2: Required battery pack energy (size) for EVs
Assumed usable SOC (battery state-of-charge) windows were 80% for EVs (10-90%)
and 70% for PHEVs (15%-85%). The battery pack sizes are listed in orange in Table 3-63 for
the 100-mile EV case and show both the onroad energy consumption ("EV adj Wh/mi"
column) and the nominal battery energy content or "battery pack size".
3-147
-------
Technologies Considered in the Agencies' Analysis
Table 3-63: Battery pack sizes for EV100 based on inertia weight, 20% applied weight reduction
Class
Baseline curb wt
(Ib)
Inertia wt
(Ib)
EV unadj
(Wh/mi)
EVadj
(Wh/mi)
100 mile batt pack size
(kWh)
2008 Baseline
Small car
Std car
Large car
Small MPV
Large MPV
Truck
2633
3306
3897
3474
4351
5108
2933
3606
4197
3774
4651
5408
158
194
230
205
260
297
225
277
329
293
372
424
28.2
34.7
41.1
36.7
46.5
53.0
20 10 Baseline
Small car
Stdcar
Large car
Small MPV
Large MPV
Truck
2753
3387
4035
3528
4313
5346
3053
3687
4335
3828
4613
5646
164
200
241
209
257
307
234
286
344
298
367
439
29.2
35.7
43.0
37.3
45.8
54.8
EPA used the following formula to determine weight of an EV (Equation 3-3):
Equation 3-3: EV weight calculation
Any weight reduction technology was applied only to the glider (baseline vehicle
absent powertrain) as defined in Equation 3-4.
Equation 3-4: Weight reduction of the glider
In the case of a PHEV, it was assumed that the base ICE powertrain remains so it is
not deducted; the proper equation for PHEVs is shown in Equation 3-5.
Equation 3-5: Weight calculation for PHEV
class:
Listed in Table 3-64 are the assumed baseline ICE-powertrain weights, by vehicle
3-148
-------
Technologies Considered in the Agencies' Analysis
Table 3-64: Baseline ICE-powertrain weight assumptions, by class
Class
Small car
Std car
Large car
Small
MPV
Large
MPV
Truck
Engine
250
300
375
300
400
550
Trans
(diffnot
included)
125
150
175
150
200
200
Fuel sys
(50%
fill)
50
60
70
60
80
100
Engine mounts/
NVH
treatments
25
25
25
25
25
25
Exhaust
20
25
30
25
30
40
12V
batt
25
30
35
30
40
50
Total ICE powertrain
weight
495
590
710
590
775
965
EPA then estimated the weight of the electric drive subsystem using the energy
content of the battery pack as an input. EPA scaled the weight by applying a specific energy
for the electric drive subsystem-including the battery pack, drive motor, wiring, power
electronics, etc.-of 120 Wh/kg (or 18.33 Ib/kWh). This specific energy value is based on
adding components to an assumed battery pack specific energy of 150 Wh/kgww. Then, the
gearbox (the only subsystem excluded from the electric drive scaling) was added to the
weight of the electric drive subsystem; this total was included into the electric vehicle weight
calculation as Weiectric drive™ • A summary table of electric drive weights for 100-mile EVs is
shown as Table 3-65.
ww 150 Wh/kg is a conservative estimate for year 2017 and beyond: outputs from ANL's battery cost model,
which shows specific energy values of 160-180 Wh/kg for a similar timeframe.
xx Applies only to the EV. Because the baseline ICE powertrain weight (which includes gearbox weight) was not
deducted from the PHEV, it is not added back in for the PHEV.
3-149
-------
Technologies Considered in the Agencies' Analysis
Table 3-65: Total electric drive weights for 100-mile EVs
Class
Batt pack size
(kWh)
2020 electric content
(Ibs)
Gearbox
(power-split or other )
2020 EV
powertrain total
2008 Baseline
Small car
Std car
Large car
Small MPV
Large MPV
Truck
28.2
34.7
41.1
36.7
46.5
53.0
517
635
754
672
853
972
50
60
70
60
80
100
567
695
824
732
933
1072
20 10 Baseline
Small car
Std car
Large car
Small MPV
Large MPV
Truck
29.2
35.7
43.0
37.3
45.8
54.8
536
655
788
683
840
1005
50
60
70
60
80
100
586
715
858
743
920
1105
The difference between the actual weight and the predicted or nominal weight should
be zero. However, if not then a revised weight reduction was used for another iteration of
steps 2-6 until the two vehicle weights match. Spreadsheet tools such as "solver" in MS
Excel were used for automating this iterative process.
Table 3-66 shows example results for 100-mile range EVs; in this case a 20% applied
glider weight reduction for a variety of vehicle classes.
Table 3-66: Sample calculation sheet for 100-mile EVs for the 2008 Baseline
Class
Small
car
Std
car
Large
car
Small
MPV
Large
MPV
Truck
Base
curb
wt
(Ib)
2633
3306
3897
3474
4351
5108
Base
power/wt
ratio
0.0486
0.0575
0.0872
0.0463
0.0565
0.0617
Powertrain
weight
(Ib)
495
590
710
590
775
965
Base
glider
wt
(Ib)
2138
2716
3187
2884
3576
4143
WR
of
glider
428
543
637
577
715
829
NewEV
wt
(nominal
Ib)
2205
2763
3260
2897
3636
4279
Energy
cons
adjusted
(Wh/mi)
225
277
329
293
372
424
Batt
pack
size
(kWh)
28.2
34.7
41.1
36.7
46.5
53.0
Electric
drive wt
(Ib)
567
695
824
732
933
1072
New
EV
weight
(Ib)
2277
2868
3374
3039
3794
4387
Error
0
0
0
0
0
0
%WR
from
curb
13.5%
13.2%
13.4%
12.5%
12.8%
14.1%
%RL
vs
base
88%
88%
88%
89%
89%
88%
Table 3-67 shows the effect on net electric vehicle weight reduction after 20% glider
weight reduction was applied to EVs and PHEVs. As battery pack size increases for larger-
range EVs and PHEVs, the overall realized vehicle weight reduction decreases (because it
requires more energy to carry the extra battery weight). In this example, EVs with a 150 mile
range require almost 20% weight reduction to the glider to make up for the additional weight
of the electric drive and battery pack compared to a conventional ICE-based powertrain.
3-150
-------
Technologies Considered in the Agencies' Analysis
Table 3-67: Actual weight reduction percentages for EVs and PHEVs with 20% weight reduction applied
to glider
75 Mile EV
Actual %WR
vs.
base vehicle
100 Mile EV
Actual %WR
vs.
base vehicle
150 Mile EV
Actual %WR
vs.
base vehicle
20 Mile PHEV
Actual %WR
vs.
base vehicle
40 Mile PHEV
Actual %WR
vs.
base vehicle
2008 Baseline
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%
20 10 Baseline
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
18%
18%
18%
18%
18%
19%
13%
13%
13%
12%
13%
14%
1%
1%
1%
1%
1%
3%
12%
12%
12%
12%
12%
11%
7%
7%
7%
8%
7%
6%
Because there is no "all-electric range" requirement for HEVs, battery pack sizes were
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 rather
insignificant. For these reasons, a more direct approach (rather than an iterative process)
works for battery sizing of HEVs.
• FIEV batteries were scaled similar to the 2010 Fusion Hybrid based on nominal
battery energy per Ib ETW (equivalent test weight), at 0.37 Wh/lb.
• A higher usable SOC window of 40% (compared to 30% for Fusion Hybrid)
reduced the required Li-Ion battery size to 75% of the Fusion Hybrid's NiMH
battery. This resulted in a 0.28 Wh/lb ETW ratio.
• In comparing anecdotal data for HEVs, the agencies assumed a slight weight
increase of 4-5% 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.
3.4.3.9 HEV, PHEV and EV battery pack design and cost analysis using the ANL
BatPaC 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.
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
3-151
-------
Technologies Considered in the Agencies' Analysis
DOE National Laboratories responsible for basic and applied battery energy storage
technologies for future HEV, PHEV and EV applications. A basic description of the ANL Li-
ion battery cost model and initial modeling results for PHEV applications were published in a
peer-reviewed technical paper presented at EVS-2461. ANL has extended modeling inputs
and pack design criteria within the battery cost model to include analysis of manufacturing
costs for EVs and HEVs as well has PHEVs.62 In early 2011, ANL issued a draft report
detailing the methodology, inputs and outputs of their Battery Performance and Cost (BatPaC)
model.63 A complete independent peer-review of the BatPaC model and its inputs and results
for HEV, PHEV and EV applications has been completed64. ANL recently provided the
agencies with an updated report documenting the BatPaC model that fully addresses the issues
raised within the peer review.65 Based on the feedback from peer-reviewers, ANL updated the
model in the following areas.
1. Battery pack cost is adjusted upward. This adjustment is based on the feedback
from several peer-reviewers, and changes are related to limiting electrode
thickness to 100 microns, changing allocation of overhead cost to more closely
represent a Tier 1 auto supplier, increasing cost of tabs, changing capital cost
of material preparation, etc;
2. Battery management system cost is increased to represent the complete
monitoring and control needs for proper battery operation and safety as shown
in Table 5.3 in the report;
3. Battery automatic and manual disconnect unit cost is added based on safety
considerations as shown in Table 5.3 in the report;
4. Liquid thermal management system is added. ANL stated in the report that the
liquid-cooled closure design it uses in the model would not have sufficient
surface area and cell spacing to be cooled by air effectively as shown in Table
5.3 in the report.
Subsequently, the agencies requested that an option be added to select between liquid
or air thermal management and that adequate surface area and cell spacing be determined
accordingly. Also, the agencies requested a feature to allow battery packs to be configured as
subpacks in parallel or modules in parallel, as additional options for staying within voltage
and cell size limits for large packs.
ANL added these features in a version of the model distributed March 1, 2012. This
version of the model is used for the battery cost estimates in the final rule. This model and the
peer review report are available in the public dockets for this rulemaking.64'66
NHTSA and EPA decided to use the ANL BatPaC model for estimating large-format
lithium-ion batteries for this final rule, consistent with the proposal, for the following reasons.
First, the ANL model 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
3-152
-------
Technologies Considered in the Agencies' Analysis
this area. The model uses a bill of materials methodology which the agencies believe is the
preferred method for developing cost estimates. The ANL model appropriately considers the
vehicle applications power and energy requirements, which are two of the fundamental
parameters when designing a lithium-ion battery for an HEV, PHEV, or EV. The ANL model
can estimate high volume production costs, which the agencies believe is appropriate for the
2025 time frame. Finally, the ANL model's cost estimates, while generally lower than the
estimates we received from the OEMs, is consistent with some of the supplier cost estimates
the agencies received from large-format lithium-ion battery pack manufacturers. A portion of
those data was received from on-site visits to vehicle manufacturers and battery suppliers
done by the EPA in 2008.
The ANL battery cost model 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 as well as the design criteria include a vehicle application's power
and energy storage capacity requirements, the battery's cathode and anode chemistry, and the
number of cells per module and modules per battery pack. The model assumes use of a
laminated multi-layer prismatic cell and battery modules consisting of double-seamed rigid
containers. The model also assumes that the battery modules are liquid-cooled. 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.
Basic user inputs to BatPaC include performance goals (power and energy capacity),
choice of battery chemistry (of five predefined chemistries), the vehicle type for which the
battery is intended (HEV, PHEV, or EV), the desired number of cells and modules, and the
volume of production. BatPaC then designs the cells, modules, and battery pack, and
provides an 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 encouraged to change
these defaults 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.
Also inherent in BatPaC are certain modeling assumptions that are still open to some
uncertainty or debate in the industry. For some, such as the available portion of total battery
3-153
-------
Technologies Considered in the Agencies' Analysis
energy (aka "SOC window") for a PHEV/EV/HEV, the user can easily modify a single
parameter to represent a value other than the default. For others, such as specific unit costs
for thermal management or battery monitoring components, changes can often be made by
replacing the relevant components of the model outputs.
The cost outputs used by the agencies to determine 2025 HEV, PHEV and EV battery
costs were based on the following inputs and assumptions.
EPA selected basic user inputs as follows. For performance goals, EPA used the
power and energy requirements derived from the scaling analysis described in the previous
section. Specifically, these covered each of the six classes of vehicles (Small Car, Standard
Car, Large Car, Small MPV, Large MPV and Truck) under each of the five weight reduction
scenarios (0%, 2%, 7.5%, 10%, and 20%). The chosen battery chemistries were NMC441-G
(for EVs and PHEV40) and LMO-G (for P2 HEVs and PHEV20). Vehicle types were EV75,
EV100, EV150 (using the BatPaC "EV" setting); PHEV20 and PHEV40 (using the "PHEV"
setting), and P2 HEV (using the "HEV-HP" setting). All modules were composed of 32 cells,
with each pack having a varying number of modules. Cost outputs were generated for annual
production volumes of 50K, 125K, 250K, and 450K packs. The cost outputs for the 450K
production volume are used in the FRM analysis, consistent with the proposal, as being
applicable in MY 2017 (HEVs) and MY 2025 (EVs and PHEVs).
For engineering properties and material costs, and for other parameters not identified
below, EPA used the defaults provided in the model.
For design guidelines regarding pack voltage and cell capacity, EPA chose guidelines
based on knowledge of current practices and developing trends of battery manufacturers and
OEMs, supplemented by discussions with the BatPaC authors. Specifically: (1) allowable
pack voltage was targeted to approximately 120V for HEVs and approximately 350-400V for
EVs and PHEVs (with some EV150 packs for larger vehicles allowed to about 460-600V); (2)
allowable cell capacity was limited to less than approximately 80 A-hr.
EPA made several modeling assumptions that differed from the default model: (1) The
SOC window for HEVs was increased to 40% rather than the default 25%. (2) HEV packs
were modeled as air cooled instead of liquid cooled (except for Truck and MPV with Towing,
which are modeled as liquid-cooled). EPA replaced the model's projected costs for air cooling
components (blower motor, ducting, and temperature feedback) with costs derived from
FEV's teardown studies, which may be more representative of volume production than the
default values provided in the model.
Additionally, EPA did not include warranty costs computed by BatPaC in the total
battery cost because these are accounted for elsewhere by means of indirect cost multipliers
(ICMs).
Table 3-68 Summary of Inputs and Assumptions Used with BatPaC
Category of
BatPaC Default or Suggested
3-154
Agency Inputs for FRM Analysis
-------
Technologies Considered in the Agencies' Analysis
input/Assumptions
Annual production volume
Battery chemistry
Allowable pack voltage
Allowable cell capacity
Cells per module
SOC window for HEVs
Thermal management
Values
n/a
n/a
forHEV: 1 60-260 V
for PHEV, EV: 290-360 V
< 60 A-hr
16-32
25%
Liquid
450,000
forHEV, PHEV20: LMO-G
for PHEV40, EV: NMC441-G
forHEV:- 120V
for PHEV, EV: ~ 360-600 V
< 80 A-hr
32
40%
Air, for small/medium HEVs
Liquid for all others
The cost projections produced by BatPaC are sensitive to the inputs and assumptions
the user provides. Significant uncertainty remains regarding which will best represent
manufacturer practice in the year 2020. The battery pack cost projection from BatPaC model
ranges from $161/kWh for EV150 truck to $296/kWh for PHEV40 large car with NMC as
chemistry and to $373/kWh for PHEV20 small car as shown in Table 3-69 to Table 3-74. The
agencies note that costs used in the analysis are lower than the costs generally reported in
stakeholder meetings, which ranged from $300/kW-hour to $400/kW-hour range for 2020 and
$250 to $300/kW-hour range for 2025. A comparison of BatPaC modeling results to the costs
used in the 2012-2016 final rule and to cost estimates compiled by EPA from battery suppliers
and auto OEMs is shown in Figure 3-24.
The agencies also reviewed publically available PHEV and EV battery cost literature
including reports from Anderman67, Frost & Sullivan68, TIAX69, Boston Consulting Group70,
and NRC71. Due to the uncertainties inherent in estimating battery costs through the MY 2025
model year, a sensitivity analysis will be provided in each agency's RIA using a range of
costs estimated by DOE technical experts to represent a reasonable outer bounds to the results
from the BatPaC model. In a recent report to NHTSA and EPA, DOE and ANL suggested the
following range for the sensitivity study with 95% confidence interval after analyzing the
confidence bound using the BatPaC model. The agencies describe their respective sensitivities
surrounding BatPaC costs in their respective RIAs (see Chapter 3.11 of EPA's final RIA and
Chapter X of NHTSA's FRIA).
Suggested confidence bounds as percentage of the calculated
point estimate for a graphite based Li-ion battery using the
default inputs in BatPaC
Battery type
HEV
PHEV, EV
PHEV, EV
Cathodes
LMO, LFP, NCA,
NMC
NMC, NCA
LMO, LFP
Confidence Interval
lower
-10%
-10%
-20%
upper
10%
20%
35%
3-155
-------
Technologies Considered in the Agencies' Analysis
Figure 3-23 Table from ANL Recommendation
72
Estimated Battery Pack Costs
$1,200.00
$1,000.00
3" $800.00
— $600.00
D $400.00
w
$200.00
Total Range of OEM Supplier Estimates (>5sources) collected by EPA from 2008 -2010
| Range from the majority of OEM stakeholder meetings in June-August 2010
D2012-2016 EPACost Estimate
OBatPac Model (PHEV20]
4-BatPac Model (EV75]
* BatPac Model (PHEV40]
ABatPac Model (EV150)
D BatPac Model (EV100)
I
±
+
2008
2010
2012
2014
2016
2018 2020
Calendar Year
2022
2024
2026
2028
2030
Figure 3-24 Comparison of direct manufacturing costs per unit of energy storage (S/kW-hr) between the
estimates used by EPA in the 2012-2016 GHG final rule, the BatPaC model results for PHEV20, PHEV40,
EV75, EV100 and EV150 packages compared to estimates from OEM battery suppliers (2009 dollars,
markups not included). Multiple points shown for the BatPaC model results for PHEV 20, PHEV40,
EV75, EV100 and EV150 reflect the range of energy-specific costs for EPA's subcompact through large-
car package categories (see Table 3-70 through Table 3-74for details). A range of OEM estimated battery
costs from stakeholder meetings is also shown for comparison (red bars) which may or may not reflect
additional cost markups.
While it is expected that other Li-ion battery chemistries with higher energy density,
higher power density and lower cost will likely be available in the 2017-2025 timeframe, the
specific chemistries used for the cost analysis were chosen due to their known characteristics
and to be consistent with both public available information on current and near term HEV,
PHEV and EV product offerings from Hyundai, GM and Nissan as well as confidential
business information on future products currently under development.73'74'75'76 The cost
outputs from the BatPaC model used by the agencies in this analysis are shown in Table 3-69
through Table 3-74 for different levels of applied weight reduction technology. We
differentiate between "applied" weight reduction and "net" weight reduction in this analysis
because to achieve the same amount of mass reduction, more mass reduction technologies
might need to be applied to vehicles with electrification than with traditional powertrains
because of the added weight of the electrification systems (i.e., the battery, electric motors,
etc.). This also makes it clear that we have estimated vehicle level battery pack costs—and
motor and other electrified vehicle specific costs—based on the net weight reduction of the
3-156
-------
Technologies Considered in the Agencies' Analysis
vehicle. For example, a typical EV150 battery back and associated motors and other EV-
specific equipment increases vehicle weight roughly 18 percent. As a result, an EV150 that
applied 20 percent mass reduction technology (see section 3.4.5.5 for a full discussion of mass
reduction technologies and costs) would have a net weight reduction of only 2 percent. In
such a case, the agencies would estimate mass reduction costs associated with a 20 percent
applied mass reduction, and EV150 costs associated with only a 2 percent net mass reduction
(lower net mass reduction results in higher battery pack and motor costs). Similarly, HEV
battery packs increase vehicle weight by roughly 5 or 6 percent. Therefore, for an HEV with
20 percent applied mass reduction technology—and costs associated with 20 percent applied
mass reduction—would have HEV costs associated with a 15 percent net mass reduction.
Furthermore, such an HEV would have an effectiveness level improvement associated with a
15 percent net mass reduction rather than a 20 percent net reduction.
Table 3-69 Direct Manufacturing Costs for P2 HEV battery packs at different levels of applied vehicle
weight reduction (2010 dollars, markups not included)
P2 HEV (LMO)
@ 450K/yr
volume
0% weight
reduction
Pack
$/kWh
2% weight
reduction
Pack
$/kWh
7.5% weight
reduction
Pack
$/kWh
10% weight
reduction
Pack
$/kWh
20% weight
reduction
Pack
$/kWh
2008 Baseline
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
$726
$801
$938
$779
$876
$1,010
$896
$804
$809
$747
$682
$676
$722
$796
$929
$775
$870
$1,003
$909
$815
$817
$758
$691
$685
$712
$783
$909
$762
$853
$983
$950
$849
$848
$790
$718
$711
$708
$777
$900
$757
$846
$974
$970
$866
$862
$806
$731
$724
$700
$765
$882
$746
$830
$957
$1,008
$901
$894
$839
$760
$747
20 10 Baseline
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
$732
$809
$950
$788
$878
$1,019
$904
$813
$819
$756
$683
$682
$729
$805
$943
$784
$872
$1,012
$918
$824
$830
$767
$692
$691
$718
$791
$920
$771
$855
$992
$958
$858
$858
$800
$720
$718
$714
$785
$911
$765
$847
$983
$978
$875
$873
$816
$733
$731
$705
$773
$893
$754
$832
$967
$1,017
$909
$904
$848
$762
$754
Table 3-70 Direct Manufacturing Costs for PHEV20 battery packs at different levels of applied vehicle
weight reduction (2010 dollars, markups not included)
PHEV20
(LMO)
@ 450K/yr
volume
0% weight
reduction
Pack
$/kWh
2% weight
reduction
Pack
$/kWh
7.5% weight
reduction
Pack
$/kWh
10% weight
reduction
Pack
$/kWh
20% weight
reduction
Pack
$/kWh
2008 Baseline
Small Car
Standard Car
$2,531
$2,962
$364
$347
$2,517
$2,938
$364
$348
$2,469
$2,835
$370
$345
$2,447
$2,808
$371
$346
$2,431
$2,784
$373
$347
3-157
-------
Technologies Considered in the Agencies' Analysis
Large Car
Small MPV
Large MPV
Truck
$3,734
$2,835
$3,424
$3,874
$368
$316
$300
$295
$3,696
$2,813
$3,393
$3,834
$369
$317
$301
$295
$3,592
$2,754
$3,309
$3,732
$369
$319
$302
$295
$3,546
$2,730
$3,274
$3,681
$368
$320
$303
$297
$3,510
$2,703
$3,244
$3,671
$369
$323
$303
$296
20 10 Baseline
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
$2,572
$3,019
$3,813
$2,933
$3,434
$3,922
$370
$353
$376
$326
$301
$298
$2,554
$2,992
$3,773
$2,911
$3,403
$3,881
$370
$354
$376
$328
$302
$298
$2,507
$2,927
$3,668
$2,811
$3,319
$3,778
$376
$357
$376
$326
$303
$299
$2,487
$2,858
$3,621
$2,783
$3,282
$3,732
$377
$352
$376
$326
$303
$301
$2,468
$2,829
$3,575
$2,754
$3,253
$3,706
$379
$353
$376
$329
$304
$298
Table 3-71 Direct Manufacturing Costs for PHEV40 battery pack at
weight reduction (2010 dollars, markups not
different levels of applied vehicle
included)
PHEV40
(NMC)
@ 450K/yr
volume
0% weight
reduction
Pack
$/kWh
2% weight
reduction
Pack
$/kWh
7.5% weight
reduction
Pack
$/kWh
10% weight
reduction
Pack
$/kWh
20% weight
reduction
Pack
$/kWh
2008 Baseline
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
$3,644
$4,390
$6,006
$4,247
$5,269
$6,122
$262
$257
$296
$236
$231
$233
$3,619
$4,343
$5,921
$4,207
$5,212
$6,050
$262
$257
$295
$237
$231
$233
$3,542
$4,228
$5,671
$4,101
$5,065
$5,900
$264
$258
$291
$238
$231
$232
$3,542
$4,228
$5,671
$4,100
$5,065
$5,900
$264
$258
$291
$237
$231
$232
$3,542
$4,228
$5,671
$4,100
$5,065
$5,900
$264
$258
$291
$237
$231
$232
20 10 Baseline
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
$3,722
$4,494
$6,158
$4,351
$5,286
$6,215
$268
$263
$304
$242
$232
$236
$3,690
$4,447
$6,073
$4,309
$5,228
$6,142
$267
$263
$303
$243
$232
$236
$3,606
$4,324
$5,850
$4,198
$5,080
$5,980
$269
$263
$300
$243
$232
$235
$3,606
$4,324
$5,850
$4,198
$5,080
$5,980
$269
$263
$300
$243
$232
$235
$3,606
$4,324
$5,850
$4,198
$5,080
$5,980
$269
$263
$300
$243
$232
$235
Table 3-72 Direct Manufacturing Costs for EV75 battery packs at different levels of applied vehicle
weight reduction (2010 dollars, markups not included)
EV75 (NMC)
@ 450K/yr
volume
0% weight
reduction
Pack $/kWh
2% weight
reduction
Pack $/kWh
7.5% weight
reduction
Pack $/kWh
10% weight
reduction
Pack $/kWh
20% weight
reduction
Pack $/kWh
3-158
-------
Technologies Considered in the Agencies' Analysis
2008 Baseline
Small Car
Standard Car
Large Car
Small MPV
Large MPV
$5,115
$6,021
$7,724
$5,995
$7,310
$224
$215
$232
$203
$195
$5,098
$5,965
$7,635
$5,952
$7,237
$225
$215
$232
$204
$196
$4,996
$5,818
$7,397
$5,843
$7,045
$228
$216
$231
$206
$196
$4,962
$5,755
$7,295
$5,800
$6,963
$229
$216
$231
$207
$196
$4,768
$5,509
$6,907
$5,625
$6,610
$233
$219
$231
$211
$197
Truck
8,332 $193 $8,242 $193 $8,005 $193 $7,883 $194 $7,474
$194
20 10 Baseline
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
$5,232
$6,152
$7,923
$6,070
$7,312
$8,472
$221
$214
$229
$203
$197
$191
$5,195
$6,092
$7,832
$6,016
$7,238
$8,380
$222
$214
$229
$203
$198
$191
$5,106
$5,940
$7,586
$5,904
$7,046
$8,141
$225
$215
$229
$205
$198
$191
$5,071
$5,874
$7,479
$5,860
$6,962
$8,036
$226
$215
$228
$206
$198
$191
$4,912
$5,624
$7,092
$5,684
$6,605
$7,629
$231
$218
$228
$210
$198
$191
Table 3-73 Direct Manufacturing Costs for EV100 battery packs at different levels of applied vehicle
weight reduction (2010 dollars, markups not included)
EV100 (NMC)
@ 450K/yr
volume
0% weight
reduction
Pack
$/kWh
2% weight
reduction
Pack
$/kWh
7. 5% weight
reduction
Pack
$/kWh
10% weight
reduction
Pack
$/kWh
20% weight
reduction
Pack
$/kWh
2008 Baseline
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
$6,105
$7,054
$8,630
$7,293
$8,641
$9,962
$201
$189
$195
$186
$173
$173
$6,083
$7,001
$8,535
$7,237
$8,571
$9,879
$201
$189
$195
$186
$174
$174
$5,950
$6,826
$8,283
$7,096
$8,392
$9,676
$204
$190
$194
$188
$175
$175
$5,906
$6,770
$8,175
$7,039
$8,321
$9,554
$205
$191
$194
$189
$176
$176
$5,817
$6,662
$7,999
$6,953
$8,215
$9,392
$206
$192
$194
$190
$177
$177
20 10 Baseline
Small Car
Standard Car
Large Car
Small MPV
Large MPV
Truck
$6,255
$7,173
$8,863
$7,375
$8,586
$10,158
$198
$187
$192
$185
$174
$172
$6,209
$7,118
$8,765
$7,318
$8,516
$10,075
$199
$188
$192
$185
$174
$172
$6,094
$6,980
$8,504
$7,174
$8,338
$9,865
$201
$190
$192
$187
$176
$174
$6,048
$6,884
$8,393
$7,117
$8,268
$9,782
$202
$189
$192
$188
$176
$174
$5,956
$6,802
$8,251
$7,031
$8,128
$9,615
$204
$190
$192
$189
$177
$175
3-159
-------
Technologies Considered in the Agencies' Analysis
Table 3-74 Direct Manufacturing Costs for EV150 battery packs at different levels of applied vehicle
weight reduction (2010 dollars, markups not included)
EV150
(NMQ
@
450K/yr
volume
0% weight
reduction
Pack
$/kWh
2% weight
reduction
Pack
$/kWh
7.5% weight
reduction
Pack
$/kWh
10% weight
reduction
Pack
$/kWh
20% weight
reduction
Pack
$/kWh
2008 Baseline
Small
Car
Standard
Car
Large
Car
Small
MPV
Large
MPV
Truck
$8,080
$9,753
$11,120
$10,109
$12,114
$13,878
$177
$174
$167
$171
$162
$161
$8,048
$9,714
$11,073
$10,109
$12,112
$13,818
$178
$174
$167
$171
$162
$161
$8,048
$9,714
$11,073
$10,109
$12,112
$13,759
$178
$174
$167
$171
$162
$161
$8,048
$9,714
$11,073
$10,109
$12,112
$13,759
$178
$174
$167
$171
$162
$161
$8,048
$9,714
$11,073
$10,109
$12,112
$13,759
$178
$174
$167
$171
$162
$161
20 10 Baseline
Small
Car
Standard
Car
Large
Car
Small
MPV
Large
MPV
Truck
$8,298
$9,928
$11,432
$10,228
$12,032
$14,166
$175
$173
$166
$171
$162
$160
$8,265
$9,888
$11,384
$10,228
$11,981
$14,045
$176
$173
$166
$171
$163
$160
$8,265
$9,888
$11,384
$10,228
$11,981
$14,044
$176
$173
$166
$171
$163
$160
$8,265
$9,888
$11,384
$10,228
$11,981
$14,044
$176
$173
$166
$171
$163
$160
$8,265
$9,888
$11,384
$10,228
$11,981
$14,044
$176
$173
$166
$171
$163
$160
Specifically for modeling purposes, both agencies wanted HEV/PHEV/EV battery
pack costs based on net weight reduction rather than applied weight reduction as shown in
Table 3-69 through Table 3-74 above. The agencies did this by first determining the average
weight differences (applied weight reduction vs net weight reduction) for each of the 6 major
vehicle classes (small car, standard car, large car, small MPV, large MPV & truck) and each
of the electrification types (P2 HEV, PHEV & EV). Due to the weight increases of adding the
electrification system and battery pack and the weight decreases by applying smaller or no
conventional internal combustion engine, the net mass reduction for HEV, PHEV and EV
varies for different electrification packages and vehicle classes. For example, for a 20-mile
small car PHEV, a 5% mass reduction of the glider is offset by the additional weight of the
electrification system. Said another way, a 5% mass reduction needs to be applied to the
glider to achieve a net 0% overall vehicle mass reduction for a PHEV20 small car. Those
weight reduction differences are shown in Table 3-75.
Table 3-75 EPA and NHTSA Weight Reduction Offset Associated with Electrification Technologies
Vehicle Class
P2HEV
PHEV20
PHEV40
EV75
EV100
EV150
2008 Baseline
Small car
Standard car
Large car
Small MPV
Large MPV
5%
5%
5%
5%
5%
7%
7%
8%
7%
7%
13%
12%
14%
12%
12%
0%
0%
-1%
1%
0%
6%
6%
5%
7%
6%
18%
18%
17%
19%
18%
3-160
-------
Technologies Considered in the Agencies' Analysis
Truck
4%
7%
12%
1%
7%
19%
20 10 Baseline
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
5%
5%
5%
5%
5%
4%
7%
7%
8%
7%
7%
7%
12%
12%
13%
12%
12%
12%
0%
1%
0%
1%
1%
0%
6%
7%
6%
7%
7%
6%
19%
19%
17%
19%
19%
19%
Notes:
For example, PHEV40-specific technologies add 12-14% to vehicle weight so that a 20%
applied weight reduction would result in a 6-8% net weight reduction.
While an EV75 can actually reduce vehicle weight by 1-2% (i.e., battery packs and motors
weigh less than the removed internal combustion engine and transmission), the agencies used a
value of 0% where negative entries are shown.
The agencies then generated linear regressions of battery pack costs against percentage
net weight reduction using the costs shown in Table 3-69 through Table 3-74 and the weight
reduction offsets shown in Table 3-75. These results are shown in Table 3-76.
Table 3-76 EPA and NHTSA Linear Regressions of Battery Pack Direct Manufacturing Costs vs Net
Weight Reduction (2010$)
Vehicle
Class
P2HEV
PHEV20
PHEV40
EV75
EV100
EV150
2008 Baseline
Small car
Standard
car
Large car
Small
MPV
Large
MPV
Truck
-$181x+$726
-$240x+$801
-$369x+$937
-$224x+$779
-$303x+$876
-$367x+$l,010
-$861x+$2,533
-$l,543x+$2,962
-$l,881x+$3,734
-$l,073x+$2,835
-$l,517x+$3,646
-$2,195x+$4,389
-$4,700x+$6,010
-$l,957x+$4,247
-$l,859x+$5,131
-$2,754x+$6,023
-$4,356x+$7,725
-$2,061x+$5,997
-$2,168x+$6,115
-$2,958x+$7,056
-$4,647x+$8,630
-$2,649x+$7,293
-$2,045x+$8,080
-$2,552x+$9,753
-$2,840x+$l 1,120
-$19x+$10,109
20 10 Baseline
Small car
Standard
car
Large car
Small
MPV
Large
MPV
Truck
-$188x+$733
-$248x+$810
-$387x+$950
-$233x+$789
-$305x+$878
-$364x+$l,019
-$866x+$2,572
-$l,573x+$3,024
-$l,957x+$3,813
-$l,516x+$2,934
-$l,612x+$3,722
-$2,291x+$4,494
-$4,217x+$6,158
-$2,022x+$4,350
-$l,717x+$5,233
-$2,887x+$6,154
-$4,543x+$7,925
-$2,155x+$6,067
-$2,209x+$6,256
-$2,883x+$7,178
-$4,744x+$8,862
-$2,706x+$7,375
-$2,700x+$8,298
-$3,242x+$9,928
-$4,250x+$l 1,432
-$21x+$10,228
Notes:
"x" in the equations represents the net weight reduction as a percentage, so a small car P2 HEV battery pack (2008 baseline) with a 20%
applied weight reduction and, therefore, a 15% net weight reduction would cost (-$181)x(15%)+$726=$698.
The agencies did not regress PHEV or EV costs for the large MPV and truck vehicle classes since we do not believe these vehicle classes
would use the technologies.
For P2 HEV battery packs, the direct manufacturing costs shown in Table 3-76 are
considered applicable to the 2017MY. The agencies consider the P2 battery packs technology
to be on the flat portion of the learning curve during the 2017-2025 timeframe. The agencies
have applied a highl complexity ICM of 1.56 through 2024 then 1.35 thereafter. For PHEV
3-161
-------
Technologies Considered in the Agencies' Analysis
and EV battery packs, the direct manufacturing costs shown in Table 3-76 are considered
applicable to the 2025MY. For the PHEV and EV battery packs, the agencies have applied
the learning curve discussed in Section 3.2.3. The agencies have applied a high2 complexity
ICM of 1.77 through 2024 then 1.50 thereafter. The resultant costs for P2 HEV, PHEV20,
PHEV40, EV75, EV100 and EV150 battery packs for the 2008 and 2010 baselines are shown
in Table 3-77 through Table 3-87, respectively.
Table 3-77 Costs for P2 HEV Battery Packs for the 2008 Baseline (2010$)
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
Vehicle class
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Applied
WR
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%
Net
WR
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
11%
16%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
11%
16%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
2017
$717
$707
$698
$789
$777
$765
$919
$900
$882
$768
$757
$745
$861
$846
$830
$988
$970
$951
$404
$399
$394
$444
$438
$431
$518
$507
$497
$433
$426
$420
$485
$477
$468
$557
$546
$536
$1,120
$1,106
$1,092
$1,233
$1,214
$1,196
$1,436
$1,407
$1,379
$1,201
$1,183
2018
$695
$686
$677
$765
$753
$742
$891
$873
$855
$745
$734
$723
$835
$820
$805
$958
$941
$923
$402
$397
$392
$443
$436
$429
$516
$506
$495
$431
$425
$419
$483
$475
$466
$555
$545
$534
$1,097
$1,084
$1,070
$1,208
$1,190
$1,171
$1,407
$1,379
$1,350
$1,176
$1,159
2019
$674
$666
$657
$742
$731
$719
$864
$847
$830
$722
$712
$701
$810
$796
$781
$930
$912
$895
$401
$396
$391
$441
$435
$428
$514
$504
$494
$430
$424
$417
$482
$473
$465
$553
$543
$533
$1,075
$1,062
$1,048
$1,183
$1,165
$1,147
$1,378
$1,351
$1,323
$1,152
$1,135
2020
$654
$646
$637
$720
$709
$698
$838
$821
$805
$701
$690
$680
$786
$772
$758
$902
$885
$868
$400
$395
$390
$440
$433
$427
$512
$502
$492
$428
$422
$416
$480
$472
$463
$551
$541
$531
$1,054
$1,040
$1,027
$1,160
$1,142
$1,125
$1,351
$1,324
$1,297
$1,129
$1,113
2021
$634
$626
$618
$698
$688
$677
$813
$797
$781
$680
$670
$660
$762
$749
$735
$875
$858
$842
$399
$393
$388
$439
$432
$425
$511
$501
$490
$427
$421
$415
$479
$470
$462
$549
$539
$529
$1,033
$1,020
$1,007
$1,137
$1,119
$1,102
$1,324
$1,297
$1,271
$1,107
$1,091
2022
$615
$608
$600
$677
$667
$657
$789
$773
$757
$659
$650
$640
$739
$726
$713
$848
$833
$817
$397
$392
$387
$437
$431
$424
$509
$499
$489
$426
$419
$413
$477
$469
$460
$548
$538
$527
$1,013
$1,000
$987
$1,114
$1,098
$1,081
$1,298
$1,272
$1,246
$1,085
$1,069
2023
$597
$589
$582
$657
$647
$637
$765
$750
$734
$640
$630
$621
$717
$704
$692
$823
$808
$792
$396
$391
$386
$436
$429
$423
$508
$498
$487
$424
$418
$412
$476
$467
$459
$546
$536
$526
$993
$980
$968
$1,093
$1,076
$1,060
$1,273
$1,247
$1,222
$1,064
$1,048
2024
$579
$572
$564
$637
$628
$618
$742
$727
$712
$620
$611
$602
$695
$683
$671
$798
$783
$769
$395
$390
$385
$435
$428
$421
$506
$496
$486
$423
$417
$411
$474
$466
$458
$545
$534
$524
$974
$962
$949
$1,072
$1,056
$1,039
$1,248
$1,223
$1,198
$1,044
$1,028
2025
$562
$554
$547
$618
$609
$599
$720
$705
$691
$602
$593
$584
$675
$663
$651
$774
$760
$746
$243
$239
$236
$267
$263
$259
$311
$305
$298
$260
$256
$252
$291
$286
$281
$334
$328
$322
$804
$794
$784
$885
$872
$858
$1,031
$1,010
$989
$862
$849
3-162
-------
Technologies Considered in the Agencies' Analysis
TC
TC
TC
TC
TC
TC
TC
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
20%
10%
15%
20%
10%
15%
20%
15%
5%
10%
15%
6%
11%
16%
$1,165
$1,346
$1,322
$1,298
$1,545
$1,516
$1,487
$1,142
$1,318
$1,295
$1,272
$1,513
$1,485
$1,457
$1,119
$1,292
$1,269
$1,246
$1,483
$1,455
$1,428
$1,096
$1,266
$1,243
$1,221
$1,453
$1,426
$1,399
$1,074
$1,241
$1,219
$1,197
$1,424
$1,398
$1,371
$1,053
$1,216
$1,195
$1,174
$1,396
$1,370
$1,344
$1,033
$1,193
$1,172
$1,151
$1,369
$1,344
$1,318
$1,013
$1,170
$1,149
$1,129
$1,343
$1,318
$1,293
$836
$966
$949
$932
$1,109
$1,088
$1,068
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-78 Costs for P2 HEV Battery Packs for the 2010 Baseline (2010$)
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
Vehicle class
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Applied
WR
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%
Net
WR
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
11%
16%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
11%
16%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
2017
$723
$714
$704
$797
$785
$772
$931
$911
$892
$777
$765
$754
$863
$847
$832
$997
$979
$961
$408
$402
$397
$449
$442
$435
$524
$514
$503
$438
$431
$425
$486
$478
$469
$562
$552
$541
$1,131
$1,116
$1,101
$1,246
$1,227
$1,208
$1,455
$1,425
$1,394
$1,215
2018
$701
$692
$683
$773
$761
$749
$903
$884
$865
$754
$742
$731
$837
$822
$807
$967
$949
$932
$406
$401
$396
$448
$441
$434
$523
$512
$501
$436
$430
$423
$485
$476
$467
$560
$550
$540
$1,108
$1,093
$1,079
$1,221
$1,202
$1,183
$1,425
$1,396
$1,366
$1,190
2019
$680
$672
$663
$750
$738
$727
$876
$857
$839
$731
$720
$709
$812
$797
$783
$938
$921
$904
$405
$400
$394
$446
$439
$432
$521
$510
$499
$435
$428
$422
$483
$474
$466
$558
$548
$538
$1,085
$1,071
$1,057
$1,196
$1,178
$1,159
$1,396
$1,367
$1,338
$1,166
2020
$660
$651
$643
$727
$716
$705
$849
$832
$814
$709
$698
$688
$787
$773
$759
$910
$893
$877
$403
$398
$393
$445
$438
$431
$519
$508
$498
$433
$427
$420
$481
$473
$464
$556
$546
$536
$1,063
$1,050
$1,036
$1,172
$1,154
$1,136
$1,369
$1,340
$1,312
$1,142
2021
$640
$632
$624
$706
$695
$684
$824
$807
$790
$688
$677
$667
$764
$750
$737
$883
$867
$850
$402
$397
$392
$443
$436
$430
$518
$507
$496
$432
$426
$419
$480
$471
$463
$555
$544
$534
$1,042
$1,029
$1,015
$1,149
$1,131
$1,113
$1,341
$1,313
$1,286
$1,120
2022
$621
$613
$605
$685
$674
$663
$799
$782
$766
$667
$657
$647
$741
$728
$715
$856
$841
$825
$401
$396
$391
$442
$435
$428
$516
$505
$494
$431
$424
$418
$478
$470
$461
$553
$543
$533
$1,022
$1,009
$995
$1,126
$1,109
$1,091
$1,315
$1,288
$1,260
$1,098
2023
$602
$595
$587
$664
$654
$643
$775
$759
$743
$647
$637
$628
$719
$706
$693
$831
$815
$800
$400
$395
$389
$441
$434
$427
$514
$504
$493
$429
$423
$417
$477
$468
$460
$551
$541
$531
$1,002
$989
$976
$1,105
$1,087
$1,070
$1,290
$1,263
$1,236
$1,077
2024
$584
$577
$569
$644
$634
$624
$752
$736
$721
$628
$618
$609
$697
$685
$672
$806
$791
$776
$399
$393
$388
$439
$433
$426
$513
$502
$492
$428
$422
$415
$475
$467
$459
$550
$540
$529
$983
$970
$957
$1,083
$1,067
$1,050
$1,265
$1,238
$1,212
$1,056
2025
$567
$559
$552
$625
$615
$605
$729
$714
$699
$609
$600
$591
$676
$664
$652
$781
$767
$753
$245
$242
$238
$270
$266
$261
$315
$308
$302
$263
$259
$255
$292
$287
$282
$338
$331
$325
$812
$801
$790
$895
$881
$867
$1,044
$1,023
$1,001
$872
3-163
-------
Technologies Considered in the Agencies' Analysis
TC
TC
TC
TC
TC
TC
TC
TC
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
5%
10%
15%
6%
11%
16%
$1,196
$1,178
$1,349
$1,325
$1,301
$1,559
$1,531
$1,502
$1,172
$1,154
$1,321
$1,298
$1,275
$1,527
$1,499
$1,471
$1,148
$1,131
$1,295
$1,272
$1,249
$1,496
$1,469
$1,442
$1,125
$1,108
$1,269
$1,246
$1,224
$1,466
$1,440
$1,413
$1,103
$1,086
$1,243
$1,222
$1,200
$1,437
$1,411
$1,385
$1,081
$1,065
$1,219
$1,198
$1,176
$1,409
$1,383
$1,358
$1,060
$1,044
$1,195
$1,174
$1,153
$1,382
$1,356
$1,331
$1,040
$1,024
$1,172
$1,152
$1,131
$1,355
$1,330
$1,306
$859
$846
$968
$951
$934
$1,119
$1,099
$1,078
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-79 Costs for PHEV20 Battery Packs for the 2008 Baseline (2010$)
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
Vehicle
class
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Applied
WR
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%
NetWR
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
2017
$4,896
$4,812
$4,728
$5,696
$5,545
$5,394
$7,219
$7,035
$6,851
$5,474
$5,370
$5,265
$2,106
$2,070
$2,034
$2,450
$2,385
$2,321
$3,105
$3,026
$2,947
$2,355
$2,310
$2,265
$7,003
$6,883
$6,762
$8,146
$7,930
$7,715
$10,324
$10,061
$9,799
$7,829
2018
$3,917
$3,850
$3,783
$4,557
$4,436
$4,315
$5,775
$5,628
$5,481
$4,379
$4,296
$4,212
$2,034
$1,999
$1,964
$2,366
$2,304
$2,241
$2,999
$2,923
$2,846
$2,274
$2,231
$2,187
$5,951
$5,849
$5,747
$6,923
$6,740
$6,556
$8,774
$8,551
$8,327
$6,654
2019
$3,917
$3,850
$3,783
$4,557
$4,436
$4,315
$5,775
$5,628
$5,481
$4,379
$4,296
$4,212
$2,034
$1,999
$1,964
$2,366
$2,304
$2,241
$2,999
$2,923
$2,846
$2,274
$2,231
$2,187
$5,951
$5,849
$5,747
$6,923
$6,740
$6,556
$8,774
$8,551
$8,327
$6,654
2020
$3,134
$3,080
$3,026
$3,645
$3,549
$3,452
$4,620
$4,502
$4,385
$3,504
$3,436
$3,369
$1,977
$1,943
$1,909
$2,299
$2,238
$2,178
$2,914
$2,840
$2,766
$2,210
$2,168
$2,125
$5,110
$5,023
$4,935
$5,944
$5,787
$5,630
$7,534
$7,342
$7,151
$5,713
2021
$3,134
$3,080
$3,026
$3,645
$3,549
$3,452
$4,620
$4,502
$4,385
$3,504
$3,436
$3,369
$1,977
$1,943
$1,909
$2,299
$2,238
$2,178
$2,914
$2,840
$2,766
$2,210
$2,168
$2,125
$5,110
$5,023
$4,935
$5,944
$5,787
$5,630
$7,534
$7,342
$7,151
$5,713
2022
$3,134
$3,080
$3,026
$3,645
$3,549
$3,452
$4,620
$4,502
$4,385
$3,504
$3,436
$3,369
$1,977
$1,943
$1,909
$2,299
$2,238
$2,178
$2,914
$2,840
$2,766
$2,210
$2,168
$2,125
$5,110
$5,023
$4,935
$5,944
$5,787
$5,630
$7,534
$7,342
$7,151
$5,713
2023
$3,134
$3,080
$3,026
$3,645
$3,549
$3,452
$4,620
$4,502
$4,385
$3,504
$3,436
$3,369
$1,977
$1,943
$1,909
$2,299
$2,238
$2,178
$2,914
$2,840
$2,766
$2,210
$2,168
$2,125
$5,110
$5,023
$4,935
$5,944
$5,787
$5,630
$7,534
$7,342
$7,151
$5,713
2024
$3,134
$3,080
$3,026
$3,645
$3,549
$3,452
$4,620
$4,502
$4,385
$3,504
$3,436
$3,369
$1,977
$1,943
$1,909
$2,299
$2,238
$2,178
$2,914
$2,840
$2,766
$2,210
$2,168
$2,125
$5,110
$5,023
$4,935
$5,944
$5,787
$5,630
$7,534
$7,342
$7,151
$5,713
2025
$2,507
$2,464
$2,421
$2,916
$2,839
$2,762
$3,696
$3,602
$3,508
$2,803
$2,749
$2,696
$1,245
$1,224
$1,202
$1,448
$1,410
$1,372
$1,835
$1,789
$1,742
$1,392
$1,365
$1,339
$3,752
$3,688
$3,623
$4,364
$4,249
$4,133
$5,531
$5,391
$5,250
$4,195
3-164
-------
Technologies Considered in the Agencies' Analysis
TC
TC
Small MPV
Small MPV
15%
20%
8%
13%
$7,679
$7,530
$6,526
$6,399
$6,526
$6,399
$5,604
$5,495
$5,604
$5,495
$5,604
$5,495
$5,604
$5,495
$5,604
$5,495
$4,114
$4,034
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-80 Costs for PHEV20 Battery Packs for the 2010 Baseline (2010$)
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
Vehicle
class
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Applied
WR
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%
NetWR
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
2017
$4,973
$4,888
$4,804
$5,815
$5,661
$5,508
$7,371
$7,180
$6,989
$5,643
$5,494
$5,346
$2,139
$2,103
$2,066
$2,501
$2,435
$2,369
$3,171
$3,089
$3,007
$2,427
$2,364
$2,300
$7,112
$6,991
$6,870
$8,316
$8,097
$7,877
$10,542
$10,269
$9,996
$8,070
$7,858
$7,646
2018
$3,978
$3,910
$3,843
$4,652
$4,529
$4,406
$5,897
$5,744
$5,591
$4,514
$4,396
$4,277
$2,066
$2,031
$1,996
$2,416
$2,352
$2,288
$3,063
$2,983
$2,904
$2,344
$2,283
$2,221
$6,044
$5,941
$5,838
$7,068
$6,881
$6,694
$8,960
$8,727
$8,495
$6,858
$6,678
$6,498
2019
$3,978
$3,910
$3,843
$4,652
$4,529
$4,406
$5,897
$5,744
$5,591
$4,514
$4,396
$4,277
$2,066
$2,031
$1,996
$2,416
$2,352
$2,288
$3,063
$2,983
$2,904
$2,344
$2,283
$2,221
$6,044
$5,941
$5,838
$7,068
$6,881
$6,694
$8,960
$8,727
$8,495
$6,858
$6,678
$6,498
2020
$3,182
$3,128
$3,074
$3,722
$3,623
$3,525
$4,718
$4,595
$4,473
$3,611
$3,516
$3,422
$2,007
$1,973
$1,939
$2,347
$2,285
$2,223
$2,976
$2,899
$2,821
$2,278
$2,218
$2,158
$5,190
$5,102
$5,013
$6,069
$5,908
$5,748
$7,693
$7,494
$7,294
$5,889
$5,734
$5,580
2021
$3,182
$3,128
$3,074
$3,722
$3,623
$3,525
$4,718
$4,595
$4,473
$3,611
$3,516
$3,422
$2,007
$1,973
$1,939
$2,347
$2,285
$2,223
$2,976
$2,899
$2,821
$2,278
$2,218
$2,158
$5,190
$5,102
$5,013
$6,069
$5,908
$5,748
$7,693
$7,494
$7,294
$5,889
$5,734
$5,580
2022
$3,182
$3,128
$3,074
$3,722
$3,623
$3,525
$4,718
$4,595
$4,473
$3,611
$3,516
$3,422
$2,007
$1,973
$1,939
$2,347
$2,285
$2,223
$2,976
$2,899
$2,821
$2,278
$2,218
$2,158
$5,190
$5,102
$5,013
$6,069
$5,908
$5,748
$7,693
$7,494
$7,294
$5,889
$5,734
$5,580
2023
$3,182
$3,128
$3,074
$3,722
$3,623
$3,525
$4,718
$4,595
$4,473
$3,611
$3,516
$3,422
$2,007
$1,973
$1,939
$2,347
$2,285
$2,223
$2,976
$2,899
$2,821
$2,278
$2,218
$2,158
$5,190
$5,102
$5,013
$6,069
$5,908
$5,748
$7,693
$7,494
$7,294
$5,889
$5,734
$5,580
2024
$3,182
$3,128
$3,074
$3,722
$3,623
$3,525
$4,718
$4,595
$4,473
$3,611
$3,516
$3,422
$2,007
$1,973
$1,939
$2,347
$2,285
$2,223
$2,976
$2,899
$2,821
$2,278
$2,218
$2,158
$5,190
$5,102
$5,013
$6,069
$5,908
$5,748
$7,693
$7,494
$7,294
$5,889
$5,734
$5,580
2025
$2,546
$2,503
$2,459
$2,977
$2,899
$2,820
$3,774
$3,676
$3,578
$2,889
$2,813
$2,737
$1,264
$1,243
$1,221
$1,479
$1,439
$1,400
$1,874
$1,826
$1,777
$1,435
$1,397
$1,359
$3,810
$3,746
$3,681
$4,456
$4,338
$4,220
$5,648
$5,502
$5,356
$4,324
$4,210
$4,097
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3-165
-------
Technologies Considered in the Agencies' Analysis
Table 3-81 Costs for PHEV40 Battery Packs for the 2008 Baseline (2010$)
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
Vehicle
class
Small car
Small car
Standard
car
Standard
car
Large car
Large car
Small
MPV
Small
MPV
Small car
Small car
Standard
car
Standard
car
Large car
Large car
Small
MPV
Small
MPV
Small car
Small car
Standard
car
Standard
car
Large car
Large car
Small
MPV
Small
MPV
Applied
WR
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
Net
WR
2%
7%
3%
8%
1%
6%
3%
8%
2%
7%
3%
8%
1%
6%
3%
8%
2%
7%
3%
8%
1%
6%
3%
8%
2017
$7,063
$6,915
$8,443
$8,229
$11,646
$11,187
$8,179
$7,988
$3,038
$2,975
$3,632
$3,540
$5,010
$4,813
$3,519
$3,436
$10,101
$9,889
$12,075
$11,769
$16,656
$16,000
$11,698
$11,425
2018
$5,650
$5,532
$6,754
$6,583
$9,317
$8,950
$6,544
$6,391
$2,934
$2,873
$3,508
$3,419
$4,838
$4,648
$3,398
$3,319
$8,584
$8,404
$10,262
$10,002
$14,155
$13,597
$9,942
$9,709
2019
$5,650
$5,532
$6,754
$6,583
$9,317
$8,950
$6,544
$6,391
$2,934
$2,873
$3,508
$3,419
$4,838
$4,648
$3,398
$3,319
$8,584
$8,404
$10,262
$10,002
$14,155
$13,597
$9,942
$9,709
2020
$4,520
$4,425
$5,404
$5,266
$7,453
$7,160
$5,235
$5,113
$2,851
$2,791
$3,408
$3,322
$4,701
$4,516
$3,302
$3,225
$7,371
$7,217
$8,812
$8,588
$12,155
$11,676
$8,537
$8,337
2021
$4,520
$4,425
$5,404
$5,266
$7,453
$7,160
$5,235
$5,113
$2,851
$2,791
$3,408
$3,322
$4,701
$4,516
$3,302
$3,225
$7,371
$7,217
$8,812
$8,588
$12,155
$11,676
$8,537
$8,337
2022
$4,520
$4,425
$5,404
$5,266
$7,453
$7,160
$5,235
$5,113
$2,851
$2,791
$3,408
$3,322
$4,701
$4,516
$3,302
$3,225
$7,371
$7,217
$8,812
$8,588
$12,155
$11,676
$8,537
$8,337
2023
$4,520
$4,425
$5,404
$5,266
$7,453
$7,160
$5,235
$5,113
$2,851
$2,791
$3,408
$3,322
$4,701
$4,516
$3,302
$3,225
$7,371
$7,217
$8,812
$8,588
$12,155
$11,676
$8,537
$8,337
2024
$4,520
$4,425
$5,404
$5,266
$7,453
$7,160
$5,235
$5,113
$2,851
$2,791
$3,408
$3,322
$4,701
$4,516
$3,302
$3,225
$7,371
$7,217
$8,812
$8,588
$12,155
$11,676
$8,537
$8,337
2025
$3,616
$3,540
$4,323
$4,213
$5,963
$5,728
$4,188
$4,090
$1,796
$1,758
$2,147
$2,092
$2,961
$2,844
$2,080
$2,031
$5,412
$5,298
$6,470
$6,305
$8,924
$8,572
$6,268
$6,121
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-82 Costs for PHEV40 Battery Packs for the 2010 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
Vehicle
class
Small car
Small car
Standard
car
Standard
car
Large car
Applied
WR
15%
20%
15%
20%
15%
Net
WR
3%
8%
3%
8%
2%
2017
$7,175
$7,018
$8,642
$8,418
$11,862
2018
$5,740
$5,614
$6,914
$6,735
$9,490
2019
$5,740
$5,614
$6,914
$6,735
$9,490
2020
$4,592
$4,491
$5,531
$5,388
$7,592
2021
$4,592
$4,491
$5,531
$5,388
$7,592
2022
$4,592
$4,491
$5,531
$5,388
$7,592
2023
$4,592
$4,491
$5,531
$5,388
$7,592
2024
$4,592
$4,491
$5,531
$5,388
$7,592
2025
$3,674
$3,593
$4,425
$4,310
$6,073
3-166
-------
Technologies Considered in the Agencies' Analysis
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
Large car
Small
MPV
Small
MPV
Small car
Small car
Standard
car
Standard
car
Large car
Large car
Small
MPV
Small
MPV
Small car
Small car
Standard
car
Standard
car
Large car
Large car
Small
MPV
Small
MPV
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
7%
3%
8%
3%
8%
3%
8%
2%
7%
3%
8%
3%
8%
3%
8%
2%
7%
3%
8%
$11,450
$8,378
$8,180
$3,087
$3,019
$3,718
$3,622
$5,103
$4,926
$3,604
$3,519
$10,262
$10,037
$12,360
$12,040
$16,965
$16,376
$11,982
$11,700
$9,160
$6,702
$6,544
$2,981
$2,916
$3,591
$3,498
$4,928
$4,757
$3,481
$3,399
$8,721
$8,530
$10,504
$10,232
$14,418
$13,918
$10,183
$9,943
$9,160
$6,702
$6,544
$2,981
$2,916
$3,591
$3,498
$4,928
$4,757
$3,481
$3,399
$8,721
$8,530
$10,504
$10,232
$14,418
$13,918
$10,183
$9,943
$7,328
$5,362
$5,235
$2,896
$2,833
$3,489
$3,398
$4,789
$4,622
$3,382
$3,302
$7,488
$7,324
$9,020
$8,786
$12,380
$11,951
$8,744
$8,538
$7,328
$5,362
$5,235
$2,896
$2,833
$3,489
$3,398
$4,789
$4,622
$3,382
$3,302
$7,488
$7,324
$9,020
$8,786
$12,380
$11,951
$8,744
$8,538
$7,328
$5,362
$5,235
$2,896
$2,833
$3,489
$3,398
$4,789
$4,622
$3,382
$3,302
$7,488
$7,324
$9,020
$8,786
$12,380
$11,951
$8,744
$8,538
$7,328
$5,362
$5,235
$2,896
$2,833
$3,489
$3,398
$4,789
$4,622
$3,382
$3,302
$7,488
$7,324
$9,020
$8,786
$12,380
$11,951
$8,744
$8,538
$7,328
$5,362
$5,235
$2,896
$2,833
$3,489
$3,398
$4,789
$4,622
$3,382
$3,302
$7,488
$7,324
$9,020
$8,786
$12,380
$11,951
$8,744
$8,538
$5,863
$4,290
$4,188
$1,824
$1,784
$2,197
$2,141
$3,016
$2,911
$2,130
$2,080
$5,498
$5,377
$6,622
$6,451
$9,090
$8,774
$6,420
$6,268
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-83 Costs for EV75 Battery Packs for the 2008 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
Vehicle
class
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small
MPV
Small
MPV
Small
MPV
Small car
Small car
Applied
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
Net
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
9%
14%
19%
10%
15%
2017
$9,658
$9,476
$9,294
$11,226
$10,957
$10,688
$14,236
$13,811
$13,385
$11,350
$11,149
$10,947
$4,155
$4,076
2018
$7,726
$7,581
$7,436
$8,980
$8,765
$8,550
$11,389
$11,049
$10,708
$9,080
$8,919
$8,758
$4,012
$3,937
2019
$7,726
$7,581
$7,436
$8,980
$8,765
$8,550
$11,389
$11,049
$10,708
$9,080
$8,919
$8,758
$4,012
$3,937
2020
$6,181
$6,065
$5,948
$7,184
$7,012
$6,840
$9,111
$8,839
$8,567
$7,264
$7,135
$7,006
$3,899
$3,825
2021
$6,181
$6,065
$5,948
$7,184
$7,012
$6,840
$9,111
$8,839
$8,567
$7,264
$7,135
$7,006
$3,899
$3,825
2022
$6,181
$6,065
$5,948
$7,184
$7,012
$6,840
$9,111
$8,839
$8,567
$7,264
$7,135
$7,006
$3,899
$3,825
2023
$6,181
$6,065
$5,948
$7,184
$7,012
$6,840
$9,111
$8,839
$8,567
$7,264
$7,135
$7,006
$3,899
$3,825
2024
$6,181
$6,065
$5,948
$7,184
$7,012
$6,840
$9,111
$8,839
$8,567
$7,264
$7,135
$7,006
$3,899
$3,825
2025
$4,945
$4,852
$4,759
$5,747
$5,610
$5,472
$7,289
$7,071
$6,853
$5,811
$5,708
$5,605
$2,456
$2,409
3-167
-------
Technologies Considered in the Agencies' Analysis
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small
MPV
Small
MPV
Small
MPV
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small
MPV
Small
MPV
Small
MPV
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
20%
10%
15%
20%
10%
15%
20%
9%
14%
19%
10%
15%
20%
10%
15%
20%
10%
15%
20%
9%
14%
19%
$3,998
$4,829
$4,713
$4,598
$6,124
$5,941
$5,758
$4,883
$4,796
$4,709
$13,812
$13,552
$13,293
$16,055
$15,670
$15,285
$20,360
$19,752
$19,144
$16,232
$15,945
$15,657
$3,861
$4,664
$4,552
$4,440
$5,915
$5,738
$5,561
$4,715
$4,632
$4,548
$11,738
$11,518
$11,297
$13,644
$13,317
$12,990
$17,303
$16,786
$16,269
$13,795
$13,551
$13,306
$3,861
$4,664
$4,552
$4,440
$5,915
$5,738
$5,561
$4,715
$4,632
$4,548
$11,738
$11,518
$11,297
$13,644
$13,317
$12,990
$17,303
$16,786
$16,269
$13,795
$13,551
$13,306
$3,752
$4,532
$4,423
$4,314
$5,747
$5,575
$5,403
$4,582
$4,501
$4,419
$10,079
$9,890
$9,700
$11,716
$11,435
$11,154
$14,858
$14,414
$13,970
$11,846
$11,636
$11,426
$3,752
$4,532
$4,423
$4,314
$5,747
$5,575
$5,403
$4,582
$4,501
$4,419
$10,079
$9,890
$9,700
$11,716
$11,435
$11,154
$14,858
$14,414
$13,970
$11,846
$11,636
$11,426
$3,752
$4,532
$4,423
$4,314
$5,747
$5,575
$5,403
$4,582
$4,501
$4,419
$10,079
$9,890
$9,700
$11,716
$11,435
$11,154
$14,858
$14,414
$13,970
$11,846
$11,636
$11,426
$3,752
$4,532
$4,423
$4,314
$5,747
$5,575
$5,403
$4,582
$4,501
$4,419
$10,079
$9,890
$9,700
$11,716
$11,435
$11,154
$14,858
$14,414
$13,970
$11,846
$11,636
$11,426
$3,752
$4,532
$4,423
$4,314
$5,747
$5,575
$5,403
$4,582
$4,501
$4,419
$10,079
$9,890
$9,700
$11,716
$11,435
$11,154
$14,858
$14,414
$13,970
$11,846
$11,636
$11,426
$2,363
$2,854
$2,786
$2,717
$3,620
$3,512
$3,403
$2,886
$2,835
$2,784
$7,400
$7,261
$7,122
$8,602
$8,396
$8,190
$10,909
$10,583
$10,257
$8,697
$8,543
$8,389
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-84 Costs for EV75 Battery Packs for the 2010 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
Vehicle
class
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small
MPV
Small
MPV
Small
MPV
Applied
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
Net
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
9%
14%
19%
2017
$9,886
$9,718
$9,551
$11,456
$11,174
$10,892
$14,592
$14,148
$13,704
$11,470
$11,260
$11,049
2018
$7,909
$7,775
$7,640
$9,164
$8,939
$8,713
$11,673
$11,318
$10,964
$9,176
$9,008
$8,840
2019
$7,909
$7,775
$7,640
$9,164
$8,939
$8,713
$11,673
$11,318
$10,964
$9,176
$9,008
$8,840
2020
$6,327
$6,220
$6,112
$7,332
$7,151
$6,971
$9,339
$9,055
$8,771
$7,341
$7,206
$7,072
2021
$6,327
$6,220
$6,112
$7,332
$7,151
$6,971
$9,339
$9,055
$8,771
$7,341
$7,206
$7,072
2022
$6,327
$6,220
$6,112
$7,332
$7,151
$6,971
$9,339
$9,055
$8,771
$7,341
$7,206
$7,072
2023
$6,327
$6,220
$6,112
$7,332
$7,151
$6,971
$9,339
$9,055
$8,771
$7,341
$7,206
$7,072
2024
$6,327
$6,220
$6,112
$7,332
$7,151
$6,971
$9,339
$9,055
$8,771
$7,341
$7,206
$7,072
2025
$5,062
$4,976
$4,890
$5,865
$5,721
$5,577
$7,471
$7,244
$7,017
$5,873
$5,765
$5,657
3-168
-------
Technologies Considered in the Agencies' Analysis
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small
MPV
Small
MPV
Small
MPV
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small
MPV
Small
MPV
Small
MPV
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%
9%
14%
19%
10%
15%
20%
10%
15%
20%
10%
15%
20%
9%
14%
19%
$4,253
$4,181
$4,109
$4,928
$4,807
$4,685
$6,277
$6,086
$5,895
$4,934
$4,844
$4,753
$14,139
$13,899
$13,659
$16,384
$15,980
$15,577
$20,869
$20,234
$19,600
$16,405
$16,104
$15,803
$4,107
$4,038
$3,968
$4,759
$4,642
$4,525
$6,062
$5,878
$5,694
$4,765
$4,678
$4,591
$12,016
$11,812
$11,608
$13,924
$13,581
$13,238
$17,736
$17,196
$16,657
$13,942
$13,686
$13,430
$4,107
$4,038
$3,968
$4,759
$4,642
$4,525
$6,062
$5,878
$5,694
$4,765
$4,678
$4,591
$12,016
$11,812
$11,608
$13,924
$13,581
$13,238
$17,736
$17,196
$16,657
$13,942
$13,686
$13,430
$3,991
$3,923
$3,855
$4,624
$4,511
$4,397
$5,890
$5,711
$5,532
$4,630
$4,545
$4,460
$10,318
$10,143
$9,968
$11,956
$11,662
$11,367
$15,229
$14,766
$14,303
$11,971
$11,752
$11,532
$3,991
$3,923
$3,855
$4,624
$4,511
$4,397
$5,890
$5,711
$5,532
$4,630
$4,545
$4,460
$10,318
$10,143
$9,968
$11,956
$11,662
$11,367
$15,229
$14,766
$14,303
$11,971
$11,752
$11,532
$3,991
$3,923
$3,855
$4,624
$4,511
$4,397
$5,890
$5,711
$5,532
$4,630
$4,545
$4,460
$10,318
$10,143
$9,968
$11,956
$11,662
$11,367
$15,229
$14,766
$14,303
$11,971
$11,752
$11,532
$3,991
$3,923
$3,855
$4,624
$4,511
$4,397
$5,890
$5,711
$5,532
$4,630
$4,545
$4,460
$10,318
$10,143
$9,968
$11,956
$11,662
$11,367
$15,229
$14,766
$14,303
$11,971
$11,752
$11,532
$3,991
$3,923
$3,855
$4,624
$4,511
$4,397
$5,890
$5,711
$5,532
$4,630
$4,545
$4,460
$10,318
$10,143
$9,968
$11,956
$11,662
$11,367
$15,229
$14,766
$14,303
$11,971
$11,752
$11,532
$2,514
$2,471
$2,428
$2,913
$2,841
$2,769
$3,710
$3,597
$3,485
$2,917
$2,863
$2,809
$7,575
$7,447
$7,318
$8,778
$8,562
$8,346
$11,181
$10,841
$10,501
$8,789
$8,628
$8,467
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-85 Costs for EV100 Battery Packs for the 2008 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
Vehicle
class
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Applied
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
Net
WR
4%
9%
14%
4%
9%
14%
5%
10%
15%
3%
2017
$11,774
$11,563
$11,351
$13,550
$13,261
$12,973
$16,403
$15,949
$15,495
$14,089
2018
$9,420
$9,250
$9,081
$10,840
$10,609
$10,378
$13,122
$12,759
$12,396
$11,271
2019
$9,420
$9,250
$9,081
$10,840
$10,609
$10,378
$13,122
$12,759
$12,396
$11,271
2020
$7,536
$7,400
$7,265
$8,672
$8,487
$8,302
$10,498
$10,207
$9,917
$9,017
2021
$7,536
$7,400
$7,265
$8,672
$8,487
$8,302
$10,498
$10,207
$9,917
$9,017
2022
$7,536
$7,400
$7,265
$8,672
$8,487
$8,302
$10,498
$10,207
$9,917
$9,017
2023
$7,536
$7,400
$7,265
$8,672
$8,487
$8,302
$10,498
$10,207
$9,917
$9,017
2024
$7,536
$7,400
$7,265
$8,672
$8,487
$8,302
$10,498
$10,207
$9,917
$9,017
2025
$6,028
$5,920
$5,812
$6,938
$6,790
$6,642
$8,398
$8,166
$7,933
$7,214
3-169
-------
Technologies Considered in the Agencies' Analysis
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
Small MPV
Small MPV
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
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%
8%
13%
4%
9%
14%
4%
9%
14%
5%
10%
15%
3%
8%
13%
4%
9%
14%
4%
9%
14%
5%
10%
15%
3%
8%
13%
$13,830
$13,572
$5,065
$4,974
$4,883
$5,829
$5,705
$5,581
$7,056
$6,861
$6,666
$6,061
$5,950
$5,838
$16,840
$16,537
$16,234
$19,380
$18,966
$18,553
$23,459
$22,810
$22,161
$20,150
$19,780
$19,410
$11,064
$10,857
$4,892
$4,804
$4,716
$5,630
$5,510
$5,390
$6,815
$6,626
$6,438
$5,853
$5,746
$5,638
$14,311
$14,054
$13,797
$16,470
$16,119
$15,768
$19,937
$19,385
$18,833
$17,125
$16,810
$16,496
$11,064
$10,857
$4,892
$4,804
$4,716
$5,630
$5,510
$5,390
$6,815
$6,626
$6,438
$5,853
$5,746
$5,638
$14,311
$14,054
$13,797
$16,470
$16,119
$15,768
$19,937
$19,385
$18,833
$17,125
$16,810
$16,496
$8,851
$8,686
$4,753
$4,668
$4,582
$5,470
$5,353
$5,237
$6,621
$6,438
$6,255
$5,687
$5,583
$5,479
$12,289
$12,068
$11,847
$14,142
$13,841
$13,539
$17,119
$16,645
$16,172
$14,705
$14,434
$14,164
$8,851
$8,686
$4,753
$4,668
$4,582
$5,470
$5,353
$5,237
$6,621
$6,438
$6,255
$5,687
$5,583
$5,479
$12,289
$12,068
$11,847
$14,142
$13,841
$13,539
$17,119
$16,645
$16,172
$14,705
$14,434
$14,164
$8,851
$8,686
$4,753
$4,668
$4,582
$5,470
$5,353
$5,237
$6,621
$6,438
$6,255
$5,687
$5,583
$5,479
$12,289
$12,068
$11,847
$14,142
$13,841
$13,539
$17,119
$16,645
$16,172
$14,705
$14,434
$14,164
$8,851
$8,686
$4,753
$4,668
$4,582
$5,470
$5,353
$5,237
$6,621
$6,438
$6,255
$5,687
$5,583
$5,479
$12,289
$12,068
$11,847
$14,142
$13,841
$13,539
$17,119
$16,645
$16,172
$14,705
$14,434
$14,164
$8,851
$8,686
$4,753
$4,668
$4,582
$5,470
$5,353
$5,237
$6,621
$6,438
$6,255
$5,687
$5,583
$5,479
$12,289
$12,068
$11,847
$14,142
$13,841
$13,539
$17,119
$16,645
$16,172
$14,705
$14,434
$14,164
$7,081
$6,949
$2,994
$2,940
$2,886
$3,445
$3,372
$3,298
$4,171
$4,055
$3,940
$3,582
$3,517
$3,451
$9,022
$8,860
$8,698
$10,383
$10,162
$9,940
$12,569
$12,221
$11,873
$10,796
$10,598
$10,399
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-86 Costs for EV100 Battery Packs for the 2010 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
Vehicle
class
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Applied
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
Net
WR
4%
9%
14%
4%
9%
14%
5%
10%
15%
3%
8%
13%
4%
9%
2017
$12,046
$11,831
$11,615
$13,794
$13,512
$13,231
$16,845
$16,382
$15,919
$14,246
$13,982
$13,718
$5,182
$5,089
2018
$9,637
$9,465
$9,292
$11,035
$10,810
$10,585
$13,476
$13,106
$12,735
$11,397
$11,185
$10,974
$5,005
$4,915
2019
$9,637
$9,465
$9,292
$11,035
$10,810
$10,585
$13,476
$13,106
$12,735
$11,397
$11,185
$10,974
$5,005
$4,915
2020
$7,710
$7,572
$7,434
$8,828
$8,648
$8,468
$10,781
$10,484
$10,188
$9,117
$8,948
$8,779
$4,863
$4,776
2021
$7,710
$7,572
$7,434
$8,828
$8,648
$8,468
$10,781
$10,484
$10,188
$9,117
$8,948
$8,779
$4,863
$4,776
2022
$7,710
$7,572
$7,434
$8,828
$8,648
$8,468
$10,781
$10,484
$10,188
$9,117
$8,948
$8,779
$4,863
$4,776
2023
$7,710
$7,572
$7,434
$8,828
$8,648
$8,468
$10,781
$10,484
$10,188
$9,117
$8,948
$8,779
$4,863
$4,776
2024
$7,710
$7,572
$7,434
$8,828
$8,648
$8,468
$10,781
$10,484
$10,188
$9,117
$8,948
$8,779
$4,863
$4,776
2025
$6,168
$6,057
$5,947
$7,062
$6,918
$6,774
$8,625
$8,388
$8,150
$7,294
$7,159
$7,023
$3,063
$3,008
3-170
-------
Technologies Considered in the Agencies' Analysis
1C
1C
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Standard
car
Standard
car
Standard
car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
14%
4%
9%
14%
5%
10%
15%
3%
8%
13%
4%
9%
14%
4%
9%
14%
5%
10%
15%
3%
8%
13%
$4,997
$5,934
$5,813
$5,692
$7,247
$7,047
$6,848
$6,128
$6,015
$5,901
$17,229
$16,920
$16,612
$19,728
$19,325
$18,923
$24,092
$23,429
$22,767
$20,375
$19,997
$19,619
$4,826
$5,731
$5,614
$5,497
$6,999
$6,806
$6,614
$5,919
$5,809
$5,699
$14,642
$14,380
$14,118
$16,766
$16,424
$16,082
$20,475
$19,912
$19,348
$17,316
$16,994
$16,673
$4,826
$5,731
$5,614
$5,497
$6,999
$6,806
$6,614
$5,919
$5,809
$5,699
$14,642
$14,380
$14,118
$16,766
$16,424
$16,082
$20,475
$19,912
$19,348
$17,316
$16,994
$16,673
$4,689
$5,568
$5,455
$5,341
$6,800
$6,613
$6,426
$5,751
$5,644
$5,538
$12,573
$12,348
$12,122
$14,396
$14,103
$13,809
$17,581
$17,097
$16,614
$14,868
$14,593
$14,317
$4,689
$5,568
$5,455
$5,341
$6,800
$6,613
$6,426
$5,751
$5,644
$5,538
$12,573
$12,348
$12,122
$14,396
$14,103
$13,809
$17,581
$17,097
$16,614
$14,868
$14,593
$14,317
$4,689
$5,568
$5,455
$5,341
$6,800
$6,613
$6,426
$5,751
$5,644
$5,538
$12,573
$12,348
$12,122
$14,396
$14,103
$13,809
$17,581
$17,097
$16,614
$14,868
$14,593
$14,317
$4,689
$5,568
$5,455
$5,341
$6,800
$6,613
$6,426
$5,751
$5,644
$5,538
$12,573
$12,348
$12,122
$14,396
$14,103
$13,809
$17,581
$17,097
$16,614
$14,868
$14,593
$14,317
$4,689
$5,568
$5,455
$5,341
$6,800
$6,613
$6,426
$5,751
$5,644
$5,538
$12,573
$12,348
$12,122
$14,396
$14,103
$13,809
$17,581
$17,097
$16,614
$14,868
$14,593
$14,317
$2,953
$3,507
$3,436
$3,364
$4,283
$4,165
$4,048
$3,622
$3,555
$3,488
$9,231
$9,066
$8,900
$10,570
$10,354
$10,138
$12,908
$12,553
$12,198
$10,916
$10,714
$10,511
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-87 Costs for EV150 Battery Packs for the 2008 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
Vehicle
class
Small car
Standard
car
Large car
Small MPV
Small car
Standard
car
Large car
Small MPV
Small car
Standard
car
Large car
Small MPV
Applied
WR
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
Net
WR
2%
2%
3%
1%
2%
2%
3%
1%
2%
2%
3%
1%
2017
$15,701
$18,950
$21,552
$19,744
$6,755
$8,152
$9,272
$8,493
$22,456
$27,102
$30,824
$28,237
2018
$12,561
$15,160
$17,242
$15,795
$6,523
$7,873
$8,954
$8,203
$19,084
$23,033
$26,196
$23,998
2019
$12,561
$15,160
$17,242
$15,795
$6,523
$7,873
$8,954
$8,203
$19,084
$23,033
$26,196
$23,998
2020
$10,049
$12,128
$13,793
$12,636
$6,338
$7,650
$8,700
$7,970
$16,387
$19,777
$22,494
$20,606
2021
$10,049
$12,128
$13,793
$12,636
$6,338
$7,650
$8,700
$7,970
$16,387
$19,777
$22,494
$20,606
2022
$10,049
$12,128
$13,793
$12,636
$6,338
$7,650
$8,700
$7,970
$16,387
$19,777
$22,494
$20,606
2023
$10,049
$12,128
$13,793
$12,636
$6,338
$7,650
$8,700
$7,970
$16,387
$19,777
$22,494
$20,606
2024
$10,049
$12,128
$13,793
$12,636
$6,338
$7,650
$8,700
$7,970
$16,387
$19,777
$22,494
$20,606
2025
$8,039
$9,702
$11,035
$10,109
$3,992
$4,818
$5,480
$5,020
$12,031
$14,520
$16,515
$15,129
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3-171
-------
Technologies Considered in the Agencies' Analysis
Table 3-88 Costs for EV150 Battery Packs for the 2010 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
Vehicle
class
Small car
Standard
car
Large car
Small MPV
Small car
Standard
car
Large car
Small MPV
Small car
Standard
car
Large car
Small MPV
Applied
WR
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
Net
WR
2%
2%
3%
1%
2%
2%
3%
1%
2%
2%
3%
1%
2017
$16,102
$19,265
$22,080
$19,976
$6,927
$8,287
$9,498
$8,593
$23,028
$27,552
$31,578
$28,569
2018
$12,881
$15,412
$17,664
$15,981
$6,690
$8,004
$9,173
$8,299
$19,571
$23,415
$26,837
$24,280
2019
$12,881
$15,412
$17,664
$15,981
$6,690
$8,004
$9,173
$8,299
$19,571
$23,415
$26,837
$24,280
2020
$10,305
$12,329
$14,131
$12,784
$6,500
$7,777
$8,913
$8,064
$16,805
$20,106
$23,044
$20,848
2021
$10,305
$12,329
$14,131
$12,784
$6,500
$7,777
$8,913
$8,064
$16,805
$20,106
$23,044
$20,848
2022
$10,305
$12,329
$14,131
$12,784
$6,500
$7,777
$8,913
$8,064
$16,805
$20,106
$23,044
$20,848
2023
$10,305
$12,329
$14,131
$12,784
$6,500
$7,777
$8,913
$8,064
$16,805
$20,106
$23,044
$20,848
2024
$10,305
$12,329
$14,131
$12,784
$6,500
$7,777
$8,913
$8,064
$16,805
$20,106
$23,044
$20,848
2025
$8,244
$9,863
$11,305
$10,228
$4,094
$4,898
$5,614
$5,079
$12,338
$14,762
$16,919
$15,307
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
For Mild HEV batteries, the agencies used a similar approach to estimating the cost of
the battery pack but used a different approach to determining its size. Our Mild HEV system
used in the analyses is based, largely, on the Buick eAssist system.37 According to the press
releases, it includes a 15 kW motor and a 15 kW/0.5kWh/l 15 Volt two-module battery. For
the agencies' analyses, a 15kW/0.25kWh/l 10 Volt single-module battery was selected for
several reasons. First, the Buick system uses a 20% state-of-charge (SOC) swing for the
battery. We believe that, in the 2017-2025 timeframe, a 40% SOC swing is reasonable. As
such, the energy capacity of the battery can be halved (from 0.5 to 0.25 kWh).22 The 110V
system used in the analysis is essentially the same as Buick's 115V system. The voltage
change is due to our use of a 28 cell single-module battery pack rather than the 32 cell double-
module battery pack which is used in the eAssist system. Such changes are consistent with
our expectation that cells will increase in size allowing for fewer cells and fewer modules.
Further, for the Mild HEV technology, the agencies are using the same system regardless of
vehicle class or subclass. In other words, the Mild HEV system is a stand-alone technology
that can be applied to any subclass without unique modifications for each class or subclass.
As such, it adds more weight as a percentage to a smaller vehicle than to a larger vehicle but it
provides more effectiveness to a smaller vehicle than to a larger vehicle. Since the same
system is used regardless of vehicle class or subclass, the costs are identical regardless of
vehicle class or subclass. Using the ANL BatPaC model, the Mild HEV battery DMC was
calculated as $553 and is considered applicable to the MY 2017. The agencies derived the
Mild HEV battery pack cost using the same methodology that was used for the P2 HEV
yy "eAssist" is a Buick (or General Motors) term and is not a generic term for this technology, hence our use of
the term mild hybrid.
zz Note that projected battery cost is relatively insensitive to kWh capacity at the high power-to-energy ratio of
these batteries. A 0.5 kWh battery could alternatively be specified at a similar cost.
3-172
-------
Technologies Considered in the Agencies' Analysis
battery pack, and consider cost to be on the flat portion of the learning curve during the 2017-
2025 timeframe. The agencies have applied a highl complexity ICM of 1.56 through 2024
then 1.35 thereafter. The resultant Mild HEV battery pack costs are as shown in Table 3-89.
The associated weight penalties are as shown in Table 3-90.
Table 3-89 Costs for Mild Hybrid (MHEV) Battery Packs for both the 2008 and 2010 Baselines (2010$)
Cost type
DMC
1C
TC
Vehicle class
All
All
All
2017
$553
$312
$865
2018
$536
$311
$847
2019
$520
$310
$830
2020
$505
$309
$813
2021
$490
$308
$797
2022
$475
$307
$782
2023
$461
$306
$766
2024
$447
$305
$752
2025
$433
$187
$621
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-90 EPA and NHTSA Weight Reduction Offset Associated with MHEV for both the 2008 and 2010
Baselines
Vehicle class
Small car
Standard car
Large car
Small MPV
Large MPV
Truck
Weight penalty
3.5%
3.0%
2.5%
2.5%
2.5%
2.0%
The CAFE model does not use pre-built packages and it applies technologies
incrementally as necessary to meet the fuel consumption reduction requirement, so the cost
interaction between any particular technology and other technologies (cost synergies) must be
defined. This allows flexibility so that when a technology is picked, the model will
automatically look through the cost synergy defined in a table and apply cost adjustments
accordingly. The total cost for mass reduction and electrification is composed of the following
four parts:
(1) Cost of net mass reduction;
(2) Cost of electrification with zero mass reduction;
(3) Mass reduction cost synergy for increased or decreased amount of mass
reduction due to switching from conventional powertrain to electrification
systems as defined in Figure 3-25. For an example, if a midsize passenger car
needs both 10 percent net mass reduction and P2 hybrid to meet the CAFE
target, the model will need to find the cost of additional 5 percent of mass
reduction to consider the vehicle weight increase due to switching from
conventional powertrain system to P2 electrification packages. This additional
5 percent of mass reduction is calculated starting from 10 percent mass
reduction, not zero as shown in Figure 3-25 because mass reduction cost versus
mass reduction percent is not a linear function. The cost increases faster as the
3-173
-------
Technologies Considered in the Agencies' Analysis
amount of mass reduction becomes higher.
(4) Electrification system cost synergies (battery and non-battery components) due
to mass reduction as defined in Table 3-76 and Table 3-103: Continuing the
example in the steps above, if a midsize passenger car needs both 10 percent
net mass reduction and P2 hybrid to meet the CAFE target, after calculating
the costs above, the model will need to find the cost of electrification systems,
including battery system and non-battery system, with the required net amount
of mass reduction using the equations in Table 3-76 and Table 3-103. Then the
delta cost between this cost and the cost calculated in step 2, i.e. electrification
system cost with zero applied mass reduction is calculated and treated as a cost
synergy. These cost deltas are normally negative, i.e., a cost reduction, due to
the downsizing of the electrification system resulting from mass reduction
The sum of item (3) and (4) in the above list are calculated as cost synergies and
stored in the cost synergy table as defined in NHTSA's RIA.
Figure 3-25 Mass Reduction Cost Example for Applied and Net Mass Reduction
$600
5 $500
+-*
m
2 $400
o
'•g $300
3
1 $200
| $100
$-
0
Example of Applied and Net Mass Reduction
Costs
j?
. ro* ,
^>; o^' ^\
"f // ^
^ ^ff ^
C0^%H
^^
K^^^
^^
X
fc 5% 10% 15'\. 20%
Amount of Mass Reduction [%]
The agencies have also carefully reconsidered the power and energy requirements for
each electrified vehicle type, which has a significant impact on the cost estimates for HEVs,
PHEVs, and EVs as compared to the estimates used in the 2012-2016 rulemaking.
The agencies note that, for this analysis, the agencies have assumed batteries will be
capable of lasting the lifetime of the vehicle, which is consistent with the expected customer
demands from this technology (as manufacturers have confirmed). Lastly, the agencies have
focused attention on an emerging HEV technology known as a P2-hybrid, a technology not
considered in the 2012-2016 light-duty rule.
3-174
-------
Technologies Considered in the Agencies' Analysis
The agencies have also considered, for this analysis, the costs associated with in-home
chargers expected to be necessary for PHEVs and EVs. Further details on in-home chargers
and their estimated costs are presented in Section 3.4.4.
3.4.3.10 Non-battery costs for MHEVs, HEVs, PHEVs, EVs and FCEVs
This section addresses the costs of non-battery components which are required for
electric drive vehicles. Some of these components are not found in every electric-drive
vehicle (e.g. an HEV does not have an on-board battery charger as found in a PHEV or EV).
Others are found in all electric drive vehicles and/or must be scaled to the vehicle type or
class to properly represent the cost. The agencies derived the costs of these components from
the FEV teardown study and the 2010 TAR. Where appropriate, costs were scaled to vehicle
class and in the case of the motor and inverter, the sizing methodology used for battery sizing
was applied.
The electric drive motor and inverter provide the motive power for any electric-drive
vehicle converting electrical energy from the battery into kinetic energy for propulsion. In an
electric-drive vehicle, energy stored in the battery is routed to the inverter which converts it to
a voltage and wave form that can be used by the motor.
In many cases, such as HEVs, the combined cost of the motor and inverter exceed the
battery cost. As batteries become larger in PHEVs and EVs, the battery cost grows faster than
motor and inverter cost. For this analysis, the agencies used the vehicle power requirement
calculation discussed in 3.4.3.8 to calculate the required motor and inverter size for each
vehicle class at each weight reduction point. Then, for the HEVs and PHEVs, a regression
was created from the FEV teardown data for motors and inverters and this regression was
used to calculate the motor and inverter cost for each combination of vehicle class and weight
reduction. This regression for use with the 2008 baseline was $13.78x(motor size in
kW)+$781.50 (values in 2010$), and for use with the 2010 baseline was $14.13x(motor size
in kW)+$771.21 (values in 2010$). The results are shown as the "Motor assembly" line item
in Table 3-91 through Table 3-96, which show our scaled DMC for P2 HEV, PHEV20 and
PHEV40, respectively, for both the 2008 and 2010 baselines.
For EVs, the agencies used the motor and inverter cost regression from the 2010 TAR
(see 2010 TAR at page B-21) and we used that regression for both the 2008 and 2010
baselines. Since the FEV teardown was conducted on an HEV Ford Fusion, the agencies
believe the technology for an EV is different enough to warrant using the TAR regression.
The regression presented in the TAR showed the DMC being equal to $8.45x(motor size in
kW)+$185.05 (values in 2010$). The results are presented as separate line items for "Motor
inverter" and "Motor assembly" in Table 3-97 through Table 3-102, which show our scaled
DMC for EV75, EV100 and EV150, respectively, for both the 2008 and 2010 baselines.
In addition to electric drive motors and inverters, there are several other components
in electric drive vehicles that are required. These components include the following:
• Body Modifications which are required on HEVs and PHEVs include changes
to sheet metal to accommodate electric drive components and the addition of fasteners to
secure components such as electric cables. These costs come from the FEV teardown and are
scaled by vehicle class. For EVs, these costs are assumed to be included in the base vehicle
because they are less likely to be adapted from conventional vehicles.
3-175
-------
Technologies Considered in the Agencies' Analysis
• Brake System changes include the addition of a braking system that can control
the vehicle's regenerative braking system—a key enabler of electric drive vehicle efficiency.
The brake system costs are from the FEV teardown and are scaled to vehicle class.
• Climate Control System includes components such as an electric air
conditioning compressor that enables operation while the engine is off for HEVs and PHEVs
as well as for an EV which has no engine. Climate control system costs come from the FEV
teardown and are scaled to vehicle class.
• Conventional vehicle battery and alternator are deleted in these vehicles, for a
cost savings, replaced by the DC-DC converter which converts the high-voltage traction
battery to a nominal 12V DC to operate the vehicle's accessories. This credit comes from the
FEV teardown study and is scaled to vehicle class.
• DC-DC converter converts the high-voltage battery voltage to a nominal 12V
battery voltage to run vehicle accessories such as the radio, lights and wipers. This cost
comes from the FEV teardown study and is scaled to vehicle class.
• Power distribution and Control consists of those components which route
electricity to the motor, inverter and contains the controllers to operate and monitor the
electric drive system. This cost applies to HEVs and PHEVs and comes from the FEV
teardown study. It is scaled to vehicle class.
• On-Vehicle Charger consists of the components necessary to charge a PHEV
or EV from an outlet. It includes the charging port, wiring and electronics necessary to
convert a 120V or 240V AC input to the high-voltage DC power necessary to charge the
battery. Because the FEV teardown study subject vehicle did not have an on-vehicle charger,
the costs from the TAR were used for this item. It is not scaled to vehicle class, however the
EV charger is assumed to cost twice the amount of the PHEV charger to account for a higher
current capacity. This cost does not include off-vehicle charger components which are
discussed in Section 3.4.4, below.
• Supplemental heating is required for passenger comfort on PHEVs and EVs
which may operate for long periods with no engine heat available. This cost comes from the
FEV teardown study and is scaled to vehicle class. The supplemental heater on the EV is
assumed to be three times more costly than the PHEV because the entire cabin comfort is
dependent on the supplemental heater. In a PHEV, it is assumed that in extreme conditions,
the internal combustion engine will start to provide additional cabin heat and defrost
functions.
• High Voltage Wiring is an item used on EVs only. It includes the high voltage
cabling from the battery to the inverter and motor as well as control components. It is
equivalent to the power distribution and control used on HEVs and PHEVs and comes from
the FEV teardown study. It is scaled to vehicle class.
• Delete Internal Combustion Engine and Transmission For EVs, the engine and
transmission are deleted and a credit is applied. These credits come from work done in
support of the 2010 TAR and are scaled to vehicle class.
• Battery Discharge System For HEVs, PHEVs and EVs, it is expected that
manufacturers will provide the means to safely discharge battery packs following a vehicle
3-176
-------
Technologies Considered in the Agencies' Analysis
crash. The agencies have assumed that this would include dedicated DC terminals, an access
panel for the terminals, and a diagnostics port. The estimated cost of this capability is the
same for all vehicle classes, but is different for HEVs than for PHEVs and EVs.
The results of the scaling exercise applied to non-battery components are presented in
Table 3-91 through Table 3-102 for P2 HEVs, PHEV20, PHEV40, EV75, EV100 and EV150,
for the 2008 and 2010 baselines, respectively.
Table 3-91 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for P2 HEV for the 2008
Baseline (2010$)
System
0%WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Battery discharge system
Motor assembly
Total
2%WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Battery discharge system
Motor assembly
Total
7.5% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Battery discharge system
Motor assembly
Total
10% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Battery discharge system
Motor assembly
Small
car
$6
$221
$140
-$60
$121
$196
$6
$1,045
$1,675
$6
$221
$140
-$60
$121
$196
$6
$1,039
$1,670
$6
$221
$140
-$60
$121
$196
$6
$1,025
$1,655
$6
$221
$140
-$60
$121
$196
$6
$1,018
Standard
car
$6
$228
$157
-$65
$152
$201
$6
$1,172
$1,857
$6
$228
$157
-$65
$152
$201
$6
$1,164
$1,849
$6
$228
$157
-$65
$152
$201
$6
$1,143
$1,828
$6
$228
$157
-$65
$152
$201
$6
$1,133
Large
car
$6
$231
$168
-$82
$162
$204
$6
$1,480
$2,175
$6
$231
$168
-$82
$162
$204
$6
$1,467
$2,161
$6
$231
$168
-$82
$162
$204
$6
$1,428
$2,123
$6
$231
$168
-$82
$162
$204
$6
$1,411
Small
MPV
$6
$225
$164
-$86
$152
$200
$6
$1,112
$1,777
$6
$225
$164
-$86
$152
$200
$6
$1,106
$1,771
$6
$225
$164
-$86
$152
$200
$6
$1,088
$1,752
$6
$225
$164
-$86
$152
$200
$6
$1,079
Large
MPV
$6
$233
$250
-$86
$152
$206
$6
$1,287
$2,052
$6
$233
$250
-$86
$152
$206
$6
$1,277
$2,042
$6
$233
$250
-$86
$152
$206
$6
$1,249
$2,014
$6
$233
$250
-$86
$152
$206
$6
$1,237
Truck
$6
$240
$186
-$94
$177
$220
$6
$1,429
$2,169
$6
$240
$186
-$94
$177
$220
$6
$1,416
$2,156
$6
$240
$186
-$94
$177
$220
$6
$1,381
$2,121
$6
$240
$186
-$94
$177
$220
$6
$1,364
3-177
-------
Technologies Considered in the Agencies' Analysis
Total
20% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Battery discharge system
Motor assembly
Total
$1,649
$6
$221
$140
-$60
$121
$196
$6
$1,007
$1,637
$1,818
$6
$228
$157
-$65
$152
$201
$6
$1,115
$1,800
$2,105
$6
$231
$168
-$82
$162
$204
$6
$1,377
$2,071
$1,744
$6
$225
$164
-$86
$152
$200
$6
$1,064
$1,729
$2,002
$6
$233
$250
-$86
$152
$206
$6
$1,212
$1,977
$2,104
$6
$240
$186
-$94
$177
$220
$6
$1,337
$2,077
Table 3-92 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for P2 HEV for the 2010
Baseline (2010$)
System
0%WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Battery discharge system
Motor assembly
Total
2%WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Battery discharge system
Motor assembly
Total
7.5% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Battery discharge system
Motor assembly
Total
10% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Small car
$6
$223
$140
-$60
$121
$197
$6
$1,051
$1,683
$6
$223
$140
-$60
$121
$197
$6
$1,045
$1,677
$6
$223
$140
-$60
$121
$197
$6
$1,030
$1,662
$6
$223
$140
-$60
$121
$197
Standard car
$6
$229
$157
-$65
$152
$202
$6
$1,191
$1,878
$6
$229
$157
-$65
$152
$202
$6
$1,183
$1,869
$6
$229
$157
-$65
$152
$202
$6
$1,159
$1,846
$6
$229
$157
-$65
$152
$202
Large car
$6
$232
$168
-$82
$177
$205
$6
$1,512
$2,224
$6
$232
$168
-$82
$177
$205
$6
$1,497
$2,210
$6
$232
$168
-$82
$177
$205
$6
$1,457
$2,169
$6
$232
$168
-$82
$177
$205
Small MPV
$6
$225
$164
-$86
$162
$201
$6
$1,134
$1,811
$6
$225
$164
-$86
$162
$201
$6
$1,127
$1,804
$6
$225
$164
-$86
$162
$201
$6
$1,107
$1,784
$6
$225
$164
-$86
$162
$201
Large MPV
$6
$232
$250
-$86
$162
$206
$6
$1,299
$2,073
$6
$232
$250
-$86
$162
$206
$6
$1,288
$2,063
$6
$232
$250
-$86
$162
$206
$6
$1,259
$2,034
$6
$232
$250
-$86
$162
$206
Truck
$6
$242
$186
-$94
$177
$221
$6
$1,445
$2,188
$6
$242
$186
-$94
$177
$221
$6
$1,432
$2,175
$6
$242
$186
-$94
$177
$221
$6
$1,395
$2,138
$6
$242
$186
-$94
$177
$221
3-178
-------
Technologies Considered in the Agencies' Analysis
Battery discharge system
Motor assembly
Total
20% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
Battery discharge system
Motor assembly
Total
$6
$1,023
$1,655
$6
$223
$140
-$60
$121
$197
$6
$1,010
$1,642
$6
$1,149
$1,836
$6
$229
$157
-$65
$152
$202
$6
$1,129
$1,816
$6
$1,438
$2,150
$6
$232
$168
-$82
$177
$205
$6
$1,402
$2,114
$6
$1,098
$1,775
$6
$225
$164
-$86
$162
$201
$6
$1,081
$1,757
$6
$1,246
$2,021
$6
$232
$250
-$86
$162
$206
$6
$1,220
$1,994
$6
$1,378
$2,121
$6
$242
$186
-$94
$177
$221
$6
$1,350
$2,093
Table 3-93 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for PHEV20 for the 2008
Baseline (2010$)
System
0% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
2% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
7.5% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Small car
$6
$221
$140
-$60
$121
$196
$105
$38
$2,097
$13
$2,878
$6
$221
$140
-$60
$121
$196
$105
$38
$2,071
$13
$2,852
$6
$221
$140
-$60
$121
$196
$105
$38
$1,999
$13
Standard car
$6
$228
$157
-$65
$152
$201
$105
$43
$2,735
$13
$3,575
$6
$228
$157
-$65
$152
$201
$105
$43
$2,695
$13
$3,536
$6
$228
$157
-$65
$152
$201
$105
$43
$2,588
$13
Large car
$6
$231
$168
-$82
$162
$204
$105
$45
$4,276
$13
$5,129
$6
$231
$168
-$82
$162
$204
$105
$45
$4,207
$13
$5,059
$6
$231
$168
-$82
$162
$204
$105
$45
$4,014
$13
Small MPV
$6
$225
$164
-$86
$152
$200
$105
$44
$2,436
$13
$3,258
$6
$225
$164
-$86
$152
$200
$105
$44
$2,403
$13
$3,225
$6
$225
$164
-$86
$152
$200
$105
$44
$2,312
$13
3-179
-------
Technologies Considered in the Agencies' Analysis
Total
10% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
20% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
$2,780
$6
$221
$140
-$60
$121
$196
$105
$38
$1,966
$13
$2,747
$6
$221
$140
-$60
$121
$196
$105
$38
$1,943
$13
$2,724
$3,428
$6
$228
$157
-$65
$152
$201
$105
$43
$2,539
$13
$3,379
$6
$228
$157
-$65
$152
$201
$105
$43
$2,500
$13
$3,341
$4,867
$6
$231
$168
-$82
$162
$204
$105
$45
$3,927
$13
$4,780
$6
$231
$168
-$82
$162
$204
$105
$45
$3,861
$13
$4,714
$3,134
$6
$225
$164
-$86
$152
$200
$105
$44
$2,271
$13
$3,093
$6
$225
$164
-$86
$152
$200
$105
$44
$2,235
$13
$3,057
a The agencies have not estimated PHEV or EV costs for the large MPV and track vehicle classes since we do not believe these vehicle
classes would use the technologies.
Table 3-94 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for PHEV20 for the 2010
Baseline (2010$)
System
0% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
2% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Small car
$6
$223
$140
-$60
$121
$197
$105
$38
$2,169
$13
$2,951
$6
$223
$140
-$60
$121
$197
$105
$38
$2,141
Standard car
$6
$229
$157
-$65
$152
$202
$105
$43
$2,870
$13
$3,712
$6
$229
$157
-$65
$152
$202
$105
$43
$2,828
Large car
$6
$232
$168
-$82
$177
$205
$105
$45
$4,476
$13
$5,347
$6
$232
$168
-$82
$177
$205
$105
$45
$4,402
Small MPV
$6
$225
$164
-$86
$162
$201
$105
$44
$2,586
$13
$3,419
$6
$225
$164
-$86
$162
$201
$105
$44
$2,549
5-180
-------
Technologies Considered in the Agencies' Analysis
Battery discharge system
Total
7.5% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
10% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
20% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
$13
$2,924
$6
$223
$140
-$60
$121
$197
$105
$38
$2,064
$13
$2,847
$6
$223
$140
-$60
$121
$197
$105
$38
$2,029
$13
$2,812
$6
$223
$140
-$60
$121
$197
$105
$38
$2,002
$13
$2,785
$13
$3,670
$6
$229
$157
-$65
$152
$202
$105
$43
$2,712
$13
$3,554
$6
$229
$157
-$65
$152
$202
$105
$43
$2,660
$13
$3,502
$6
$229
$157
-$65
$152
$202
$105
$43
$2,616
$13
$3,458
$13
$5,272
$6
$232
$168
-$82
$177
$205
$105
$45
$4,198
$13
$5,069
$6
$232
$168
-$82
$177
$205
$105
$45
$4,106
$13
$4,976
$6
$232
$168
-$82
$177
$205
$105
$45
$4,031
$13
$4,901
$13
$3,383
$6
$225
$164
-$86
$162
$201
$105
$44
$2,450
$13
$3,283
$6
$225
$164
-$86
$162
$201
$105
$44
$2,404
$13
$o T5 o
J,ZJO
$6
$225
$164
-$86
$162
$201
$105
$44
$2,364
$13
$3,197
a The agencies have not estimated PHEV or EV costs for the large MPV and track vehicle classes since we do not believe these vehicle
classes would use the technologies.
Table 3-95 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for PHEV40 for the 2008
Baseline (2010$)
System
0% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Small car
$6
$221
$140
-$60
$121
$196
$105
Standard car
$6
$228
$157
-$65
$152
$201
$105
Large car
$6
$231
$168
-$82
$162
$204
$105
Small MPV
$6
$225
$164
-$86
$152
$200
$105
5-181
-------
Technologies Considered in the Agencies' Analysis
Supplemental heater
Motor assembly
Battery discharge system
Total
2% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
7.5% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
10% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
20% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
$38
$2,097
$13
$2,878
$6
$221
$140
-$60
$121
$196
$105
$38
$2,071
$13
$2,852
$6
$221
$140
-$60
$121
$196
$105
$38
$2,007
$13
$2,788
$6
$221
$140
-$60
$121
$196
$105
$38
$2,007
$13
$2,788
$6
$221
$140
-$60
$121
$196
$105
$38
$2,007
$13
$2,788
$43
$2,735
$13
$3,575
$6
$228
$157
-$65
$152
$201
$105
$43
$2,695
$13
$3,536
$6
$228
$157
-$65
$152
$201
$105
$43
$2,591
$13
$3,432
$6
$228
$157
-$65
$152
$201
$105
$43
$2,591
$13
$3,432
$6
$228
$157
-$65
$152
$201
$105
$43
$2,591
$13
$3,432
$45
$4,276
$13
$5,129
$6
$231
$168
-$82
$162
$204
$105
$45
$4,207
$13
$5,059
$6
$231
$168
-$82
$162
$204
$105
$45
$4,025
$13
$4,878
$6
$231
$168
-$82
$162
$204
$105
$45
$4,025
$13
$4,878
$6
$231
$168
-$82
$162
$204
$105
$45
$4,025
$13
$4,878
$44
$2,436
$13
$3,258
$6
$225
$164
-$86
$152
$200
$105
$44
$2,403
$13
$3,225
$6
$225
$164
-$86
$152
$200
$105
$44
$2,313
$13
$3,135
$6
$225
$164
-$86
$152
$200
$105
$44
$2,312
$13
$3,134
$6
$225
$164
-$86
$152
$200
$105
$44
$2,312
$13
$3,134
a The agencies have not estimated PHEV or EV costs for the large MPV and track vehicle classes since we do not believe these vehicle
classes would use the technologies.
5-182
-------
Technologies Considered in the Agencies' Analysis
Table 3-96 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for PHEV40 for the 2010
Baseline (2010$)a
System
0% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
2% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
7.5% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
10% WR
Body system
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
20% WR
Body system
Small car
$6
$223
$140
-$60
$121
$197
$105
$38
$2,169
$13
$2,951
$6
$223
$140
-$60
$121
$197
$105
$38
$2,141
$13
$2,924
$6
$223
$140
-$60
$121
$197
$105
$38
$2,068
$13
$2,851
$6
$223
$140
-$60
$121
$197
$105
$38
$2,068
$13
$2,851
$6
Standard car
$6
$229
$157
-$65
$152
$202
$105
$43
$2,870
$13
$3,712
$6
$229
$157
-$65
$152
$202
$105
$43
$2,828
$13
$3,670
$6
$229
$157
-$65
$152
$202
$105
$43
$2,714
$13
$3,556
$6
$229
$157
-$65
$152
$202
$105
$43
$2,714
$13
$3,556
$6
Large car
$6
$232
$168
-$82
$177
$205
$105
$45
$4,476
$13
$5,347
$6
$232
$168
-$82
$177
$205
$105
$45
$4,402
$13
$5,272
$6
$232
$168
-$82
$177
$205
$105
$45
$4,206
$13
$5,076
$6
$232
$168
-$82
$177
$205
$105
$45
$4,206
$13
$5,076
$6
Small MPV
$6
$225
$164
-$86
$162
$201
$105
$44
$2,586
$13
$3,419
$6
$225
$164
-$86
$162
$201
$105
$44
$2,549
$13
$3,383
$6
$225
$164
-$86
$162
$201
$105
$44
$2,450
$13
$3,283
$6
$225
$164
-$86
$162
$201
$105
$44
$2,449
$13
$3,283
$6
5-183
-------
Technologies Considered in the Agencies' Analysis
Brake system
Climate controls
Delete electrical
DC-DC converter
Power Distr & control
On vehicle charger
Supplemental heater
Motor assembly
Battery discharge system
Total
$223
$140
-$60
$121
$197
$105
$38
$2,068
$13
$2,851
$229
$157
-$65
$152
$202
$105
$43
$2,714
$13
$3,556
$232
$168
-$82
$177
$205
$105
$45
$4,206
$13
$5,076
$225
$164
-$86
$162
$201
$105
$44
$2,449
$13
$3,283
a The agencies have not estimated PHEV or EV costs for the large MPV and track vehicle classes since we do not believe these vehicle
classes would use the technologies.
Table 3-97 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for EV75 for the 2008 Baseline
(2010$) a
System
0% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
2% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
7.5% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
Small car
$221
$140
-$60
$121
$196
$76
$316
$703
$121
-$1,596
-$894
$992
$13
$350
$221
$140
-$60
$121
$196
$76
$316
$689
$121
-$1,596
-$894
$976
$13
$320
$221
$140
-$60
$121
$196
$76
Standard car
$228
$157
-$65
$152
$201
$85
$316
$1,044
$121
-$1,596
-$894
$1,383
$13
$1,145
$228
$157
-$65
$152
$201
$85
$316
$1,023
$121
-$1,596
-$894
$1,359
$13
$1,100
$228
$157
-$65
$152
$201
$85
Large car
$231
$168
-$82
$162
$204
$91
$316
$1,868
$121
-$2,466
-$894
$2,329
$13
$2,060
$231
$168
-$82
$162
$204
$91
$316
$1,831
$121
-$2,466
-$894
$2,286
$13
$1,979
$231
$168
-$82
$162
$204
$91
Small MPV
$225
$164
-$86
$152
$200
$89
$316
$885
$121
-$2,394
-$894
$1,200
$13
-$12
$225
$164
-$86
$152
$200
$89
$316
$867
$121
-$2,394
-$894
$1,180
$13
-$50
$225
$164
-$86
$152
$200
$89
5-184
-------
Technologies Considered in the Agencies' Analysis
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
10% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
20% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
$316
$650
$121
-$1,596
-$894
$932
$13
$237
$221
$140
-$60
$121
$196
$76
$316
$633
$121
-$1,596
-$894
$911
$13
$199
$221
$140
-$60
$121
$196
$76
$316
$571
$121
-$1,596
-$894
$840
$13
$65
$316
$966
$121
-$1,596
-$894
$1,293
$13
$977
$228
$157
-$65
$152
$201
$85
$316
$939
$121
-$1,596
-$894
$1,263
$13
$921
$228
$157
-$65
$152
$201
$85
$316
$851
$121
-$1,596
-$894
$1,162
$13
$731
$316
$1,728
$121
-$2,466
-$894
$2,168
$13
$1,759
$231
$168
-$82
$162
$204
$91
$316
$1,681
$121
-$2,466
-$894
$2,114
$13
$1,659
$231
$168
-$82
$162
$204
$91
$316
$1,519
$121
-$2,466
-$894
$1,928
$13
$1,309
$316
$818
$121
-$2,394
-$894
$1,124
$13
-$154
$225
$164
-$86
$152
$200
$89
$316
$796
$121
-$2,394
-$894
$1,099
$13
-$202
$225
$164
-$86
$152
$200
$89
$316
$727
$121
-$2,394
-$894
$1,020
$13
-$350
a The agencies have not estimated PHEV or EV costs for the large MPV and truck vehicle classes since we do not believe these vehicle
classes would use the technologies.
Table 3-98 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for EV75 for the 2010 Baseline
(2010$) a
System
0% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Small car
$223
$140
-$60
$121
$197
$76
$316
$729
Standard car
$229
$157
-$65
$152
$202
$85
$316
$1,094
Large car
$232
$168
-$82
$177
$205
$91
$316
$1,932
Small MPV
$225
$164
-$86
$162
$201
$89
$316
$946
3-185
-------
Technologies Considered in the Agencies' Analysis
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
2% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
7.5% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
10% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
20% WR
Brake system
$121
-$1,596
-$894
$1,021
$13
$406
$223
$140
-$60
$121
$197
$76
$316
$714
$121
-$1,596
-$894
$1,004
$13
$375
$223
$140
-$60
$121
$197
$76
$316
$674
$121
-$1,596
-$894
$958
$13
$289
$223
$140
-$60
$121
$197
$76
$316
$656
$121
-$1,596
-$894
$938
$13
$250
$223
$121
-$1,596
-$894
$1,441
$13
$1,255
$229
$157
-$65
$152
$202
$85
$316
$1,072
$121
-$1,596
-$894
$1,416
$13
$1,208
$229
$157
-$65
$152
$202
$85
$316
$1,012
$121
-$1,596
-$894
$1,347
$13
$1,079
$229
$157
-$65
$152
$202
$85
$316
$985
$121
-$1,596
-$894
$1,315
$13
$1,020
$229
$121
-$2,466
-$894
$2,402
$13
$2,214
$232
$168
-$82
$177
$205
$91
$316
$1,893
$121
-$2,466
-$894
$2,358
$13
$2,131
$232
$168
-$82
$177
$205
$91
$316
$1,787
$121
-$2,466
-$894
$2,236
$13
$1,903
$232
$168
-$82
$177
$205
$91
$316
$1,739
$121
-$2,466
-$894
$2,180
$13
$1,799
$232
$121
-$2,394
-$894
$1,271
$13
$132
$225
$164
-$86
$162
$201
$89
$316
$927
$121
-$2,394
-$894
$1,249
$13
$92
$225
$164
-$86
$162
$201
$89
$316
$875
$121
-$2,394
-$894
$1,189
$13
-$20
$225
$164
-$86
$162
$201
$89
$316
$851
$121
-$2,394
-$894
$1,162
$13
-$71
$225
3-186
-------
Technologies Considered in the Agencies' Analysis
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
$140
-$60
$121
$197
$76
$316
$595
$121
-$1,596
-$894
$867
$13
$118
$157
-$65
$152
$202
$85
$316
$895
$121
-$1,596
-$894
$1,212
$13
$828
$168
-$82
$177
$205
$91
$316
$1,580
$121
-$2,466
-$894
$1,998
$13
$1,458
$164
-$86
$162
$201
$89
$316
$780
$121
-$2,394
-$894
$1,080
$13
-$225
a The agencies have not estimated PHEV or EV costs for the large MPV and track vehicle classes since we do not believe these vehicle
classes would use the technologies.
Table 3-99 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for EV100 for the 2008
Baseline (2010$)a
System
0% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
2% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
7.5% WR
Brake system
Climate controls
Delete electrical
Small car
$221
$140
-$60
$121
$196
$76
$316
$703
$121
-$1,596
-$894
$992
$13
$350
$221
$140
-$60
$121
$196
$76
$316
$689
$121
-$1,596
-$894
$976
$13
$320
$221
$140
-$60
Standard car
$228
$157
-$65
$152
$201
$85
$316
$1,044
$121
-$1,596
-$894
$1,383
$13
$1,145
$228
$157
-$65
$152
$201
$85
$316
$1,023
$121
-$1,596
-$894
$1,359
$13
$1,100
$228
$157
-$65
Large car
$231
$168
-$82
$162
$204
$91
$316
$1,868
$121
-$2,466
-$894
$2,329
$13
$2,060
$231
$168
-$82
$162
$204
$91
$316
$1,831
$121
-$2,466
-$894
$2,286
$13
$1,979
$231
$168
-$82
Small MPV
$225
$164
-$86
$152
$200
$89
$316
$885
$121
-$2,394
-$894
$1,200
$13
-$12
$225
$164
-$86
$152
$200
$89
$316
$867
$121
-$2,394
-$894
$1,180
$13
-$50
$225
$164
-$86
5-187
-------
Technologies Considered in the Agencies' Analysis
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
10% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
20% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
$121
$196
$76
$316
$650
$121
-$1,596
-$894
$932
$13
$237
$221
$140
-$60
$121
$196
$76
$316
$633
$121
-$1,596
-$894
$911
$13
$199
$221
$140
-$60
$121
$196
$76
$316
$608
$121
-$1,596
-$894
$883
$13
$146
$152
$201
$85
$316
$966
$121
-$1,596
-$894
$1,293
$13
$977
$228
$157
-$65
$152
$201
$85
$316
$939
$121
-$1,596
-$894
$1,263
$13
$921
$228
$157
-$65
$152
$201
$85
$316
$906
$121
-$1,596
-$894
$1,224
$13
$848
$162
$204
$91
$316
$1,728
$121
-$2,466
-$894
$2,168
$13
$1,759
$231
$168
-$82
$162
$204
$91
$316
$1,681
$121
-$2,466
-$894
$2,114
$13
$1,659
$231
$168
-$82
$162
$204
$91
$316
$1,617
$121
-$2,466
-$894
$2,041
$13
$1,521
$152
$200
$89
$316
$818
$121
-$2,394
-$894
$1,124
$13
-$154
$225
$164
-$86
$152
$200
$89
$316
$796
$121
-$2,394
-$894
$1,099
$13
-$202
$225
$164
-$86
$152
$200
$89
$316
$774
$121
-$2,394
-$894
$1,073
$13
-$249
a The agencies have not estimated PHEV or EV costs for the large MPV and truck vehicle classes since we do not believe these vehicle
classes would use the technologies.
Table 3-100 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for EV100 for the 2010
Baseline (2010$)a
System
0% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Small car
$223
$140
-$60
$121
$197
Standard car
$229
$157
-$65
$152
$202
Large car
$232
$168
-$82
$177
$205
Small MPV
$225
$164
-$86
$162
$201
5-188
-------
Technologies Considered in the Agencies' Analysis
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
2% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
7.5% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
10% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
$76
$316
$729
$121
-$1,596
-$894
$1,021
$13
$406
$223
$140
-$60
$121
$197
$76
$316
$714
$121
-$1,596
-$894
$1,004
$13
$375
$223
$140
-$60
$121
$197
$76
$316
$674
$121
-$1,596
-$894
$958
$13
$289
$223
$140
-$60
$121
$197
$76
$316
$656
$121
-$1,596
-$894
$938
$13
$85
$316
$1,094
$121
-$1,596
-$894
$1,441
$13
$1,255
$229
$157
-$65
$152
$202
$85
$316
$1,072
$121
-$1,596
-$894
$1,416
$13
$1,208
$229
$157
-$65
$152
$202
$85
$316
$1,012
$121
-$1,596
-$894
$1,347
$13
$1,079
$229
$157
-$65
$152
$202
$85
$316
$985
$121
-$1,596
-$894
$1,315
$13
$91
$316
$1,932
$121
-$2,466
-$894
$2,402
$13
$2,214
$232
$168
-$82
$177
$205
$91
$316
$1,893
$121
-$2,466
-$894
$2,358
$13
$2,131
$232
$168
-$82
$177
$205
$91
$316
$1,787
$121
-$2,466
-$894
$2,236
$13
$1,903
$232
$168
-$82
$177
$205
$91
$316
$1,739
$121
-$2,466
-$894
$2,180
$13
$89
$316
$946
$121
-$2,394
-$894
$1,271
$13
$132
$225
$164
-$86
$162
$201
$89
$316
$927
$121
-$2,394
-$894
$1,249
$13
$92
$225
$164
-$86
$162
$201
$89
$316
$875
$121
-$2,394
-$894
$1,189
$13
-$20
$225
$164
-$86
$162
$201
$89
$316
$851
$121
-$2,394
-$894
$1,162
$13
3-189
-------
Technologies Considered in the Agencies' Analysis
Total
20% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
$250
$223
$140
-$60
$121
$197
$76
$316
$633
$121
-$1,596
-$894
$912
$13
$201
$1,020
$229
$157
-$65
$152
$202
$85
$316
$954
$121
-$1,596
-$894
$1,280
$13
$954
$1,799
$232
$168
-$82
$177
$205
$91
$316
$1,684
$121
-$2,466
-$894
$2,118
$13
$1,682
-$71
$225
$164
-$86
$162
$201
$89
$316
$829
$121
-$2,394
-$894
$1,137
$13
-$118
a The agencies have not estimated PHEV or EV costs for the large MPV and track vehicle classes since we do not believe these vehicle
classes would use the technologies.
Table 3-101 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for EV150 for the 2008
Baseline (2010$)a
System
0% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
2% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
7.5% WR
Small car
$223
$140
-$60
$121
$196
$76
$316
$703
$121
-$1,596
-$894
$992
$13
$351
$223
$140
-$60
$121
$196
$76
$316
$692
$121
-$1,596
-$894
$979
$13
$328
Standard car
$229
$157
-$65
$152
$201
$85
$316
$1,044
$121
-$1,596
-$894
$1,383
$13
$1,146
$229
$157
-$65
$152
$201
$85
$316
$1,028
$121
-$1,596
-$894
$1,364
$13
$1,111
Large car
$232
$168
-$82
$162
$204
$91
$316
$1,868
$121
-$2,466
-$894
$2,329
$13
$2,061
$232
$168
-$82
$162
$204
$91
$316
$1,837
$121
-$2,466
-$894
$2,293
$13
$1,995
Small MPV
$225
$164
-$86
$152
$200
$89
$316
$885
$121
-$2,394
-$894
$1,200
$13
-$11
$225
$164
-$86
$152
$200
$89
$316
$878
$121
-$2,394
-$894
$1,193
$13
-$26
5-190
-------
Technologies Considered in the Agencies' Analysis
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
10% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
20% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
$223
$140
-$60
$121
$196
$76
$316
$692
$121
-$1,596
-$894
$979
$13
$328
$223
$140
-$60
$121
$196
$76
$316
$692
$121
-$1,596
-$894
$979
$13
$328
$223
$140
-$60
$121
$196
$76
$316
$692
$121
-$1,596
-$894
$979
$13
$328
$229
$157
-$65
$152
$201
$85
$316
$1,028
$121
-$1,596
-$894
$1,364
$13
$1,111
$229
$157
-$65
$152
$201
$85
$316
$1,028
$121
-$1,596
-$894
$1,364
$13
$1,111
$229
$157
-$65
$152
$201
$85
$316
$1,028
$121
-$1,596
-$894
$1,364
$13
$1,111
$232
$168
-$82
$162
$204
$91
$316
$1,837
$121
-$2,466
-$894
$2,293
$13
$1,995
$232
$168
-$82
$162
$204
$91
$316
$1,837
$121
-$2,466
-$894
$2,293
$13
$1,995
$232
$168
-$82
$162
$204
$91
$316
$1,837
$121
-$2,466
-$894
$2,293
$13
$1,995
$225
$164
-$86
$152
$200
$89
$316
$878
$121
-$2,394
-$894
$1,193
$13
-$26
$225
$164
-$86
$152
$200
$89
$316
$878
$121
-$2,394
-$894
$1,193
$13
-$26
$225
$164
-$86
$152
$200
$89
$316
$878
$121
-$2,394
-$894
$1,193
$13
-$26
a The agencies have not estimated PHEV or EV costs for the large MPV and truck vehicle classes since we do not believe these vehicle
classes would use the technologies.
Table 3-102 Scaled Non-battery DMC by Applied Vehicle Weight Reduction for EV150 for the 2010
Baseline (2010$)a
System
0% WR
Brake system
Climate controls
Small car
$223
$140
Standard car
$229
$157
Large car
$232
$168
Small MPV
$225
$164
5-191
-------
Technologies Considered in the Agencies' Analysis
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
2% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
7.5% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
10% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
-$60
$121
$197
$76
$316
$729
$121
-$1,596
-$894
$1,021
$13
$406
$223
$140
-$60
$121
$197
$76
$316
$720
$121
-$1,596
-$894
$1,011
$13
$387
$223
$140
-$60
$121
$197
$76
$316
$720
$121
-$1,596
-$894
$1,011
$13
$387
$223
$140
-$60
$121
$197
$76
$316
$720
$121
-$1,596
-$65
$152
$202
$85
$316
$1,094
$121
-$1,596
-$894
$1,441
$13
$1,255
$229
$157
-$65
$152
$202
$85
$316
$1,081
$121
-$1,596
-$894
$1,425
$13
$1,226
$229
$157
-$65
$152
$202
$85
$316
$1,081
$121
-$1,596
-$894
$1,425
$13
$1,226
$229
$157
-$65
$152
$202
$85
$316
$1,081
$121
-$1,596
-$82
$177
$205
$91
$316
$1,932
$121
-$2,466
-$894
$2,402
$13
$2,214
$232
$168
-$82
$177
$205
$91
$316
$1,910
$121
-$2,466
-$894
$2,377
$13
$2,167
$232
$168
-$82
$177
$205
$91
$316
$1,910
$121
-$2,466
-$894
$2,377
$13
$2,167
$232
$168
-$82
$177
$205
$91
$316
$1,910
$121
-$2,466
-$86
$162
$201
$89
$316
$946
$121
-$2,394
-$894
$1,271
$13
$132
$225
$164
-$86
$162
$201
$89
$316
$941
$121
-$2,394
-$894
$1,265
$13
$121
$225
$164
-$86
$162
$201
$89
$316
$941
$121
-$2,394
-$894
$1,265
$13
$121
$225
$164
-$86
$162
$201
$89
$316
$941
$121
-$2,394
3-192
-------
Technologies Considered in the Agencies' Analysis
Delete transmission
Motor assembly
Battery discharge system
Total
20% WR
Brake system
Climate controls
Delete electrical
DC-DC converter
High voltage wiring
Supplemental heater
On vehicle charger
Motor inverter
Controls
Delete 1C engine
Delete transmission
Motor assembly
Battery discharge system
Total
-$894
$1,011
$13
$387
$223
$140
-$60
$121
$197
$76
$316
$720
$121
-$1,596
-$894
$1,011
$13
$387
-$894
$1,425
$13
$1,226
$229
$157
-$65
$152
$202
$85
$316
$1,081
$121
-$1,596
-$894
$1,425
$13
$1,226
-$894
$2,377
$13
$2,167
$232
$168
-$82
$177
$205
$91
$316
$1,910
$121
-$2,466
-$894
$2,377
$13
$2,167
-$894
$1,265
$13
$121
$225
$164
-$86
$162
$201
$89
$316
$941
$121
-$2,394
-$894
$1,265
$13
$121
a The agencies have not estimated PHEV
classes would use the technologies.
or EV costs for the large MPV and truck vehicle classes since we do not believe these vehicle
Similar to the approach taken for battery pack costs, the agencies generated linear
regressions of non-battery system costs against percent of net mass reduction and the results
are shown in Table 3-103. This was done using the same weight reduction offsets as used for
battery packs as presented in Table 3-75. The agencies separated battery pack costs from the
remainder of the systems for each type of electrified vehicle. The advantage of separating the
battery pack costs from other system costs is that it allows each to carry unique indirect cost
multipliers and learning effects which are important given that battery technology is an
emerging technology, while electric motors and inverters are more stable technologies.
Table 3-103 Linear Regressions of Non-Battery System Direct Manufacturing Costs vs Net Mass
reduction (2010$)
Vehicle
Class
P2HEV
PHEV20
PHEV40
EV75
EV100
EV150
2008 Baseline
Small car
Standard
car
Large car
Small
MPV
Large
MPV
Truck
-$263x+$ 1,675
-$391x+$l,857
-$699x+$2,175
-$331x+$l,777
-$506x+$2,052
-$648x+$2,169
-$l,316x+$2,878
-$l,953x+$3,575
-$3,495x+$5,129
-$l,655x+$3,258
-$l,316x+$2,878
-$l,953x+$3,575
-$3,495x+$5,129
-$l,655x+$3,258
-$l,510x+$350
-$2,242x+$l,145
-$4,012x+$2,060
-$l,900x+-$12
-$l,510x+$350
-$2,242x+$l,145
-$4,012x+$2,060
-$l,900x+-$12
-$l,510x+$351
-$2,242x+$l,146
-$4,012x+$2,061
-$l,900x+-$ll
20 10 Baseline
Small car
Standard
car
Large car
Small
MPV
Large
MPV
Truck
-$279x+$ 1,683
-$420x+$l,878
-$741x+$2,224
-$363x+$l,811
-$528x+$2,073
-$674x+$2,188
-$l,397x+$2,951
-$2,099x+$3,712
-$3,705x+$5,347
-$l,814x+$3,419
-$l,397x+$2,951
-$2,099x+$3,712
-$3,705x+$5,347
-$l,814x+$3,419
-$l,565x+$406
-$2,350x+$l,255
-$4,149x+$2,214
-$2,032x+$132
-$l,565x+$406
-$2,350x+$l,255
-$4,149x+$2,214
-$2,032x+$132
-$l,565x+$406
-$2,350x+$l,255
-$4,149x+$2,214
-$2,032x+$132
Notes:
3-193
-------
Technologies Considered in the Agencies' Analysis
"x" in the equations represents the net weight reduction as a percentage, so the non-battery components for a small car P2 HEV (2008
baseline) with a 20% applied weight reduction and, therefore, a 15% net weight reduction would cost (-$263)x(15%)+$l,675=$l,635.The
agencies did not regress PHEV or EV costs for the large MPV and truck vehicle classes since we do not believe these vehicle classes would
use the technologies.
For P2 HEV and PHEV non-battery components, the direct manufacturing costs
shown in Table 3-103 are considered applicable to the 2012MY. The agencies consider the
P2 and PHEV non-battery component technologies to be on the flat portion of the learning
curve during the 2017-2025 timeframe. The agencies have applied a highl complexity ICM
of 1.56 through 2018 then 1.35 thereafter. For EV non-battery components, the direct
manufacturing costs shown in Table 3-103 are considered applicable to the 2017MY. The
agencies consider the EV non-battery component technologies to be on the flat portion of the
learning curve during the 2017-2025 timeframe. The agencies have applied a high2
complexity ICM of 1.77 through 2024 then 1.50 thereafter. The resultant costs for P2 HEV,
PHEV20, PHEV40, EV75, EV100 and EV150 non-battery components for the 2008 and 2010
baselines are shown in Table 3-104 through Table 3-115, respectively.aaa
Table 3-104 Costs for P2 HEV Non-Battery Components for the 2008 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
Vehicle class
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
Small car
Applied
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
Net
WR
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
11%
16%
5%
2017
$1,442
$1,430
$1,419
$1,594
$1,577
$1,560
$1,857
$1,826
$1,796
$1,528
$1,513
$1,499
$1,759
$1,737
$1,715
$1,848
$1,820
$1,792
$922
2018
$1,413
$1,402
$1,391
$1,562
$1,546
$1,529
$1,820
$1,790
$1,760
$1,497
$1,483
$1,469
$1,723
$1,702
$1,680
$1,811
$1,784
$1,756
$920
2019
$1,385
$1,374
$1,363
$1,531
$1,515
$1,498
$1,783
$1,754
$1,725
$1,467
$1,453
$1,440
$1,689
$1,668
$1,647
$1,775
$1,748
$1,721
$565
2020
$1,357
$1,346
$1,335
$1,500
$1,484
$1,468
$1,747
$1,719
$1,690
$1,438
$1,424
$1,411
$1,655
$1,634
$1,614
$1,739
$1,713
$1,686
$564
2021
$1,330
$1,319
$1,309
$1,470
$1,455
$1,439
$1,713
$1,685
$1,657
$1,409
$1,396
$1,383
$1,622
$1,602
$1,582
$1,705
$1,679
$1,653
$563
2022
$1,303
$1,293
$1,283
$1,441
$1,426
$1,410
$1,678
$1,651
$1,623
$1,381
$1,368
$1,355
$1,590
$1,570
$1,550
$1,670
$1,645
$1,620
$563
2023
$1,277
$1,267
$1,257
$1,412
$1,397
$1,382
$1,645
$1,618
$1,591
$1,353
$1,340
$1,328
$1,558
$1,538
$1,519
$1,637
$1,612
$1,587
$562
2024
$1,252
$1,242
$1,232
$1,384
$1,369
$1,354
$1,612
$1,585
$1,559
$1,326
$1,314
$1,301
$1,527
$1,508
$1,489
$1,604
$1,580
$1,556
$561
2025
$1,227
$1,217
$1,207
$1,356
$1,342
$1,327
$1,580
$1,554
$1,528
$1,300
$1,287
$1,275
$1,496
$1,477
$1,459
$1,572
$1,548
$1,524
$560
aaa Note that, in the draft Joint TSD, we inadvertently stated the following with respect to the years in which costs
were considered valid and the years for which near term and long term ICMs were applied: "For P2 HEV non-
battery components, the direct manufacturing costs shown in Table 3-103 are considered applicable to the
2017MY. The agencies consider the P2 non-battery component technologies to be on the flat portion of the
learning curve during the 2017-2025 timeframe. The agencies have applied a highl complexity ICM of 1.56
through 2018 then 1.35 thereafter. For PHEV and EV non-battery components, the direct manufacturing costs
shown in Table 3-103 are considered applicable to the 2025MY. The agencies consider the PHEV and EV non-
battery component technologies to be on the flat portion of the learning curve during the 2017-2025 timeframe.
The agencies have applied a high2 complexity ICM of 1.77 through 2024 then 1.50 thereafter." Importantly, the
costs then (and now) were calculated according to the corrected text shown in this final Joint TSD.
3-194
-------
Technologies Considered in the Agencies' Analysis
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
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
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%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
11%
16%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
11%
16%
$915
$908
$1,020
$1,009
$998
$1,188
$1,168
$1,149
$977
$968
$959
$1,125
$1,111
$1,097
$1,182
$1,164
$1,146
$2,364
$2,345
$2,327
$2,614
$2,586
$2,558
$3,044
$2,995
$2,945
$2,505
$2,481
$2,458
$2,884
$2,848
$2,812
$3,030
$2,984
$2,938
$913
$906
$1,018
$1,007
$996
$1,185
$1,166
$1,147
$975
$966
$957
$1,123
$1,109
$1,095
$1,180
$1,162
$1,144
$2,333
$2,315
$2,296
$2,580
$2,552
$2,525
$3,005
$2,956
$2,907
$2,472
$2,449
$2,426
$2,846
$2,811
$2,775
$2,991
$2,946
$2,900
$561
$556
$625
$618
$612
$728
$716
$704
$599
$593
$588
$689
$681
$672
$724
$713
$702
$1,950
$1,934
$1,919
$2,156
$2,133
$2,110
$2,511
$2,470
$2,429
$2,066
$2,046
$2,027
$2,378
$2,349
$2,319
$2,499
$2,461
$2,423
$560
$555
$624
$617
$611
$727
$715
$703
$598
$592
$587
$688
$680
$671
$723
$712
$701
$1,921
$1,906
$1,891
$2,124
$2,102
$2,079
$2,474
$2,434
$2,393
$2,036
$2,016
$1,997
$2,343
$2,314
$2,285
$2,463
$2,425
$2,388
$559
$554
$623
$616
$610
$726
$714
$702
$597
$591
$586
$687
$679
$670
$722
$711
$700
$1,893
$1,878
$1,863
$2,093
$2,071
$2,049
$2,438
$2,398
$2,358
$2,006
$1,987
$1,968
$2,309
$2,280
$2,252
$2,427
$2,390
$2,353
$558
$554
$622
$615
$609
$724
$713
$701
$596
$590
$585
$686
$678
$669
$721
$710
$699
$1,866
$1,851
$1,836
$2,063
$2,041
$2,019
$2,403
$2,364
$2,324
$1,977
$1,958
$1,940
$2,276
$2,247
$2,219
$2,392
$2,355
$2,319
$557
$553
$621
$614
$608
$723
$712
$700
$595
$590
$584
$685
$677
$668
$720
$709
$698
$1,839
$1,824
$1,810
$2,033
$2,012
$1,990
$2,368
$2,329
$2,291
$1,948
$1,930
$1,912
$2,243
$2,215
$2,187
$2,357
$2,321
$2,285
$556
$552
$620
$614
$607
$722
$711
$699
$594
$589
$583
$684
$676
$667
$719
$708
$697
$1,813
$1,798
$1,784
$2,004
$1,983
$1,961
$2,334
$2,296
$2,258
$1,920
$1,902
$1,884
$2,211
$2,183
$2,156
$2,323
$2,288
$2,253
$556
$551
$619
$613
$606
$721
$710
$698
$593
$588
$582
$683
$675
$666
$718
$707
$696
$1,787
$1,773
$1,758
$1,975
$1,954
$1,933
$2,301
$2,263
$2,226
$1,893
$1,875
$1,857
$2,179
$2,152
$2,125
$2,290
$2,255
$2,221
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-105 Costs for P2 HEV Non-Battery Components for the 2010 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
Vehicle class
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Applied
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
Net
WR
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
2017
$1,448
$1,436
$1,423
$1,611
$1,593
$1,574
$1,898
$1,866
$1,833
$1,555
$1,540
$1,524
$1,776
$1,753
$1,730
$1,863
2018
$1,419
$1,407
$1,395
$1,579
$1,561
$1,543
$1,860
$1,828
$1,797
$1,524
$1,509
$1,493
$1,740
$1,718
$1,696
$1,826
2019
$1,390
$1,379
$1,367
$1,547
$1,529
$1,512
$1,823
$1,792
$1,761
$1,494
$1,479
$1,464
$1,706
$1,684
$1,662
$1,790
2020
$1,363
$1,351
$1,340
$1,516
$1,499
$1,482
$1,786
$1,756
$1,726
$1,464
$1,449
$1,434
$1,672
$1,650
$1,628
$1,754
2021
$1,335
$1,324
$1,313
$1,486
$1,469
$1,452
$1,750
$1,721
$1,691
$1,435
$1,420
$1,406
$1,638
$1,617
$1,596
$1,719
2022
$1,309
$1,298
$1,287
$1,456
$1,440
$1,423
$1,715
$1,686
$1,657
$1,406
$1,392
$1,377
$1,605
$1,585
$1,564
$1,684
2023
$1,282
$1,272
$1,261
$1,427
$1,411
$1,395
$1,681
$1,653
$1,624
$1,378
$1,364
$1,350
$1,573
$1,553
$1,533
$1,651
2024
$1,257
$1,246
$1,236
$1,398
$1,383
$1,367
$1,647
$1,620
$1,592
$1,350
$1,337
$1,323
$1,542
$1,522
$1,502
$1,618
2025
$1,232
$1,221
$1,211
$1,370
$1,355
$1,339
$1,614
$1,587
$1,560
$1,323
$1,310
$1,296
$1,511
$1,491
$1,472
$1,585
3-195
-------
Technologies Considered in the Agencies' Analysis
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
Truck
Truck
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Large MPV
Large MPV
Large MPV
Truck
Truck
Truck
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%
11%
16%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
11%
16%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
5%
10%
15%
6%
11%
16%
$1,834
$1,805
$926
$918
$911
$1,030
$1,019
$1,007
$1,214
$1,193
$1,173
$995
$985
$975
$1,136
$1,122
$1,107
$1,192
$1,173
$1,155
$2,374
$2,354
$2,334
$2,641
$2,611
$2,582
$3,112
$3,059
$3,006
$2,550
$2,525
$2,499
$2,912
$2,875
$2,837
$3,056
$3,008
$2,960
$1,798
$1,769
$924
$917
$909
$1,028
$1,017
$1,005
$1,212
$1,191
$1,171
$993
$983
$973
$1,134
$1,119
$1,105
$1,190
$1,171
$1,152
$2,343
$2,323
$2,304
$2,607
$2,577
$2,548
$3,071
$3,019
$2,967
$2,517
$2,492
$2,466
$2,874
$2,837
$2,800
$3,016
$2,969
$2,921
$1,762
$1,733
$568
$563
$558
$631
$624
$617
$744
$731
$719
$610
$604
$597
$696
$687
$678
$730
$719
$708
$1,958
$1,942
$1,925
$2,178
$2,154
$2,129
$2,566
$2,523
$2,479
$2,103
$2,082
$2,061
$2,402
$2,371
$2,340
$2,520
$2,481
$2,441
$1,726
$1,699
$567
$562
$557
$630
$623
$616
$743
$730
$718
$609
$603
$596
$695
$686
$677
$729
$718
$706
$1,929
$1,913
$1,897
$2,146
$2,122
$2,098
$2,529
$2,486
$2,443
$2,073
$2,052
$2,031
$2,367
$2,336
$2,306
$2,483
$2,444
$2,405
$1,692
$1,665
$566
$561
$556
$629
$622
$615
$742
$729
$716
$608
$602
$596
$694
$685
$676
$728
$717
$705
$1,901
$1,885
$1,869
$2,115
$2,091
$2,067
$2,492
$2,450
$2,408
$2,042
$2,022
$2,001
$2,332
$2,302
$2,272
$2,447
$2,409
$2,370
$1,658
$1,632
$565
$560
$555
$629
$621
$614
$740
$728
$715
$607
$601
$595
$693
$684
$675
$727
$716
$704
$1,874
$1,858
$1,842
$2,085
$2,061
$2,037
$2,456
$2,414
$2,373
$2,013
$1,992
$1,972
$2,298
$2,269
$2,239
$2,412
$2,374
$2,336
$1,625
$1,599
$564
$559
$555
$628
$620
$613
$739
$727
$714
$606
$600
$594
$692
$683
$674
$726
$715
$703
$1,847
$1,831
$1,816
$2,054
$2,031
$2,008
$2,420
$2,379
$2,338
$1,984
$1,964
$1,944
$2,265
$2,236
$2,207
$2,377
$2,339
$2,302
$1,592
$1,567
$563
$559
$554
$627
$620
$613
$738
$726
$713
$605
$599
$593
$691
$682
$673
$725
$714
$702
$1,820
$1,805
$1,790
$2,025
$2,002
$1,979
$2,386
$2,345
$2,305
$1,955
$1,936
$1,916
$2,233
$2,204
$2,175
$2,343
$2,306
$2,269
$1,560
$1,536
$562
$558
$553
$626
$619
$612
$737
$725
$712
$604
$598
$592
$690
$681
$672
$724
$713
$701
$1,794
$1,779
$1,764
$1,996
$1,974
$1,951
$2,352
$2,312
$2,272
$1,927
$1,908
$1,888
$2,201
$2,173
$2,144
$2,309
$2,273
$2,237
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-106 Costs for PHEV20 Non-Battery Components for the 2008 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
Vehicle
class
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Applied
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
NetWR
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
2017
$2,463
$2,406
$2,349
$3,050
$2,966
$2,881
$4,389
$4,238
$4,086
$2,784
2018
$2,414
$2,358
$2,302
$2,989
$2,906
$2,823
$4,301
$4,153
$4,004
$2,728
2019
$2,365
$2,311
$2,256
$2,930
$2,848
$2,767
$4,215
$4,070
$3,924
$2,673
2020
$2,318
$2,264
$2,211
$2,871
$2,791
$2,712
$4,131
$3,988
$3,846
$2,620
2021
$2,272
$2,219
$2,166
$2,814
$2,735
$2,657
$4,049
$3,909
$3,769
$2,568
2022
$2,226
$2,175
$2,123
$2,757
$2,681
$2,604
$3,968
$3,831
$3,693
$2,516
2023
$2,182
$2,131
$2,081
$2,702
$2,627
$2,552
$3,888
$3,754
$3,620
$2,466
2024
$2,138
$2,089
$2,039
$2,648
$2,575
$2,501
$3,810
$3,679
$3,547
$2,417
2025
$2,095
$2,047
$1,998
$2,595
$2,523
$2,451
$3,734
$3,605
$3,476
$2,368
3-196
-------
Technologies Considered in the Agencies' Analysis
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
Small MPV
Small MPV
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
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%
8%
13%
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
$2,712
$2,640
$1,576
$1,539
$1,503
$1,951
$1,897
$1,843
$2,808
$2,711
$2,614
$1,781
$1,735
$1,689
$4,039
$3,945
$3,851
$5,002
$4,863
$4,724
$7,197
$6,949
$6,700
$4,565
$4,447
$4,329
$2,658
$2,587
$1,572
$1,536
$1,500
$1,948
$1,893
$1,839
$2,802
$2,706
$2,609
$1,777
$1,731
$1,686
$3,986
$3,894
$3,801
$4,937
$4,800
$4,663
$7,104
$6,858
$6,613
$4,505
$4,389
$4,273
$2,604
$2,536
$965
$943
$921
$1,196
$1,163
$1,129
$1,721
$1,661
$1,602
$1,091
$1,063
$1,035
$3,331
$3,254
$3,177
$4,125
$4,011
$3,896
$5,936
$5,731
$5,526
$3,765
$3,668
$3,570
$2,552
$2,485
$964
$942
$919
$1,194
$1,161
$1,128
$1,718
$1,659
$1,599
$1,089
$1,061
$1,033
$3,282
$3,206
$3,130
$4,065
$3,952
$3,839
$5,849
$5,647
$5,445
$3,709
$3,614
$3,518
$2,501
$2,435
$963
$940
$918
$1,192
$1,159
$1,126
$1,715
$1,656
$1,597
$1,088
$1,060
$1,032
$3,234
$3,159
$3,084
$4,006
$3,894
$3,783
$5,764
$5,565
$5,366
$3,655
$3,561
$3,467
$2,451
$2,386
$961
$939
$917
$1,190
$1,157
$1,124
$1,713
$1,654
$1,594
$1,086
$1,058
$1,030
$3,187
$3,114
$3,040
$3,948
$3,838
$3,728
$5,680
$5,484
$5,288
$3,602
$3,510
$3,417
$2,402
$2,339
$960
$937
$915
$1,189
$1,156
$1,123
$1,710
$1,651
$1,592
$1,085
$1,057
$1,029
$3,141
$3,069
$2,996
$3,891
$3,783
$3,675
$5,598
$5,405
$5,212
$3,550
$3,459
$3,367
$2,354
$2,292
$958
$936
$914
$1,187
$1,154
$1,121
$1,708
$1,649
$1,590
$1,083
$1,055
$1,027
$3,096
$3,025
$2,953
$3,835
$3,728
$3,622
$5,518
$5,328
$5,137
$3,500
$3,409
$3,319
$2,307
$2,246
$957
$935
$913
$1,185
$1,152
$1,119
$1,705
$1,646
$1,587
$1,081
$1,054
$1,026
$3,052
$2,982
$2,911
$3,780
$3,675
$3,570
$5,440
$5,252
$5,064
$3,450
$3,361
$3,272
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-107 Costs for PHEV20 Non-Battery Components for the 2010 Baseline (2010$)
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
Vehicle
class
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Applied
WR
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%
NetWR
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
3%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
3%
2017
$2,524
$2,464
$2,403
$3,166
$3,075
$2,984
$4,574
$4,414
$4,253
$2,919
$2,841
$2,762
$1,615
$1,576
$1,537
$2,025
$1,967
$1,909
$2,926
$2,824
$2,721
$1,868
$1,817
$1,767
$4,139
2018
$2,474
$2,414
$2,355
$3,102
$3,013
$2,924
$4,483
$4,325
$4,168
$2,861
$2,784
$2,707
$1,612
$1,573
$1,534
$2,021
$1,963
$1,905
$2,920
$2,818
$2,715
$1,864
$1,814
$1,763
$4,085
2019
$2,424
$2,366
$2,308
$3,040
$2,953
$2,865
$4,393
$4,239
$4,084
$2,804
$2,728
$2,653
$990
$966
$942
$1,241
$1,205
$1,170
$1,793
$1,730
$1,667
$1,144
$1,114
$1,083
$3,414
2020
$2,376
$2,319
$2,262
$2,979
$2,894
$2,808
$4,305
$4,154
$4,003
$2,748
$2,674
$2,600
$988
$964
$941
$1,239
$1,203
$1,168
$1,790
$1,727
$1,664
$1,143
$1,112
$1,081
$3,364
2021
$2,328
$2,272
$2,217
$2,920
$2,836
$2,752
$4,219
$4,071
$3,923
$2,693
$2,620
$2,548
$986
$963
$939
$1,237
$1,202
$1,166
$1,788
$1,725
$1,662
$1,141
$1,110
$1,079
$3,315
2022
$2,282
$2,227
$2,172
$2,861
$2,779
$2,697
$4,135
$3,989
$3,844
$2,639
$2,568
$2,497
$985
$961
$938
$1,235
$1,200
$1,164
$1,785
$1,722
$1,659
$1,139
$1,108
$1,078
$3,267
2023
$2,236
$2,182
$2,129
$2,804
$2,724
$2,643
$4,052
$3,910
$3,767
$2,586
$2,516
$2,447
$984
$960
$936
$1,233
$1,198
$1,162
$1,782
$1,720
$1,657
$1,137
$1,107
$1,076
$3,220
2024
$2,191
$2,139
$2,086
$2,748
$2,669
$2,590
$3,971
$3,832
$3,692
$2,534
$2,466
$2,398
$982
$959
$935
$1,232
$1,196
$1,161
$1,780
$1,717
$1,655
$1,136
$1,105
$1,075
$3,174
2025
$2,148
$2,096
$2,044
$2,693
$2,616
$2,538
$3,892
$3,755
$3,618
$2,484
$2,417
$2,350
$981
$957
$934
$1,230
$1,195
$1,159
$1,777
$1,715
$1,652
$1,134
$1,104
$1,073
$3,128
3-197
-------
Technologies Considered in the Agencies' Analysis
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
Small car
Small car
Standard car
Standard car
Standard car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
8%
13%
3%
8%
13%
2%
7%
12%
3%
8%
13%
$4,040
$3,940
$5,191
$5,042
$4,892
$7,501
$7,237
$6,974
$4,787
$4,658
$4,529
$3,987
$3,889
$5,123
$4,976
$4,829
$7,403
$7,143
$6,883
$4,725
$4,597
$4,470
$3,332
$3,250
$4,281
$4,158
$4,035
$6,186
$5,969
$5,752
$3,948
$3,842
$3,735
$3,283
$3,202
$4,218
$4,097
$3,976
$6,096
$5,881
$5,667
$3,890
$3,785
$3,681
$3,235
$3,156
$4,157
$4,037
$3,918
$6,007
$5,796
$5,585
$3,834
$3,730
$3,627
$3,188
$3,110
$4,097
$3,979
$3,861
$5,920
$5,712
$5,504
$3,778
$3,676
$3,574
$3,142
$3,065
$4,038
$3,922
$3,805
$5,834
$5,629
$5,424
$3,724
$3,623
$3,523
$3,097
$3,021
$3,980
$3,865
$3,751
$5,751
$5,549
$5,347
$3,670
$3,571
$3,472
$3,053
$2,978
$3,923
$3,810
$3,697
$5,669
$5,470
$5,270
$3,618
$3,520
$3,423
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-108 Costs for PHEV40 Non-Battery Components for the 2008 Baseline (2010$)
Cost
type
Vehicle
class
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
Applied
WR
Small car
Small car
Stdcar
Stdcar
Large car
Large car
Small MPV
Small MPV
Small car
Small car
Stdcar
Stdcar
Large car
Large car
Small MPV
Small MPV
Small car
Small car
Stdcar
Stdcar
Large car
Large car
Small MPV
Small MPV
Net
WR
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
2017
2%
7%
3%
8%
1%
6%
3%
8%
2%
7%
3%
8%
1%
6%
3%
8%
2%
7%
3%
8%
1%
6%
3%
8%
2018
$2,474
$2,417
$3
$2
050
966
$4,420
$4
$2
$2
$1
$1
$1
$1
$2
$2
$1
$1
$4
$3
$5
$4
268
784
712
583
546
951
897
827
730
781
735
057
964
002
863
$7,247
$6
$4
998
565
$4,447
2019
$2,425
$2,369
$2
989
$2,906
$4,331
$4
183
$2,728
$2,658
$1,580
$1,543
$1
948
$1,893
$2
822
$2,725
$1,777
$1,731
$4,005
$3
912
$4,937
$4,800
$7
153
$6,907
$4
$4
505
389
$2
$2
$2
$2
2020
376
322
930
848
$4,245
$4
$2
$2
099
673
604
$970
$948
$1
$1
$1
$1
$1
$1
$3
$3
$4
$4
$5
$5
$3
$3
196
163
732
673
091
063
346
269
125
Oil
977
772
765
668
$2
2021
329
$2,275
$2
871
$2,791
$4,160
$4,017
$2,620
$2
552
$968
$946
$1
194
$1,161
$1,730
$1,670
$1,089
$1,061
$3,297
$3,221
$4,065
$3,952
$5
889
$5,687
$3,709
$3,614
2022
$2,282
$2,230
$2
814
$2,735
$4,076
$3
937
$2,568
$2
501
$967
$945
$1
192
$1,159
$1,727
$1,668
$1,088
$1,060
$3,249
$3
174
$4,006
$3,894
$5
804
$5,605
$3,655
$3
561
2023
$2,237
$2,185
$2,757
$2,681
$3,995
$3
858
$2,516
$2,451
$966
$943
$1
190
$1,157
$1,725
$1,665
$1,086
$1,058
$3,202
$3
128
$3,948
$3,838
$5,719
$5,523
$3,602
$3
510
$2
2024
192
$2,141
$2,702
$2,627
$3,915
$3,781
$2,466
$2,402
$964
$942
$1
189
$1,156
$1,722
$1,663
$1,085
$1,057
$3,156
$3,083
$3,891
$3,783
$5,637
$5,444
$3,550
$3,459
$2
2025
148
$2,099
$2,648
$2,575
$3,837
$3,705
$2,417
$2
354
$963
$940
$1
187
$1,154
$1,719
$1,661
$1,083
$1,055
$3,111
$3,039
$3,835
$3,728
$5
556
$5,366
$3
500
$3,409
$2
$2
$2
$2
$3
$3
$2
$2
105
057
595
523
760
631
368
307
$961
$939
$1
$1
$1
$1
$1
$1
$3
$2
$3
$3
185
152
717
658
081
054
066
996
780
675
$5,477
$5,289
$3,450
$3
361
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-109 Costs for PHEV40 Non-Battery Components for the 2010 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
Vehicle
class
Small car
Small car
Stdcar
Std car
Large car
Large car
Small MPV
Small MPV
Applied
WR
15%
20%
15%
20%
15%
20%
15%
20%
Net
WR
3%
8%
3%
8%
2%
7%
3%
8%
2017
$2,524
$2,464
$3,166
$3,075
$4,574
$4,414
$2,919
$2,841
2018
$2,474
$2,414
$3,102
$3,013
$4,483
$4,325
$2,861
$2,784
2019
$2,424
$2,366
$3,040
$2,953
$4,393
$4,239
$2,804
$2,728
2020
$2,376
$2,319
$2,979
$2,894
$4,305
$4,154
$2,748
$2,674
2021
$2,328
$2,272
$2,920
$2,836
$4,219
$4,071
$2,693
$2,620
2022
$2,282
$2,227
$2,861
$2,779
$4,135
$3,989
$2,639
$2,568
2023
$2,236
$2,182
$2,804
$2,724
$4,052
$3,910
$2,586
$2,516
2024
$2,191
$2,139
$2,748
$2,669
$3,971
$3,832
$2,534
$2,466
2025
$2,148
$2,096
$2,693
$2,616
$3,892
$3,755
$2,484
$2,417
3-198
-------
Technologies Considered in the Agencies' Analysis
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
Small car
Small car
Stdcar
Std car
Large car
Large car
Small MPV
Small MPV
Small car
Small car
Std car
Std car
Large car
Large car
Small MPV
Small MPV
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
15%
20%
3%
8%
3%
8%
2%
7%
3%
8%
3%
8%
3%
8%
2%
7%
3%
8%
$1,615
$1,576
$2,025
$1,967
$2,926
$2,824
$1,868
$1,817
$4,139
$4,040
$5,191
$5,042
$7,501
$7,237
$4,787
$4,658
$1,612
$1,573
$2,021
$1,963
$2,920
$2,818
$1,864
$1,814
$4,085
$3,987
$5,123
$4,976
$7,403
$7,143
$4,725
$4,597
$990
$966
$1,241
$1,205
$1,793
$1,730
$1,144
$1,114
$3,414
$3,332
$4,281
$4,158
$6,186
$5,969
$3,948
$3,842
$988
$964
$1,239
$1,203
$1,790
$1,727
$1,143
$1,112
$3,364
$3,283
$4,218
$4,097
$6,096
$5,881
$3,890
$3,785
$986
$963
$1,237
$1,202
$1,788
$1,725
$1,141
$1,110
$3,315
$3,235
$4,157
$4,037
$6,007
$5,796
$3,834
$3,730
$985
$961
$1,235
$1,200
$1,785
$1,722
$1,139
$1,108
$3,267
$3,188
$4,097
$3,979
$5,920
$5,712
$3,778
$3,676
$984
$960
$1,233
$1,198
$1,782
$1,720
$1,137
$1,107
$3,220
$3,142
$4,038
$3,922
$5,834
$5,629
$3,724
$3,623
$982
$959
$1,232
$1,196
$1,780
$1,717
$1,136
$1,105
$3,174
$3,097
$3,980
$3,865
$5,751
$5,549
$3,670
$3,571
$981
$957
$1,230
$1,195
$1,777
$1,715
$1,134
$1,104
$3,128
$3,053
$3,923
$3,810
$5,669
$5,470
$3,618
$3,520
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-110 Costs for EV75 Non-Battery Components for the 2008 Baseline (2010$)
Cost
type
Vehicle
class
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
Applied
WR
Small car
Small car
Small car
Stdcar
Stdcar
Stdcar
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Stdcar
Stdcar
Stdcar
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Stdcar
Stdcar
Stdcar
Large car
Large car
Large car
Small MPV
Small MPV
Net 2017
WR
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%
10%
15%
20%
10%
15%
20%
10%
15%
20%
9%
14%
19%
10%
15%
20%
10%
15%
20%
10%
15%
20%
9%
14%
19%
10%
15%
20%
10%
15%
20%
10%
15%
20%
9%
14%
2018
$199
$124
$48
$921
$809
$697
$1,659
$1,458
$1,257
-$183
-$278
-$373
$153
$95
$37
$709
$623
$536
$1,277
$1,123
$968
-$141
-$214
-$287
$352
$219
$85
$1,630
$1,431
$1,233
$2,936
$2,581
$2,226
-$324
-$492
2019
$193
$120
$47
$893
$784
$676
$1,609
$1,414
$1,220
-$177
-$269
-$362
$153
$95
$37
$707
$621
$535
$1,273
$1,119
$965
-$140
-$213
-$286
$346
$215
$84
$1,600
$1,405
$1,211
$2,882
$2,534
$2,185
-$318
-$483
2020
$187
$116
$45
$866
$761
$655
$1,560
$1,372
$1,183
-$172
-$261
-$351
$152
$95
$37
$705
$619
$533
$1,270
$1,116
$963
-$140
-$213
-$285
$340
$211
$82
$1,571
$1,380
$1,189
$2,830
$2,488
$2,146
-$312
-$474
2021
$182
$113
$44
$840
$738
$636
$1
$1
$1
,514
,331
,148
-$167
-$254
-$340
$152
$94
$37
$703
$618
$532
$1
$1
,266
,113
$960
-$140
-$212
-$285
$334
$207
$81
$1
$1
$1
,543
,356
,168
$2,780
$2,444
$2,108
-$306
-$466
2022
$176
$109
$43
$815
$716
$617
$1,468
$1,291
$1,113
-$162
-$246
-$330
$152
$94
$37
$701
$616
$531
$1,263
$1,110
$958
-$139
-$212
-$284
$328
$204
$79
$1,516
$1,332
$1,147
$2,731
$2,401
$2,071
-$301
-$458
2023
$171
$106
$41
$791
$694
$598
$1,424
$1
$1
,252
,080
-$157
-$239
-$320
$151
$94
$37
$699
$614
$529
$1
$1
,260
,107
$955
-$139
-$211
-$283
$322
$200
$78
$1,490
$1
$1
,309
,127
$2,684
$2,359
$2,035
-$296
-$450
2024
$168
$104
$40
$775
$681
$586
$1,396
$1,227
$1,058
-$154
-$234
-$314
$151
$94
$36
$698
$613
$528
$1,258
$1,106
$954
-$139
-$211
-$283
$319
$198
$77
$1,473
$1,294
$1,114
$2,654
$2,333
$2,012
-$293
-$444
2025
$164
$102
$40
$759
$667
$574
$1
$1
$1
,368
,202
,037
-$151
-$229
-$307
$151
$94
$36
$697
$612
$527
$1
$1
,256
,104
$952
-$138
-$210
-$282
$315
$195
$76
$1,457
$1
$1
,279
,102
$2,624
$2,306
$1
,989
-$289
-$439
$161
$100
$39
$744
$654
$563
$1
$1
$1
,340
,178
,016
-$148
-$225
-$301
$97
$60
$23
$449
$394
$339
$808
$710
$613
-$89
-$135
-$182
$258
$160
$62
$1
$1
,193
,048
$902
$2,149
$1
$1
,889
,629
-$237
-$360
3-199
-------
Technologies Considered in the Agencies' Analysis
TC
Small MPV
20% 19%
-$660
-$636
-$625
-$614 -$603 -$596
-$590 -$483
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-111 Costs for EV75 Non-Battery Components for the 2010 Baseline (2010$)
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
Vehicle
class
Small car
Small car
Small car
Std car
Std car
Std car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Std car
Std car
Std car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Std car
Std car
Std car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Applied
WR
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%
Net
WR
10%
15%
20%
9%
14%
19%
10%
15%
20%
9%
14%
19%
10%
15%
20%
9%
14%
19%
10%
15%
20%
9%
14%
19%
10%
15%
20%
9%
14%
19%
10%
15%
20%
9%
14%
19%
2017
$250
$172
$93
$1,043
$926
$808
$1,799
$1,592
$1,385
-$51
-$152
-$254
$192
$132
$72
$803
$713
$623
$1,386
$1,226
$1,066
-$39
-$117
-$195
$442
$304
$165
$1,847
$1,639
$1,431
$3,185
$2,818
$2,451
-$90
-$270
-$449
2018
$242
$166
$91
$1,012
$898
$784
$1,745
$1,544
$1,343
-$49
-$148
-$246
$192
$132
$72
$801
$711
$621
$1,382
$1,222
$1,063
-$39
-$117
-$195
$434
$298
$162
$1,813
$1,609
$1,405
$3,127
$2,767
$2,406
-$88
-$265
-$441
2019
$235
$161
$88
$982
$871
$761
$1,693
$1,498
$1,303
-$48
-$143
-$239
$191
$131
$72
$799
$709
$619
$1,378
$1,219
$1,060
-$39
-$117
-$194
$426
$293
$159
$1,781
$1,580
$1,380
$3,071
$2,717
$2,363
-$86
-$260
-$433
2020
$228
$157
$85
$952
$845
$738
$1,642
$1,453
$1,264
-$46
-$139
-$232
$191
$131
$71
$797
$707
$617
$1,374
$1,216
$1,057
-$39
-$116
-$194
$419
$288
$157
$1,749
$1,552
$1,355
$3,016
$2,669
$2,321
-$85
-$255
-$426
2021
$221
$152
$83
$924
$820
$716
$1,593
$1,409
$1,226
-$45
-$135
-$225
$190
$131
$71
$795
$705
$616
$1,370
$1,212
$1,054
-$39
-$116
-$193
$412
$283
$154
$1,718
$1,525
$1,331
$2,963
$2,622
$2,280
-$83
-$251
-$418
2022
$215
$147
$80
$896
$795
$694
$1,545
$1,367
$1,189
-$44
-$131
-$218
$190
$130
$71
$793
$703
$614
$1,367
$1,209
$1,052
-$38
-$116
-$193
$404
$278
$151
$1,689
$1,498
$1,308
$2,912
$2,576
$2,241
-$82
-$246
-$411
2023
$210
$144
$79
$878
$779
$680
$1,514
$1,340
$1,165
-$43
-$128
-$214
$189
$130
$71
$791
$702
$613
$1,365
$1,207
$1,050
-$38
-$115
-$193
$400
$275
$149
$1,669
$1,481
$1,293
$2,879
$2,547
$2,215
-$81
-$244
-$406
2024
$206
$142
$77
$861
$764
$667
$1,484
$1,313
$1,142
-$42
-$126
-$209
$189
$130
$71
$790
$701
$612
$1,362
$1,205
$1,048
-$38
-$115
-$192
$395
$271
$148
$1,651
$1,465
$1,279
$2,846
$2,518
$2,190
-$80
-$241
-$402
2025
$202
$139
$75
$843
$748
$653
$1,454
$1,287
$1,119
-$41
-$123
-$205
$122
$84
$46
$508
$451
$394
$877
$776
$675
-$25
-$74
-$124
$324
$222
$121
$1,352
$1,199
$1,047
$2,331
$2,062
$1,794
-$66
-$197
-$329
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-112 Costs for EV100 Non-Battery Components for the 2008 Baseline (2010$)
Cost
type
EPA
Vehicle
class
DMC
DMC
DMC
DMC
DMC
Applied
WR
Small car
Small car
Small car
Std car
Std car
Net
WR
10%
15%
20%
10%
15%
2017
4%
9%
14%
4%
9%
2018
$290
$214
$139
$1,055
$943
2019
$281
$208
$135
$1,024
$915
2020
$273
$202
$130
$993
$887
2021
$264
$195
$127
$963
$861
2022
$256
$190
$123
$934
$835
2023
$249
$184
$119
$906
$810
2024
$244
$180
$117
$888
$794
2025
$239
$177
$114
$870
$778
$234
$173
$112
$853
$762
3-200
-------
Technologies Considered in the Agencies' Analysis
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
Stdcar
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Stdcar
Stdcar
Stdcar
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Stdcar
Stdcar
Stdcar
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
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%
14%
5%
10%
15%
3%
8%
13%
4%
9%
14%
4%
9%
14%
5%
10%
15%
3%
8%
13%
4%
9%
14%
4%
9%
14%
5%
10%
15%
3%
8%
13%
$831
$1,859
$1,659
$1,458
-$69
-$164
-$259
$223
$165
$107
$813
$726
$640
$1,432
$1,277
$1,123
-$53
-$126
-$199
$513
$379
$245
$1,868
$1,669
$1,471
$3,291
$2,936
$2,581
-$122
-$290
-$458
$806
$1,803
$1,609
$1,414
-$67
-$159
-$251
$222
$164
$106
$810
$724
$638
$1,427
$1,273
$1,119
-$53
-$126
-$199
$503
$372
$241
$1,834
$1,639
$1,444
$3,231
$2,882
$2,534
-$120
-$285
-$450
$782
$1,749
$1,560
$1,372
-$65
-$154
-$244
$222
$164
$106
$808
$722
$636
$1,423
$1,270
$1,116
-$53
-$125
-$198
$494
$366
$237
$1,801
$1,610
$1,418
$3,173
$2,830
$2,488
-$117
-$280
-$442
$759
$1,697
$1,514
$1,331
-$63
-$150
-$236
$221
$164
$106
$806
$720
$635
$1,420
$1,266
$1,113
-$53
-$125
-$198
$486
$359
$232
$1,769
$1,581
$1,393
$3,116
$2,780
$2,444
-$115
-$275
-$434
$736
$1,646
$1,468
$1,291
-$61
-$145
-$229
$221
$163
$106
$804
$718
$633
$1,416
$1,263
$1,110
-$52
-$125
-$197
$477
$353
$228
$1,738
$1,553
$1,369
$3,062
$2,731
$2,401
-$113
-$270
-$426
$714
$1,596
$1,424
$1,252
-$59
-$141
-$222
$220
$163
$105
$802
$716
$631
$1,412
$1,260
$1,107
-$52
-$124
-$197
$469
$347
$224
$1,708
$1,526
$1,345
$3,009
$2,684
$2,359
-$111
-$265
-$419
$699
$1,565
$1,396
$1,227
-$58
-$138
-$218
$220
$162
$105
$800
$715
$630
$1,410
$1,258
$1,106
-$52
-$124
-$196
$464
$343
$222
$1,688
$1,509
$1,330
$2,974
$2,654
$2,333
-$110
-$262
-$414
$685
$1,533
$1,368
$1,202
-$57
-$135
-$213
$219
$162
$105
$799
$714
$629
$1,408
$1,256
$1,104
-$52
-$124
-$196
$458
$339
$219
$1,669
$1,492
$1,315
$2,941
$2,624
$2,306
-$109
-$259
-$409
$672
$1,503
$1,340
$1,178
-$56
-$132
-$209
$141
$104
$68
$514
$460
$405
$906
$808
$710
-$34
-$80
-$126
$375
$277
$180
$1,367
$1,222
$1,077
$2,408
$2,149
$1,889
-$89
-$212
-$335
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-113 Costs for EV100 Non-Battery Components for the 2010 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
EPA
Vehicle
class
Small car
Small car
Small car
Std car
Std car
Std car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Std car
Std car
Std car
Large car
Applied
WR
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
Net
WR
4%
9%
14%
3%
8%
13%
4%
9%
14%
3%
8%
13%
4%
9%
14%
3%
8%
13%
4%
2017
$344
$266
$187
$1,184
$1,067
$949
$2,048
$1,841
$1,633
$71
-$30
-$132
$265
$204
$144
$912
$822
$731
$1,577
2018
$333
$258
$182
$1,149
$1,035
$921
$1,987
$1,786
$1,584
$69
-$29
-$128
$264
$204
$144
$909
$819
$729
$1,573
2019
$323
$250
$176
$1,114
$1,004
$893
$1,927
$1,732
$1,537
$67
-$29
-$124
$263
$203
$143
$907
$817
$727
$1,568
2020
$314
$242
$171
$1,081
$974
$866
$1,869
$1,680
$1,491
$65
-$28
-$120
$262
$203
$143
$904
$815
$725
$1,564
2021
$304
$235
$166
$1,049
$945
$841
$1,813
$1,630
$1,446
$63
-$27
-$117
$262
$202
$143
$902
$813
$723
$1,560
2022
$295
$228
$161
$1,017
$916
$815
$1,759
$1,581
$1,403
$61
-$26
-$113
$261
$202
$142
$900
$810
$721
$1,556
2023
$289
$223
$158
$997
$898
$799
$1,724
$1,549
$1,375
$60
-$26
-$111
$261
$201
$142
$898
$809
$720
$1,553
2024
$284
$219
$154
$977
$880
$783
$1,689
$1,518
$1,347
$59
-$25
-$109
$260
$201
$142
$897
$808
$719
$1,551
2025
$278
$215
$151
$957
$862
$767
$1,656
$1,488
$1,320
$58
-$25
-$107
$167
$129
$91
$577
$520
$463
$998
3-201
-------
Technologies Considered in the Agencies' Analysis
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
TC
Large car
Large car
Small MPV
Small MPV
Small MPV
Small car
Small car
Small car
Std car
Std car
Std car
Large car
Large car
Large car
Small MPV
Small MPV
Small MPV
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
10%
15%
20%
9%
14%
3%
8%
13%
4%
9%
14%
3%
8%
13%
4%
9%
14%
3%
8%
13%
$1,418
$1,258
$55
-$23
-$102
$608
$470
$331
$2,096
$1,888
$1,680
$3,626
$3,258
$2,891
$126
-$54
-$234
$1,413
$1,254
$55
-$23
-$101
$597
$461
$325
$2,058
$1,854
$1,650
$3,560
$3,199
$2,839
$124
-$53
-$229
$1,410
$1,251
$55
-$23
-$101
$587
$453
$320
$2,021
$1,821
$1,620
$3,496
$3,142
$2,788
$122
-$52
-$225
$1,406
$1,247
$54
-$23
-$101
$576
$445
$314
$1,985
$1,788
$1,591
$3,434
$3,086
$2,738
$119
-$51
-$221
$1,402
$1,244
$54
-$23
-$100
$566
$437
$308
$1,951
$1,757
$1,564
$3,373
$3,032
$2,690
$117
-$50
-$217
$1,398
$1,241
$54
-$23
-$100
$556
$430
$303
$1,917
$1,727
$1,536
$3,315
$2,979
$2,644
$115
-$49
-$214
$1,396
$1,239
$54
-$23
-$100
$550
$425
$300
$1,895
$1,707
$1,519
$3,277
$2,945
$2,613
$114
-$49
-$211
$1,394
$1,237
$54
-$23
-$100
$544
$420
$296
$1,874
$1,688
$1,502
$3,240
$2,912
$2,584
$113
-$48
-$209
$897
$796
$35
-$15
-$64
$445
$344
$243
$1,534
$1,382
$1,230
$2,654
$2,385
$2,116
$92
-$39
-$171
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-114 Costs for EV150 Non-Battery Components for the 2008 Baseline (2010$)
Cost Vehicle
type class
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
Applied
WR
Small car
Std car
Large car
Small MPV
Small car
Std car
Large car
Small MPV
Small car
Std car
Large car
Small MPV
Net
WR
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
2017
2%
2%
3%
1%
2%
2%
3%
1%
2%
2%
3%
1%
2018
$321
$1,101
$1,941
-$30
$247
$848
$1,494
-$23
$569
$1,949
$3,435
-$54
2019
$312
$1,068
$1,882
-$29
$247
$845
$1,490
-$23
$558
$1,913
$3,373
-$53
2020
$302
$1,036
$1,826
-$29
$246
$843
$1,486
-$23
$548
$1,879
$3,312
-$52
2021
$293
$1,005
$1,771
-$28
$245
$841
$1,482
-$23
$538
$1,846
$3,253
-$51
2022 2023
$284
$975
$1,718
-$27
$245
$838
$1,478
-$23
$529
$1,813
$3,196
-$50
$276
$945
$1,667
-$26
$244
$836
$1,474
-$23
$520
$1,782
$3,141
-$49
2024 2025
$270
$927
$1,633
-$26
$244
$835
$1,472
-$23
$514
$1,761
$3,105
-$49
$265
$908
$1,601
-$25
$243
$834
$1,469
-$23
$508
$1,742
$3,070
-$48
$260
$890
$1,569
-$25
$157
$536
$946
-$15
$416
$1,426
$2,514
-$39
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 3-115 Costs for EV150 Non-Battery Components for the 2010 Baseline (2010$)
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
Vehicle
class
Small car
Std car
Large car
Small MPV
Small car
Std car
Large car
Small MPV
Small car
Std car
Large car
Small MPV
Applied
WR
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
20%
Net
WR
1%
1%
3%
1%
1%
1%
3%
1%
1%
1%
3%
1%
2017
$391
$1,231
$2,090
$112
$301
$948
$1,609
$86
$692
$2,180
$3,699
$198
2018
$379
$1,194
$2,027
$108
$300
$945
$1,605
$86
$679
$2,140
$3,632
$194
2019
$368
$1,159
$1,966
$105
$299
$943
$1,600
$86
$667
$2,101
$3,566
$191
2020
$357
$1,124
$1,907
$102
$298
$940
$1,596
$85
$655
$2,064
$3,503
$187
2021
$346
$1,090
$1,850
$99
$298
$938
$1,592
$85
$643
$2,028
$3,442
$184
2022
$336
$1,057
$1,795
$96
$297
$935
$1,587
$85
$632
$1,993
$3,382
$181
2023
$329
$1,036
$1,759
$94
$296
$934
$1,585
$85
$625
$1,970
$3,344
$179
2024
$322
$1,016
$1,724
$92
$296
$932
$1,582
$85
$618
$1,948
$3,306
$177
2025
$316
$995
$1,689
$90
$190
$600
$1,018
$54
$506
$1,595
$2,707
$145
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
For Mild HEV non-battery components, the agencies have used a combination of cost
sources which include the FEV teardown of a Saturn Vue along with estimates used for P2
3-202
-------
Technologies Considered in the Agencies' Analysis
HEVs as described above. For the electrical power distribution and control system and the
DC-DC converter, estimates presented in the NPRM for subcompacts were used with a
presumed 20% weight reduction because those systems were estimated to include a 16 kW
motor (essentially the same as the 15 kW motor assumed for the Mild HEV technology).
These costs and the FEV Saturn Vue teardown costs we used are shown in Table 3-116.
Table 3-116 FEV Teardown Results & P2 HEV Values used for MHEV Non-Battery Direct
Manufacturing Cost Estimates
System
Cooling subsystem
(including water pumps)
Accessory drive
subsystem
Body system
Brake system
Climate control system
Transmission oil pump
and filter subsystem
Generator/alternator and
regulatory subsystem
Electrical power
distribution & control system
DC-DC converter
Total
Teardown result
(2007$)
$88.71
$30.75
$14.83
$42.30
$0
$53.86
$51.94
P2HEV
(2009$)a
$203.22
$115.33
2010$
$92.37
$32.02
$15.44
$44.05
$0
$56.09
$54.09
$205.25
$116.48
$615.79
a See the draft Joint TSD, Table 3-80, 20% WR (EPA-420-D-1 1-901,
November 2011).
For Mild HEV non-battery components, the direct manufacturing costs shown in
Table 3-116 are considered applicable MY 2012. The agencies consider the Mild FIEV non-
battery component technologies to be on the flat portion of the learning curve during the
2017-2025 timeframe. The agencies have applied a medium complexity ICM of 1.39 through
2018 then 1.29 thereafter. The resultant costs used in this final analysis are shown in Table
3-117.
Table 3-117 Costs for Mild HEV Non-Battery Components for both the 2008 and 2010 Baselines (2010$)
Cost type
DMC
1C
TC
Vehicle class
All
All
All
2017
$534
$235
$769
2018
$524
$234
$758
2019
$513
$175
$688
2020
$503
$175
$678
2021
$493
$175
$667
2022
$483
$174
$657
2023
$473
$174
$647
2024
$464
$174
$637
2025
$455
$173
$628
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3.4.4 Hardware costs for charging grid-connected vehicles
Grid-connected vehicles such as EVs and PFLEVs require a means to charge their on-
board batteries to enable their electric range capabilities. These vehicles require certain
hardware to charge, both on-vehicle and off-vehicle. The agencies' September 2010 Technical
3-203
-------
Technologies Considered in the Agencies' Analysis
Assessment Report contains an in-depth analysis of the topic of charging and infrastructure.
The TAR analysis and assumptions did not receive any significant comment, and a review of
the current state of the industry indicates the assumptions in the TAR are still valid.
Therefore, the assumptions for the cost of Electric Vehicle Support Equipment (EVSE) are
unchanged. Additionally, while some of the characteristics of the modeled grid-connected
vehicles such as battery size and energy demand have changed somewhat due to further
analysis, the application of Level 1 and Level 2 charging by vehicle type based on charge time
has not changed.
Three charging levels are currently under consideration. Level 1 charging uses a
standard 120 volt (V), 15-20 amps (A) rated (12-16 A usable) circuit and is available in
standard residential and commercial buildings. Level 2 charging uses a single phase, 240 V,
20-80 A circuit and allows much shorter charge times. Level 3 charging—sometimes
colloquially called "quick" or "fast" charging—uses a 480 V, three-phase circuit, available in
mainly industrial areas, typically providing 60-150 kW of off-board charging power. It is
expected that 97 to 99% of charging will take place at home, so a cost for a home charger,
appropriate to the duty cycle of the vehicle, is added to the vehicle cost. Level 3 charging is
available to commercial users and vehicles that charge at Level 3 stations will be assumed to
pay at the charge station for the convenience of fast charging. Therefore Level 3 charger
costs are not included in overall vehicle cost.
The specific equipment required for charging a grid-connected vehicle consists of the
following:
Charger: A charger that converts electricity from alternating current (AC) from the
electricity source to direct current (DC) required for the battery, and also converts the
incoming 120 or 240 volt current to 300 or higher volts. Grid-connected vehicles carry an on-
board charger capable of accepting AC current from a wall plug (Level 1 circuit) or, from a
Level 2 charging station. On-board charger power capability ranges from 1.4 to 10 kW and is
usually proportional to the vehicle's battery capacity. The lowest charging power, 1.4 kW, is
expected only when grid-connected vehicles are connected to 120 volt (Level 1) outlets, and
all currently known PHEV and EV on-board chargers are expected to provide at least 3.3 kW
charging when connected to a Level 2 (220 volt, 20+ A) charging station. The latest SAE
connection recommended practice, J1772, allows for delivery of up to -19 kW to an on-board
vehicle charger. For higher capacity charging under Level 3, a charging station that delivers
DC current directly to the vehicle's battery is incorporated off-board in the wall or pedestal
mounted.
Charging Station: The charging station needed to safely deliver energy from the
electric circuit to the vehicle, called electric vehicle support equipment (EVSE). The EVSE
may at a minimum, be a specialized cordset that connects a household Level 1/120V socket to
the vehicle; otherwise, the EVSE will include a cordset and a charging station (a wall or
pedestal mounted box incorporating a charger and other equipment). Charging stations may
include optional advanced features such as timers to delay charging until off-peak hours,
communications equipment to allow the utility to regulate charging, or even electricity
metering capabilities. Stakeholders are working on which features are best located on the
EVSE or on the vehicle itself, and it is possible that redundant capabilities and features may
3-204
-------
Technologies Considered in the Agencies' Analysis
be present in both the vehicle and EVSEs in the near future until these issues are worked out.
EVSE and vehicle manufacturers are also working to ensure that current SAE-compliant
"basic" EVSEs are charge-compatible with future grid-connected vehicles.
Dedicated Circuit: A Level 1 circuit is standard household current, 120V AC, rated at
15 or 20 A (12 or 16 A usable). A Level 2 circuit is rated at 208 to 240V and up to 80 A and
is similar to the type of circuit that powers electric stoves (up to 50 A) and dryers (usually 30
A). Generally, Level 1 and 2 circuits used for electric vehicle recharging must be dedicated
circuits, i.e., there cannot be other appliances on that circuit. For a Level 2 circuit, the
homeowner or other user must install a charging station and will need a permit. A homeowner
may choose to install the charger on a separately-metered circuit to take advantage of special
electrical rates for off-peak charging, where available.
In addition to the costs of purchasing and installing charging equipment, charging
station installation may include the costs of upgrading existing electrical panels and installing
the electrical connection from the panel to the desired station location. These costs may be
dramatically lowered if new construction incorporates the panel box and wiring required for
charging stations, or even includes charging stations or outlets for charging stations as
standard equipment.
The current costs of charging stations are highly variable depending on the level of
service (and alternative power capabilities within these categories), location (individual
residence, grouped residences, retail or business, parking lot or garage), level of sophistication
of the station, and installation requirements, including electrical upgrading requirements.
Estimated costs for charging stations are included in Table 3-118 below.
Table 3-118: Estimated Costs for Charging Stations Used in the 2010 TAR (2008$)
Level
Location
Equipment
Installation
1
Single
Residence
$30- $200 (charge cord only, included
at no cost to consumer with EV/PHEV)
when an accessible household plug
(e.g., in a garage or adjacent to a
driveway) with a ground fault
interrupter is already available
$400-$ 1000+ may be necessary
depending on difficulty of installing a
new circuit at the desired location, but
in most cases, owners with sufficient
panel capacity would opt for a more
capable 220 VAC Level 2 installation
instead of a Level 1 dedicated circuit
because the additional installation cost
is only marginally higher
3-205
-------
Technologies Considered in the Agencies' Analysis
Residential,
Apartment
Complex, or
Fleet Depotb
3.3 kW EVSE (each): $300- $4,000
6.6 kW EVSE (each): $400- $4,000
3.3- 6.6 kW installation cost:
$400-$2,300 without wiring/service
panel upgrade, or
$2,000-$5,000 with panel upgrade
rets: 77,78,79,80,3
a Detailed information on charger cost for each charging level and location and specific sources for cost
estimates are available in the TAR, Appendix G.
b Level 2 EVSE installation costs vary considerably for single-family residences, multi-family residences, and
fleet depots, depending upon the need for wiring and service panel upgrades. The range depicted here reflects
the anticipated variability of these costs. However, EPRI estimates that the typical residential Level 2
installation costs to be approximately $1,500. See the TAR, Appendix G for additional information.
3.4.4.1 Application of charging level by vehicle type
The home charging availability for a specific consumer will need to be differentiated
among EV/PHEVs with different battery capacity. The electric outlets in existing homes are
most likely ready for Level 1 charging, which is about sufficient for fully recharging a
PHEV20 SUV during normal nighttime, provided the outlet is not being heavily utilized by
other loads. Shorter available charging time or owning a PHEV or an EV with a larger
battery make the capability to fully charge overnight with a Level 1 system less likely, but
upgrading to a Level 2 system in such cases will allow full recharge to happen more quickly.
Table 3-119 shows the application of charge level by vehicle type and range.
Charging types were chosen based on nominal time to charge a fully-depleted battery in a
vehicle with 0% net weight reduction. Charge times exceeding 9 hours for Level 1 were
deemed unacceptable and Level 2 charging was specified. For charge times between 6 hours
and 9 hours on Level 1, a mix of Level 1 and Level 2 was specified. This was done to
recognize the varying consumer value of faster, but more expensive, Level 2 charging over
Level 1 charging.
Table 3-119: Charger Type by Vehicle Technology and Class
EPA Vehicle
Class
Small car
Standard Car
Large Car
Small MPV
Large MPV
Truck
PHEV20
100% LI
100% LI
100% LI
100% LI
100% LI
50% LI
50% L2
PHEV40
25% LI
75% L2
10% LI
90% L2
100% L2
100% L2
100% L2
100% L2
EV75
100% L2
100% L2
100% L2
100% L2
100% L2
100% L2
EV100
100% L2
100% L2
100% L2
100% L2
100% L2
100% L2
EV150
100% L2
100% L2
100% L2
100% L2
100% L2
100% L2
3-206
-------
Technologies Considered in the Agencies' Analysis
For this final rule, consistent with the proposal, the resultant costs associated with in-
home chargers and installation of in-home chargers are included in the total cost for an EV
and or PHEV. However, here we summarize specially the costs for chargers and installation
labor. The agencies have estimated the DMC of a level 1 charge cord at $31 (2010$) based
on typical costs of similar electrical equipment sold to consumers today and that for a level 2
charger at $204 (2010$). Labor associated with installing either of these chargers is estimated
at $1,020 (2010$). Further, we have estimated that all PHEV20 vehicles (PHEVs with a 20
mile range) would be charged via a level 1 charger and that all EVs, regardless of range,
would be charged via a level 2 charger. For the PFLEV40 vehicles (PFLEVs with a 40 mile
range), we have estimated that: 25% of small cars would be charged with a level 1 charger
with the remainder charged via a level 2 charger; 10% of standard cars would be charged with
a level 1 charger with the remainder charged via a level 2 charger; and all remaining PFLEV
40 vehicles would be charged via a level 2 charger. All costs presented here are considered
applicable in the 2025 model year. The agencies have applied the learning curve presented in
Section 3.2.3 to all charger costs. The agencies have also applied a Highl ICM of 1.56
through 2024 then 1.34 thereafter. Installation costs, being labor costs, have no learning
impacts or ICMs applied. The resultant costs are shown in Table 3-120.
Table 3-120 Costs for EV/PHEV In-home Chargers (2010$)
Cost
type
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
TC
Technology
PHEV20
Charger
PHEV40
Charger
EV Charger
PHEV20
Charger
PHEV40
Charger
EV Charger
PHEV20
Charger
PHEV40
Charger
EV Charger
Charger
labor
EPA
Vehicle
Class
All
Small car
Std car
Large car
Small
MPV
All
All
Small car
Std car
Large car
Small
MPV
All
All
Small car
Std car
Large car
Small
MPV
All
All
2017
$60
$314
$365
$398
$398
$19
$100
$117
$128
$128
$79
$414
$481
$526
$526
$1,020
2018
$48
$251
$292
$319
$319
$18
$96
$112
$122
$122
$66
$347
$404
$441
$441
$1,020
2019
$48
$251
$292
$319
$319
$18
$96
$112
$122
$122
$66
$347
$404
$441
$441
$1,020
2020
$38
$201
$233
$255
$255
$18
$93
$108
$118
$118
$56
$294
$342
$373
$373
$1,020
2021
$38
$201
$233
$255
$255
$18
$93
$108
$118
$118
$56
$294
$342
$373
$373
$1,020
2022
$38
$201
$233
$255
$255
$18
$93
$108
$118
$118
$56
$294
$342
$373
$373
$1,020
2023
$38
$201
$233
$255
$255
$18
$93
$108
$118
$118
$56
$294
$342
$373
$373
$1,020
2024
$38
$201
$233
$255
$255
$18
$93
$108
$118
$118
$56
$294
$342
$373
$373
$1,020
2025
$31
$161
$187
$204
$204
$11
$55
$64
$70
$70
$41
$216
$251
$274
$274
$1,020
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3-207
-------
Technologies Considered in the Agencies' Analysis
3.4.5 Other Technologies Assessed that Reduce CO2 and Improve Fuel Economy
In addition to the technologies already mentioned above, the other technologies
generally considered in the agencies' analysis are described below. They fall into five broad
categories: engine technologies, transmission technologies, vehicle technologies,
electrification/accessory technologies, hybrid technologies and mass reduction
3.4.5.1 Lower Rolling Resistance Tires
Tire rolling resistance is the frictional loss associated mainly with the energy
dissipated in the deformation of the tires under load and thus influences fuel economy and
CC>2 emissions. Other tire design characteristics (e.g., materials, construction, and tread
design) influence durability, traction (both wet and dry grip), vehicle handling, and ride
comfort in addition to rolling resistance. A typical low rolling resistance tire's attributes
could include: increased specified tire inflation pressure, material changes, and tire
construction with less hysteresis, geometry changes (e.g., reduced aspect ratios), and
reduction in sidewall and tread deflection. These changes would generally be accompanied
with additional changes to vehicle suspension tuning and/or suspension design.
The agencies expect that greater reductions in tire rolling resistance will be possible
during the rulemaking timeframe than are currently available, as tire manufacturers continue
to improve their products in order to meet increasing demand by auto OEMs for tires that
contribute more to their vehicles' fuel efficiency. Thus, for this final rule, consistent with the
proposal, the agencies considered two "levels" of lower rolling resistance tires. The first level
("LRR1") is defined as a 10 percent reduction in rolling resistance from a base tire, which was
estimated to be a 1 to 2 percent effectiveness improvement in MYs 2012-2016 final rule.
Based on the 2011 Ricardo study the agencies are now using 1.9 percent effectiveness
improvement for LRRlfor all vehicle classes. LRR1 tires are widely available today, and
appear to comprise a larger and larger portion of tire manufacturers' product lines as the
technology continues to improve and mature. The second level ("LRR2") is defined as a 20
percent reduction in rolling resistance from a base tire, yielding an estimated 3.9 percent
effectiveness improvement. In the CAFE model this results in a 2.0 percent incremental
effectiveness increase from LRR1. LRR2 represents an additional level of rolling resistance
improvement beyond what the agencies considered in the MYs 2012-2016 rulemaking
analysis. NHTSA assumed that the increased traction requirements for braking and handling
for performance vehicles could not be fully met with the ROLL2 designs in the MYs 2017-
2025 timeframe. For this reason the CAFE model did not apply ROLL2 to performance
vehicle classifications. However, the agency did assume that tractions requirement for
ROLL1 could be met in this timeframe and thus allowed ROLL1 to be applied to performance
vehicle classifications in the MYs 2017-2025 timeframe.
In the 2012-2016 light duty vehicle rule, the agencies estimated the incremental DMC
at an increase of $5 (2007$) per vehicle. This included costs associated with five tires per
3-208
-------
Technologies Considered in the Agencies' Analysis
vehicle, four primary and one spare tire. There is no learning applied to this technology due to
the commodity based nature of this technology. Looking forward from 2016, the agencies
continue to apply this same estimated DMC adjusted for 2010 dollars.bbb The agencies
consider LRR1 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.
To analyze the feasibility and cost for a second level of rolling resistance
improvement, EPA, NHTSA, and CARB met with a number of the largest tire suppliers in the
United States. 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 today's tires by MY 2017. As such, the agencies agreed, based on these
discussions, to consider LRR2 as initially available for purposes of this analysis in MY 2017,
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, and did not allow complete application (100 percent of a
manufacturer's fleet) until 2023. The agencies believe that this schedule aligns with the
necessary efforts for production implementation, such as system and electronic system
calibration and verification.
LRR2 technology does not yet exist in the marketplace today, making cost estimation
challenging without disclosing potentially confidential business information. To develop a
transparent cost estimate, the agencies relied on LRR1 history, costs, market implementation,
and information provided by the 2010 NAS report. The agencies assumed low rolling
resistance technology ("LRR1") 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 MY 2017 as the starting point for market
entry for LRR2 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
LRR2 at $10 (2010$) per tire, or $40 ($2010) per vehicle. This estimate is generally fairly
consistent with CBI suggestions by tire suppliers. The agencies have not included a cost for
the spare tire because we believe manufacturers are not likely to include a LRR2 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.81 The agencies consider this estimated cost for
LRR2 to be applicable in MY 2021. Further, the agencies consider LRR2 technology to be on
the steep portion of the learning curve where costs would be reduced quickly in a relative
bbb As noted elsewhere in this chapter, 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 $5 for presentation simplicity.
3-209
-------
Technologies Considered in the Agencies' Analysis
short period of time. The agencies have applied a low complexity ICM of 1.24 through 2024,
and then 1.19 thereafter. The ICM timing for LRR2 is different from that for LRR1 because
LRR2 is brand-new for this rulemaking and is not yet being implemented in the fleet. The
resultant costs are shown in Table 3-121. Note that both LRR1 and LRR2 are incremental to
the baseline system, so LRR2 is not incremental to LRR1.
Table 3-121 Costs for Lower Rolling Resistance Tires Levels 1 & 2 (2010$)
Cost type
DMC
DMC
1C
1C
TC
TC
Lower
Rolling
Resistance
Tire
Technology
Level 1
Level 2
Level 1
Level 2
Level 1
Level 2
2017
$5
$63
$1
$10
$7
$73
2018
$5
$63
$1
$10
$7
$73
2019
$5
$51
$1
$10
$6
$60
2020
$5
$51
$1
$10
$6
$60
2021
$5
$40
$1
$10
$6
$50
2022
$5
$39
$1
$10
$6
$49
2023
$5
$38
$1
$10
$6
$48
2024
$5
$37
$1
$10
$6
$47
2025
$5
$36
$1
$8
$6
$44
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Note that both levels of lower rolling resistance tires are incremental to today's baseline tires.
Given that the final standards cover such a long timeframe, the agencies also
considered introducing a third level of rolling resistance reduction ("LRR3"), defined as a 30
percent reduction in rolling resistance. The agencies evaluated the potential of LRR3 entering
the marketplace during this final rulemaking timeframe.
Tire technologies that enable improvements of 10 and 20 percent have been in
existence for many years. Achieving improvements up to 20 percent involves optimizing and
integrating multiple technologies, with a primary contributor being the adoption of a silica
tread technology.82 This approach was based on the use of a new silica along with a specific
polymer and coupling agent combination. The use of the polymer, coupling agent and silica
was known to reduce tire rolling resistance at the expense of tread wear, but new approach
using novel silica reduced the tread wear tradeoff. Tire suppliers have indicated there are one
or more innovations/inventions that they expect to occur in order to move the industry to the
next quantum reduction of rolling resistance. However, based on the historical development
and integration of tire technologies, there appears to be little evidence supporting
improvements beyond LRR2 by 2025. Therefore, the agencies decided not to incorporate
LRR3 at this time.
The agencies sought comment on whether we should consider application of a 30
percent reduction from today's rolling resistance levels being available for mass production
implementation by MY 2025 or sooner. The agencies also sought comment on the viability of
this technology, maturity by MY 2025, as well as market introduction timing and the
technological ways that this level of rolling resistance improvement will be achieved without
any tradeoffs in terms of vehicle handling capability and tire life from what consumers expect
today. Finally, the agencies sought cost information regarding the potential incorporation of
LRR3 relative to today's costs as well as during the timeframe covered by this final rule. No
comments were submitted on any of these topics.
3-210
-------
Technologies Considered in the Agencies' Analysis
3.4.5.2 Low Drag Brakes
Low 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 rotating disc either
by mechanical or electric methods
The 2012-2016 final rule and 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 and updated lumped-parameter model.
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.
The resultant costs are shown in Table 3-122.
Table 3-122 Costs for Low Drag Brakes (2010$)
Cost type
DMC
1C
TC
2017
$59
$14
$74
2018
$59
$14
$74
2019
$59
$11
$71
2020
$59
$11
$71
2021
$59
$11
$71
2022
$59
$11
$71
2023
$59
$11
$71
2024
$59
$11
$71
2025
$59
$11
$71
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3.4.5.3 Front or Secondary Axle Disconnect for Four-Wheel Drive Systems
Energy is required to continually drive the front, or secondary, axle in a four-wheel
drive system even when the system is not required during most operating conditions. This
energy loss directly results in increased fuel consumption and CC>2 emissions. Many part-time
four-wheel drive systems use some type of front axle disconnect to provide shift-on-the-fly
capabilities. The front axle disconnect is normally part of the front differential assembly. As
part of a shift-on-the-fly four-wheel drive system, the front axle disconnect serves two basic
purposes. First, in two-wheel drive mode, it disengages the front axle from the front driveline
so the front wheels do not turn the front driveline at road speed, saving wear and tear.
Second, when shifting from two- to four-wheel drive "on the fly" (while moving), the front
axle disconnect couples the front axle to the front differential side gear only when the transfer
case's synchronizing mechanism has spun the front driveshaft up to the same speed as the rear
driveshaft. Four-wheel drive systems that have a front axle disconnect typically do not have
either manual- or automatic-locking hubs. To isolate the front wheels from the rest of the
front driveline, front axle disconnects use a sliding sleeve to connect or disconnect an axle
shaft from the front differential side gear. NHTSA and EPA are not aware of any
manufacturer offering this technology in the U.S. today on unibody frame vehicles; however,
it is possible this technology could be introduced by manufacturers within the MYs 2017-
2025 time period.
3-211
-------
Technologies Considered in the Agencies' Analysis
The 2012-2016 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.
In the 2012-2016 rule, the agencies estimated the DMC at $78 (2007$) which was
considered applicable to the 2015MY. This DMC becomes $82 (2010$) for this analysis after
adjusting to 2010 dollars. The agencies consider secondary axle disconnect 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. The resultant costs are shown in Table 3-123.
Table 3-123 Costs for Secondary Axle Disconnect (2010$)
Cost type
DMC
1C
TC
2017
$78
$20
$98
2018
$76
$20
$96
2019
$75
$16
$91
2020
$73
$16
$89
2021
$72
$16
$88
2022
$70
$16
$86
2023
$69
$16
$85
2024
$68
$16
$83
2025
$66
$16
$82
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
3.4.5.4 Aerodynamic Drag Reduction
Many factors affect a vehicle's aerodynamic drag and the resulting power required to
move it through the air. The overall drag force can be simplified as proportional to vehicle's
frontal area, vehicle's drag coefficient, air density and the second order of vehicle's velocity.
Therefore reducing vehicle's frontal area and drag coefficient can reduce fuel consumption
and CC>2 emissions. Although frontal areas tend to be relatively similar within a vehicle class
(mostly due to market-competitive size requirements), significant variations in drag
coefficient can be observed. Significant changes to a vehicle's aerodynamic performance
may need to be implemented during a redesign (e.g., changes in vehicle shape). However,
shorter-term aerodynamic reductions, with a somewhat lower effectiveness, may be achieved
through the use of revised exterior components (typically at a model refresh in mid-cycle) and
add-on devices that are currently being applied. The latter list would include revised front
and rear fascias, modified front air dams and rear valances, addition of rear deck lips and
underbody panels, and lower aerodynamic drag exterior mirrors.
The 2012-2016 final rule estimated that a fleet average of 10 to 20 percent total
aerodynamic drag reduction is attainable which equates to incremental reductions in fuel
consumption and CC>2 emissions of 2 to 3 percent for both cars and trucks. These numbers
are generally supported by the Ricardo study and public technical literature and therefore
NHTSA and EPA are retaining these estimates, as confirmed by joint review, for the purposes
of this final rule, consistent with the proposal. Importantly, the effectiveness values presented
here represent two-cycle effectiveness. Because active aerodynamic technologies (i.e., aero
level 2) provide additional off-cycle benefits, both agencies apply an off-cycle credit value to
the technology. Off-cycle credits are discussed in Chapter 5 of this Joint TSD.
3-212
-------
Technologies Considered in the Agencies' Analysis
For this final rule, consistent with the proposal, the agencies considered two levels of
aero improvements. The first level is that discussed in the 2012-2016 final rule and the 2010
TAR and includes such body features as air dams, tire spats, and perhaps one underbody
panel. In the 2012-2016 rule, the agencies estimated the DMC of aero-level 1 at $39 (2007$).
This DMC becomes $41 (2010$) for this analysis, applicable in the 2015MY. The agencies
consider aero-level 1 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.
The second level of aero—level 2 which includes such body features as active grille
shutters000, rear visors, larger under body panels or low-profile roof racks —was discussed in
the 2010 TAR where the agencies estimated the DMC at $120 (2008$) incremental to the
baseline vehicle. The agencies inadvertently used that cost as inclusive of aero-level 1
technologies when it should have been incremental to aero-1 technologies. As a result, the
agencies now consider the TAR cost to more appropriately be incremental to aero-level 1 with
a DMC for this analysis of $123 (2010$). The agencies consider this cost to be applicable in
the 2015MY. Further, the agencies consider aero-level 2 technology to be on the flat portion
of the learning curve. The agencies have applied a medium complexity ICM of 1.39 through
2024 then 1.29 thereafter. The timing of the aero-level 2 ICMs is different than that for the
level 1 technology because the level 2 technology is newer and not yet being implemented in
the fleet. The resultant costs are shown in Table 3-124.
Table 3-124 Costs for Aerodynamic Drag Improvements - Levels 1 & 2 (2010$)
Cost
type
DMC
DMC
1C
1C
TC
TC
TC
Aero
Technology
Level 1
Level 2
Level 1
Level 2
Level 1
Level 2
Level 2
Incremental
to
Baseline
Aero-level 1
Baseline
Aero-level 1
Baseline
Aero-level 1
Baseline
2017
$39
$117
$10
$47
$49
$164
$213
2018
$38
$115
$10
$47
$48
$162
$210
2019
$37
$112
$8
$47
$45
$160
$205
2020
$37
$110
$8
$47
$45
$157
$202
2021
$36
$108
$8
$47
$44
$155
$199
2022
$35
$106
$8
$47
$43
$153
$196
2023
$35
$104
$8
$47
$42
$150
$193
2024
$34
$102
$8
$47
$42
$148
$190
2025
$33
$100
$8
$35
$41
$135
$176
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Because a large percent of the performance vehicles already have some level of
aerodynamic treatments, when running the CAFE model NHTSA only applies level 1 of
aerodynamic treatment to these vehicles. Also for specific vehicles, such as Toyota Prius,
which already have extensive aerodynamic treatment, the level of the aerodynamic that could
be further applied by NHTSA in the CAFE model is limited in the market input file.
! For details on how active aerodynamics are considered for off-cycle credits, see TSD Chapter 5.2.2.
3-213
-------
Technologies Considered in the Agencies' Analysis
3.4.5.5 Mass Reduction
From 1987-2011, there has been a generally increasing trend in the weight of the light
duty vehicle fleet as shown in Figure 3-26 from EPA's Fuel Economy Trends Report83. A
number of factors have contributed to this weight increase, including the choices of
manufacturers and consumers to build and purchase larger vehicles, including heavier trucks,
SUVs, and CUVs. Also contributing to this weight increase has been an increase in vehicle
content including: safety features (air bags, antilock brakes, energy absorbent and intrusion
resistant vehicle structures, etc.), noise reduction (additional damping material), added
comfort and convenience features (air conditioning, power locks and windows), luxury
features (infotainment systems, powered seats), etc.
5000
3000
1975
1980 1985
1990 1995
Model year
2000 2005
2010
Figure 3-26 Light duty fleet weight trends: 1975-2011
Despite this increase in weight, the average acceleration of vehicles has grown steadily
faster without any marked or consistent reduction in fuel economy since 1987, as shown in
Figure 3-27. This combination of increased vehicle performance, stable fuel economy, and
increased vehicle weight has been partially enabled by the development and adoption of more
efficient technologies, especially in engines and transmissions. The impressive improvements
in powertrain efficiency during this period have offset increases in energy consumption that
result from improvements in weight carrying, towing and volume capacities, safety, consumer
features, vehicle refinement, and acceleration performance.
3-214
-------
Technologies Considered in the Agencies' Analysis
Adjusted fuel economy (mpg)
0/1 ic
00 -
on j
1 S -
i
1 0 -
ft -
• -*-•
^. *t^«- •»-*" 7i*1"~*~*%»-«, fuel economy -,,/
***~ ? N^^ ---**....-
/
^^ "*~*~^s*~^ acceleration
'
^
15 **
1 /i pfl
€\
no
«
1975 1980 1985 1990 1995 2000 2005 2010
Model year
Figure 3-27 Light duty fleet trends for acceleration and fuel economy: 1975-2011
84
Motoi
1950-2C
60
40
20
1!
* vehicle crash deaths per billion miles traveled
309
^
Vx^v
^-v.
S50 55 60 65 70 75 80 85 90
************»•»
11. 3 per billion
95 2000 05
Figure 3-28 U.S. Vehicle Fatality Rates for the past 60 years
Vehicle mass reduction (also referred to as "down-weighting" or 'light-weighting"),
reduces the energy needed to overcome inertial forces, thus yielding lower fuel consumption
and GHG emissions. While keeping everything else constant, a lighter vehicle will require
less energy to operate than a heavier vehicle. Mass reduction can be achieved through a
number of approaches described below, even while maintaining vehicle size. Alternatively,
mass reduction can also be achieved by vehicle "downsizing" which involves reducing
vehicle exterior dimensions, such as shifting from a midsize vehicle to a compact vehicle.
Consistent with the proposal, the agencies did not analyze downsizing as a mass reduction
strategy in this analysis for the final rule. In part, this is because a manufacturer's ability to
downsize its vehicles is constrained by consumer preferences (such as for interior passenger
or cargo volume), which are in turn influenced by many factors that are difficult to predict in
3-215
-------
Technologies Considered in the Agencies' Analysis
the future, such as the consumer's utility needs, fuel prices, economic conditions, etc. Also,
the final CAFE and GHG emission standards are based on vehicle footprint (the area bounded
by where the four tires contract the ground), and assign higher fuel economy targets (and
lower CC>2 emission targets) for vehicles with smaller footprints and lower fuel economy
targets (and higher CO2 emission targets) for vehicles with larger footprints. As discussed in
Chapter 2 of the joint TSD, the agencies believe the shape of the footprint-based target curves
will not create incentives for manufacturers to either upsize or downsize their vehicles. Based
on these considerations, the agencies are assuming that manufacturers will favor mass
reduction through material substitution, design optimization, and adopting other advanced
manufacturing technologies rather than compromising a vehicle's attributes and functionality,
such as occupant or cargo space, vehicle safety, comfort, acceleration, etc. Consequently, the
compliance paths the agencies have investigated for the promulgated standards do not include
downsizing.
Mass reduction has an important relationship with vehicle powertrain selection and
sizing. Vehicle powertrain selection depends on an OEM's product strategy, and may include
a variety of options such as naturally aspirated engines, boosted and downsized gasoline
engines, diesel engines, or vehicle electrification (P/H/EV) Regardless of the strategy
selected, vehicle mass reduction for non-powertrain systems is an important enabler to further
reduce vehicle fuel consumption and reduce the size of the powertrain system. The term
"glider" refers to a complete vehicle minus the powertrain. Figure 3-29 illustrates the mass
o/r
breakdown by system for a typical vehicle . The non-powertrain systems normally account
for 75 percent of vehicle weight. The agencies have accounted for some of the costs of
engine mass reduction when applying engine downsizing technologies. The agencies have
also accounted for the amount of mass change due to the application of hybrid and
electrification technologies in the vehicle electrification sections. Therefore, this section
focuses on both the mass reduction of the glider as well as mass reduction technologies that
are specifically targeted at reducing the weight of the powertrain.ddd rather than on mass
reduction resulting from powertrain efficiency improvements. An example of a mass
reduction technology for the powertrain that is not related to powertrain efficiency
improvement is material substitution, such as changing the engine block from cast iron to
aluminum or changing the size of the fuel tank.). Mass reduction is calculated for both the
glider and the vehicle including powertrain in the studies sponsored by the agencies as shown
later in this section.
ddd Rather than on mass reduction resulting from powertrain efficiency improvements, such as in the case of
adding a turbocharger to a downsized engine.
3-216
-------
Technologies Considered in the Agencies' Analysis
Approximate vehicle
mass breakdown"
Misc.;
Closures, ™
fenders;
8K B°dY;
'*• 23-28%
Suspension/chassis;
22-27)4
System
Body-in-white
Powertrain
Chassis
Interior
Closures
Miscellaneous
Major components in system
Passenger compartment frame, cross and side beams, roof
structure, front-end structure, underbody floor structure, panels
Engine, transmission, exhaust system, fuel tank
Chassis, suspension, tires, wheels, steering, brakes
Seats, instrument panel, insulation, trim, airbags
Front and rear doors, hood, lift gate
Electrical, lighting, thermal, windows, glazing
"Based on Stodolsky et al, 1995a; Bjelkengren, 2008; Lotus Engineering, 2010; the actual system definitions and system
component inclusion can vary, and percentage weight breakdown can vary substantially by vehicle
Figure 3-29 Vehicle system mass approximation
A vehicle can be divided into 6 major systems, which are shown in Figure 3-29. Mass
reduction can potentially be applied to any of a vehicle's subsystems, including the engine,
exhaust system, transmission, chassis, suspension, brakes, body, closure panels, glazing, seats
and other interior components, engine cooling systems, and HVAC systems. While
manufacturers may reduce the mass of some individual components during a vehicle refresh,
they generally undertake larger amounts of mass reduction systematically and more broadly
across all vehicle systems when redesigning a vehicle. In the redesign process, OEMs
normally set weight targets by benchmarking other vehicles in the same segment and
projecting weight trends into the future, and then identifying targets for all components and
subsystems that support achieving the target. The agencies believe this holistic approach,
which takes into consideration all secondary mass savings, is likely the most effective way for
OEMs to achieve large amounts of mass reduction.
During a vehicle redesign where mass reduction is a strategic vehicle program goal,
OEMs can consider modular systems design, secondary mass effects, multi-material concepts,
and new manufacturing processes to help optimize the design. There are several studies in the
public domain that illustrate the potential for these approaches to achieve significant amounts
of mass reduction, although it is important to also recognize that the studies use some
assumptions that do not account for some of the considerations that are important to
manufacturers. One example is the need to share some components across platforms to
manage cost and part complexity for assembly and service, which limits the ability to
optimize the amount of mass reduction on every vehicle component. Care must also be taken
in any study to assure that vehicle functionality and performance, such as stiffness, NVH,
safety and vehicle dynamics, continue to meet manufacturer objectives and consumer
demands. It is important for design studies to use tools such as simulation modeling to assess
the design's ability to meet functionality and performance targets. In this rulemaking, the
agencies have targeted to preserve vehicle function and performance in their analysis of mass
reduction.
3-217
-------
Technologies Considered in the Agencies' Analysis
An example of this approach is illustrated in Figure 3-30, which summarizes the
results of the 2010 phase I Lotus Engineering mass reduction study of a Toyota Venza.
Mass-reduction
features, findings
Redesign conventional mid-size vehicle for mass optimization, with two redesign architectures
Low Development vehicle technology with industry-leading manufacturing techniques that were
deemed feasible for 2014 (for model year 2017 production) for assembly at existing facilities
High Development vehicle technology, with modifications to conventional joining and assembly
processes that were deemed feasible for 2017 (for model year 2020) production
Extensive use of material substitution with high-strength steel, advanced high-strength steel,
aluminum, magnesium, plastics and composites throughout vehicles
Conservative use of emerging design and parts integration concepts to minimize technical risk
Using synergistic total vehicle substantial mass reduction opportunities found at minimized piece costs
The Low Development vehicle was found to have likely piece cost reductions, whereas the High
Development vehicle had nominal estimated cost increase of 3% (with potential for cost reduction)
Mass-reduction
impact
• Body structure reduction for Low Development Vehicle: 55 Ib (6.6%)
• Body structure reduction for High Development Vehicle: 356 Ib (42%)
• Overall glider reduction for Low Development Vehicle: 538 Ib (19%)
• Overall glider reduction for High Development Vehicle: 1096 Ib (39%)
• Overall vehicle reduction for Low Development Vehicle (with hybrid powertrain): 657 Ib (17.6%)
» Overall vehicle reduction for High Development Vehicle (with hybrid powertrain): 1209 Ib (32%)
Status
Engineering design study conducted by Lotus Engineering
First phase of project, development of two mass-reduced vehicle designs completed in April 2010
Second phase to test structural integrity, impact load paths, crash worthiness to validate the vehicle
designs.
Source
Lotus Engineering, Inc. 2010. An Assessment of Mass Reduction Opportunities for a 2017-2020
Model Year Vehicle Program
Figure 3-30 Example of a holistic vehicle redesign study from Lotus Engineering8
Mass reduction can be considered in terms of the "percent by which the redesigned
vehicle is lighter than the previous version," recognizing that the value likely represents both
"primary" mass reduction (that which the manufacturer set out to make lighter), and
"secondary" mass reduction (from ancillary systems and components that can now be lighter
due to the primary mass reductions).
oo
As summarized by NAS in its 2011 report, there are two key strategies for primary
mass reduction: 1) changing the design to use less material or 2) substituting lighter materials
for heavier materials. The first key strategy of using less material compared to the baseline
component can be achieved by optimizing the design and structure of the component, system
or vehicle structure. For example, a number of "body on frame" vehicles have been
redesigned with a lighter "unibody" construction, eliminating components, reducing the
weight of the body structure, and resulting in significant reductions in overall mass and
related costs. The unibody design currently dominates the passenger car segment and has
increased penetration into what used to be mostly body-on-frame vehicles, such as SUVs.
5-218
-------
Technologies Considered in the Agencies' Analysis
This technique was used in the 2011 Ford Explorer redesign, which also employed the
extensive use of high strength steels.89 Figure 3-31 depicts body-on-frame and unibody
designs for two sport utility vehicles.
Unibody
Figure 3-31 Illustration of Body-on-Frame (BoF) and Unibody vehicle construction
To further reduce mass inefficiencies in vehicle design, vehicle manufacturers are
using continually-improving Computer Aided Engineering (CAE) tools. For example, the
Future Steel Vehicle (FSV) project90 sponsored by WorldAutoSteel used three levels of
optimization: topology optimization, low fidelity 3G (Geometry Grade and Gauge)
optimization, and sub-system optimization, to achieve 30 percent mass reduction in the body
structure of a vehicle with a mild steel unibody structure (see Figure 3-32). Designs similar
to those proposed in the FSV project have been applied in production vehicles, such as the B-
pillarof2010FordFocus.91
3-219
-------
Technologies Considered in the Agencies' Analysis
2.4 T4: Body Structure Sub-System Optimisation
The fmal design attained from the LF3G optimisation was used as the basts for the sub-system optimisation
as well as the source of the boundary conditions Load path mapping was conducted on the mode) to
identify the most dominant structural sub-systems in the body structure. Load path mapping considers the
dominant loads m the structural sub-systems for each of the load cases as shown m Figure 2-7.
x
figurt 2-7: T4 Load Path Mapping - Major Load Path Components
Based on load path mapping, seven structural sub-systems (Figure 2-8) were selected for further
optimisation using the spectrum of FSVs potential manufacturing technologies
•Shot Gun
Figure 2-9. Structural Sub-Sytttmt Selected
FutureSteelVehicle
^
WorldAutoStecl
Figure 3-32 Example of vehicle body load path mapping for mass optimization
Vehicle manufacturers have long used these continually-improving CAE tools to
optimize vehicle designs. But because any design must meet component and system
3-220
-------
Technologies Considered in the Agencies' Analysis
functionality and manufacturability targets, there are practical limitations to the amount of
additional mass reduction that can be achieved through optimization. For example, an
optimization program would need to account for safety, stiffness, NVH, manufacturing, and
other requirements to assure the design is suitable for its intended function and for mass
production. Additionally, ultimate optimization of vehicle design for mass reduction may be
limited by an OEM's use of shared components and common platform for multiple vehicle
models. While optimization may concentrate on the vehicle that has the largest production
volume for a platform, designs must also support the most demanding functional requirements
of all of the vehicles that share that platform, or those functional requirements will not be met.
In addition, the engineering resources and capital for tooling and equipment that would be
needed to optimize every vehicle component at each redesign affects the ability to fully
optimize a new vehicle to achieve all of the theoretically possible secondary mass reduction.
Therefore, some level of mass inefficiency will inherently exist on many or all of the vehicles
that share a platform. The agencies sought comment and information in the NPRM on the
degree to which shared vehicle components and architectures affect the feasible amount of
mass reduction and the cost for mass reduction relative to what could be achieved if mass
reduction was optimized for a single vehicle design. Volkswagen confirmed in its comments
that with platform sharing, "a weight reduction technology which may be acceptable in terms
of price or performance for one model may disrupt the economics or utility of another."92
Using less material can also be achieved through improving the manufacturing
process, such as by using improved joining technologies and parts consolidation. This
method is often used in combination with applying new materials. For example, more precise
manufacturing techniques such as laser welding may reduce the flange size necessary for
welding, and thus marginally decrease the mass of an assembly. Also, when complex
assemblies are constructed from fewer pieces, the mass of the assembly tends to be lower.
However, while synergies in mass reduction certainly exist, and while certain technologies
can enable one another (e.g., parts consolidation and molding of advanced composites), others
may be incompatible (e.g., laser welding and magnesium casting).
The second key strategy to reduce mass of an assembly or component involves the
substitution of lower density and/or higher strength materials. Table 3-125 shows material
usage typical of contemporary high-volume vehicles. Material substitution includes replacing
materials, such as mild steel, with higher-strength and advanced steels, aluminum,
magnesium, and composite materials. The substitution of advanced high strength steel
(AHSS) for mild steel can reduce the mass of a strength-critical part because the gauge of the
AHSS components can be reduced, despite the fact that the densities of the materials are not
significantly different. Aluminum has also been used over the years in a variety of
components, such as vehicle closures, suspension parts, engine cradles, etc. Aluminum has
one third the density of steel and therefore can provide a notable amount of mass reduction.
Changing parts from steel to aluminum generally requires part redesign, and extra material
may have to be added for strength or durability. Aluminum also has a shorter fatigue life than
steel, and therefore the alloy selected and the application must be carefully considered.
Magnesium can provide additional mass reduction as it has lower density than aluminum. It
has been used for instrument panel cross-car beams by several OEMs for a number of years.
It has also been used in an engine block produced by BMW for several years. Its brittle
3-221
-------
Technologies Considered in the Agencies' Analysis
nature must be considered, however, when selecting the alloy and the application within the
vehicle.
Table 3-125 Distribution of Material in Typical Contemporary Vehicles (e.g., Toyota Camry or Chevrolet
Malibu)93
Approximate Concent in Cars
Material Comments Today, by Weight (percent)
Iron and mild steci
High-strength steel
Aluminum
Plastic
Other (magnesium, tuanRim, rubber, eic.)
Under 480 Mpa 55
> 4HO Mpa (in body structure) 15
No aluminum closure panels; aluminum engine block and head and wheels 10
Miscellaneous parts, mostly interior trim, tight lenses* facia, instrument panel 10
Miscellaneous, parts 10
Automobiles also utilize a wide range of plastic types, including polypropylenes,
polyesters, and vinyl esters. These materials are utilized in hatches, roofs, interior panels,
instrument panels, and hundreds of other parts. Although primarily used in nonstructural
vehicle components, plastics have continued to make in-roads in bumper systems and in
composite beam applications, and some studies have found potential to supplant structural
beams and frame components. Lighter plastics have also been developed by the industry, and
the application of these materials has been increasing.
Included in the category of plastics are composites like glass fiber and carbon fiber
reinforced polymers. While these more costly advanced materials have primarily been used
in a limited number of low production volume vehicle applications, some manufacturers are
considering these composites for broader use. While these materials currently have the
potential to be applied to components with little or no exposure to impact pulses, the
advanced microstructure and limited industry experience may make these longer-term
solutions. For example, advanced composite materials (such as carbon fiber-reinforced
plastic), depending on the specific fiber, matrix, reinforcement architecture, and processing
method, can be subject to dozens of competing damage and failure mechanisms that may
complicate a manufacturer's ability to ensure equivalent levels of durability and
crashworthiness. As the industry gains experience with these materials, these concerns will
inevitably diminish, but may remain relevant during the timeframe of this final rulemaking.
In practice, material substitution tends to be quite specific to the manufacturer and
situation. Some materials work better than others for particular vehicle components, and a
manufacturer may invest more heavily in adjusting to a particular type of advanced material,
thus complicating its ability to consider others. The agencies recognize that like any type of
mass reduction, material substitution has to be conducted not only with consideration to
maintaining equivalent component strength, but also to maintaining all the other attributes of
that component, system or vehicle, such as crashworthiness, durability, and NVH.
If vehicle mass is reduced sufficiently through application of the two primary
strategies of using less material and material substitution described above, secondary mass
reduction options may become available. Secondary mass reduction is enabled when the load
requirements of a component are reduced as a result of primary mass reduction. If the
3-222
-------
Technologies Considered in the Agencies' Analysis
primary mass reduction reaches a sufficient level, a manufacturer may use a smaller, lighter,
and potentially more efficient powertrain while maintaining vehicle acceleration performance.
If a powertrain is downsized, approximately half of the mass reduction may be attributed to
the reduced torque requirement which results from the lower vehicle mass. The lower torque
requirement enables a reduction in engine displacement, changes to transmission including
the torque converter and gear ratios, and changes to final drive gear ratio. The reduced
powertrain torque enables the downsizing and/or mass reduction of powertrain components
and accompanying reduced rotating mass (e.g., for transmission, driveshafts/halfshafts,
wheels, and tires) without sacrificing powertrain durability. Likewise, the combined mass
reductions of the engine, drivetrain, and body in turn reduce stresses on the suspension
components, steering components, wheels, tires, and brakes, which can allow further
reductions in the mass of these subsystems. Reducing the unsprung masses such as the
brakes, control arms, wheels, and tires further reduce stresses in the suspension mounting
points, which will allow for further optimization and potential mass reduction.
Secondary mass reduction can occur for each kg of primary mass reduction, when all
subsystems are redesigned to take the initial primary mass reduction into account. In the MYs
2012-2016 rulemaking analysis, the agencies assumed that 1 kg of primary mass reduction
could enable up to 1.25 kg of secondary mass reduction. In the two most recent mass
reduction projects by EPA and NHTSA, every 1 kg of primary mass reduction enabled 0.7 kg
of secondary mass reduction. We note that these estimates may not be applicable in all real-
world instances of mass reduction, and that the literature indicates that the amount of
secondary mass reduction potentially available varies significantly from an additional 0.5 kg
to 1.25 kg per 1 kg of primary mass reduction, depending on assumptions such as which
components or systems primary mass reduction is applied to, and whether the powertrain is
available for downsizing. 94>95>96 The amount of secondary mass reduction is also affected by
the degree of component sharing that occurs among a manufacturer's models. Component
sharing is used by manufacturers to achieve production economies of scale that affect cost and
the number of unique parts that must be managed in production and for service. In addition,
the engineering resources and capital for tooling and equipment that would be needed to
optimize every vehicle component at each redesign affects the ability to fully optimize a new
vehicle to achieve all of the theoretically possible secondary mass reduction. While there is
agreement in the literature that primary mass reduction can enable secondary mass reduction,
the agencies recognize that care must be taken when reviewing reports on mass reduction
methods and practices to ascertain the manner and extent to which compounding effects have
been considered.
All manufacturers are using some or all of these methods to reduce mass in the
vehicles they are producing today, and the agencies expect that the industry will continue to
learn and improve the application of these techniques for more vehicles during the rulemaking
timeframe. We consider mass reduction in net percentage terms in our analysis not only
because effectively determining specific appropriate mass reduction methods for each vehicle
in the baseline fleet is a large task beyond the scope of this rulemaking, but also because we
recognize that even as manufacturers reduce mass to make vehicles more efficient, they may
also be adding mass in the form of increased vehicle features and safety content in response to
market forces and other governmental regulations. For these reasons, when the agencies
3-223
-------
Technologies Considered in the Agencies' Analysis
discuss the amount of mass reduction that we are assuming is feasible for purposes of our
analysis, we are implicitly balancing both the considerable opportunities that we believe exist
for mass reduction in the future, and the reality that vehicle manufacturing is complex and
that mass reduction methods must be applied thoughtfully and judiciously as safety and
content demands on vehicles continue to increase over time. Despite our considerable
discussion of the topic, the agencies' application of mass reduction in our analysis is fairly
simplified. As applied in our models, the percentage reduction for a given vehicle that is
assumed for a given year is an abstraction of all the specific mass reduction methods
described above.
How much mass reduction do the agencies believe is feasible in the rulemaking timeframe?
Feasibility, if narrowly defined as the ability to reduce mass without any constraints, is
nearly unbounded. However, in practice, the feasible amount of mass reduction is affected by
other considerations. Cost effectiveness is one of those constraints and is discussed further
below in the mass reduction cost section. In the analysis for the current rulemaking for MYs
2017-2025, the agencies reviewed a number of public reports and accompanying data, as well
as confidential information from manufacturers, and believe that mass reduction of up to 20
percent from a MY 2008 baseline vehicle can be achieved in a cost effective manner using
technologies currently in production. More detail on studies reviewed by the agencies and
additional studies currently in progress by the agencies is located below in Table 3-9 and in
the paragraphs below under the question " What additional studies are the agencies
conducting to inform our estimates of mass reduction amounts, cost, and effectiveness?"
From a general planning perspective, nearly all automakers have made some public
statement regarding vehicle mass reduction being a core part of the overall technology
strategy that they will utilize to achieve future fuel economy and CC>2 emission standards.
Estimates from Ducker Worldwide indicate that the automobile industry will see
an annual increase in AHSS of about 10% through 202097.
Ford has stated that it intends to reduce the weight of its vehicles by 250-750 Ib
QO
per model from 2011 to 2020 . For context, the midpoint of that range of
reductions would correspond to a 12% reduction from the current Ford new light
duty vehicle sales fleet.
Mazda has released a statement about achieving a 220-lb reduction per vehicle by
201699. This is equivalent to about a 6% reduction for the company's current fleet.
- Land Rover executives have stated that the company remains committed to a goal
of reducing curb weights of its SUVs by as much as 500 kilograms over the next
lOyears10^
In its comment to the NPRM, Volkswagen stated that they expect to reduce the
mass of their vehicles by 7-10% on average during the period of this regulation.
Several reports focusing on the OEM's approaches for light weighting are
summarized in the University of California Davis study as shown in Table 3-126 101.
3-224
-------
Technologies Considered in the Agencies' Analysis
Table 3-126 Automaker industry statements regarding plans for vehicle mass-reduction technology
Affiliation
General
Motors
Ford
Nissan
BMW
Volkswagen
Fiat
Volkswagen
BMW
BMW
Quote
"We use a lot of aluminum today - about 300 pounds per vehicle - and are likely to
use more lightweight materials in the future"
"The use of advanced materials such as magnesium, aluminum and ultra high-
strength boron steel offers automakers structural strength at a reduced weight to help
improve fuel economy and meet safety and durability requirements"
"We are working to reduce the thickness of steel sheet by enhancing the stregnth,
expanding the use of aluminum and other lightweight materials, and reducing vehicle
weight by rationalizing vehicle body structure"
"Lightweight construction is a core aspect for sustainable mobility improving both
fuel consumption and CO2 emissions, two key elements of our Efficient Dynamics
strategy ... we will be able to produce carbon fiber components in large volumes at
competitive costs for the first time. This is particularly relevant for electric -powered
vehicles."
"Material design and manufacturing technologies remain key technologies in vehicle
development. Only integrated approaches that work on these three key technologies
will be successful in the future. In addition to the development of metals and light
metals, the research on fibre -reinforced plastics will play a major role."
"A reduction of fuel consumption attains big importance because of the possible
economical savings. In order to achieve that, different ways are followed: alternative
engine concepts (for example electric engines instead of combustion ones) or weight
reduction of the vehicle structure. Using lightweight materials and different joining
techniques hleps to reach this aim"
"Lightweight design is a key measure for reducing vehicle fuel consumption along
with powertrain efficiency, aerodynamics and electrical power management"
"A dynamic vehicle with a low fuel consumption finally demands a stiff body with a
low weight. To achieve the initially mentioned targets, it is therefore necessary to
design a body which offers good stiffness values and a high level of passive safety at
a low weight."
"Light weight design can be achieved by engineering light weight, manufacutring
light weight and material light weight design."
Source
Keith, 2010
BMW and
SQL, 2010
Goede et al,
2009
Nunez, 2009
Goede et al,
2009
Nunez, 2009
Krinke, 2009
Prestorf,
2009
Prestorf,
2009
Although the focus on mass reduction by manufacturers is widespread, the agencies
believe the practical limits of mass reduction will be different for each vehicle model as each
model starts with a different mix of conventional and advanced materials, components, and
features intended to meet the function and price of a particular market segment. A vehicle
that already has a significant fraction of advanced high strength steel (AHSS) or any other
advanced material in its structure, for example, will not have the opportunity to realize the
same percentage of mass reduction as a vehicle of more traditional construction. Given the
myriad methods of achieving mass reduction, and the difficulty in obtaining data, accounting
for the current level of mass reduction technology for every model in production in a baseline
model year would be an impractical task. However, the agencies believe that reducing
vehicle weight to reduce fuel consumption has a continuum of solutions and the technologies
employed will have levels of effectiveness and feasibility that will vary by manufacturers and
by vehicle. In estimating the amount of mass reduction for this analysis, the agencies also
consider fleet safety effects for mass reduction. See Section II. G of the preamble for a
detailed discussion of the safety considerations in establishing CAFE and GHG standards. In
the CAFE and OMEGA analyses, the agencies considered several levels of mass reduction
applicable to all of the models in each subclass, as discussed below.
3-225
-------
Technologies Considered in the Agencies' Analysis
Based on the many aspects of mass reduction (i.e.., feasibility, cost and safety), for the
final rule, consistent with the proposal, the agencies believe that mass reduction of up to 20
percent is feasible on light trucks, CUVs and minivans, but that less mass reduction should be
implemented on other vehicle types to avoid increases in societal fatalities. While the
agencies continue to examine mass reduction, we remain alert to safety considerations and
seek to ensure that any CAFE and CC>2 standards can be achieved in a safety-neutral or
improved manner.
In the CAFE model, NHTSA applied amounts of mass reduction shown in Table
3-127, which was based on the ability to achieve overall fleet fatality estimates of close to
zero. The results are described in Preamble Section II.G and Chapter V of NHTSA's RIA.
The amount of mass reduction applied in EPA's OMEGA model follows the safety neutral
analysis is described in Section II.G of the Preamble with a variety of tables in EPA's RIA
(Chapter 3.8.2).
Table 3-127 MAXIMUM MASS REDUCTION AMOUNT APPLIED IN CAFE MODEL
Absolute
%
MR1*
MR2
MRS
MR4
MRS
Subcompact
and
Subcompact
Perf. PC
Compact
and
Compact
Perf. PC
Midsize PC
and Midsize
Perf. PC
Large PC
and Large
Perf. PC
Minivan LT
Small,
Midsize and
Large LT
0.0% 0.0% 1.5% 1.5% 1.5% 1.5%
0.0% 0.0% 3.5% 7.5% 7.5% 7.5%
0.0% 0.0% 0.0% 10.0% 10.0% 10.0%
0.0% 0.0% 0.0% 0.0% 15.0% 15.0%
0.0% 0.0% 0.0% 0.0% 20.0% 20.0%
Notes:
*MR1-MR5: different levels of mass reduction used in CAFE model
The amounts of mass reduction shown in Table 3-127 are for conventional vehicles.
The agencies assume that vehicles with hybrid and electric powertrain are heavier than
conventional vehicles because of the mass of battery systems. In comparing anecdotal data
for HEVs, EPA and NHTSA assume a slight weight increase of 4-5% for HEVs as 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 as stated in TSD
section 3.4.3.8. We believe that this assumption accurately reflects real-world HEV, PHEV
and EV construction. As an example, for a Subcompact PHEV with 20 mile range operating
on electricity, the agencies assume that to achieve no change in total vehicle mass, it would be
necessary to reduce the mass of the glider by 6 percent because of the additional weight of the
electrification system. The mass reduction for P/H/EVs can be found section 3.3.3.9 in the
joint TSD, and in EPA's RIA Chapter 1 and Chapter V, section E.3.h.4, of NHTSA's FRIA.
How much do the agencies estimate mass reduction will cost in the rulemaking timeframe?
Automakers are currently utilizing various mass reduction techniques across the light-
duty vehicle fleet, and will continue to use and in some cases expand these approaches for the
3-226
-------
Technologies Considered in the Agencies' Analysis
2017 to 2025 time frame. These approaches may include optimized design, geometry, part
consolidations, and materials substitution. Unlike the other technologies described in this
chapter, mass reduction is potentially more complex in that we cannot define it as a single
piece of equipment or hardware change to implement the technological improvement. Mass
reduction, depending upon the level of reduction targeted, has the potential to impact nearly
every system on the vehicle. Because of this complexity, there are unique challenges to
estimating the cost for mass reduction and for demonstrating the feasibility of reducing
vehicle mass by a given amount. This section describes the cost estimates used for the
agencies' analysis.
In the analysis for the MYs 2012-2016 rulemaking, the agencies assumed a constant
cost for mass reduction of $1.32 for each pound reduced up to a mass reduction level of 10
percent (or $1.48/lb using an ICM factor of 1.1 for a low-complexity technology). The
$1.32/lb estimate was based on averaging three studies: the 2002 NAS Report, a 2008 study
by Sierra Research, and a 2007 study by MIT researchers.eee
Since the MYs 2012-2016 final rule, the agencies have given further consideration to
the cost of mass reduction, and now believe that a cost that varies with the level of mass
reduction provides a better estimate. The agencies believe that as the vehicle fleet progresses
from lower to higher levels of mass reduction and becomes increasingly optimized for mass
and other attributes, the cost for mass reduction will progressively increase. The higher levels
of mass reduction may, for example, require applying more advanced materials and
technologies than lower levels of mass reduction, which means that the cost of achieving
those higher levels may increase accordingly. The unit cost of mass reduction versus the
amount of mass reduction might be linear, parabolic, or some other higher order relationship.
In the 2017-2025 Notice of Intent, 75 FR 62739 (Oct. 13, 2010), CARB, EPA and NHTSA
derived a second order curve based on a study with two vehicle redesigns conducted by Lotus
Engineering completed in 2010, such that zero mass reduction had zero cost, and the dollars
per pound increased with greater levels of mass reduction. Since the publication of the TAR,
the agencies have identified a number of additional studies in the literature relating to the
costs of vehicle mass reduction, which are discussed below. The studies show that for low or
high mass reduction, the costs can range from small cost savings to significant cost increases.
The economic costs associated with mass reduction are difficult to determine conclusively
due to the broad range of methods employed to achieve mass reduction. The costs on a
specific vehicle or component depend on many factors, such as the design, materials selected,
raw material price, appropriate manufacturing processes, production volume, component
functionality, required engineering and development, etc.
eee Specifically, the 2002 NAS Report estimated that vehicle weight could be reduced by 5 percent (without
engine downsizing) at a cost of $210-$350, which translates into $1.50/lb assuming a 3,800 Ib base vehicle and
using the midpoint cost; Sierra Research estimated that a 10 percent reduction (with compounding) could be
accomplished for $1.01/lb, and MIT researchers estimated that a 14 percent reduction (with no compounding)
could be accomplished for $1.36/lb. References for these studies are available in endnotes to Chapter 3 of the
TSD for the MYs 2012-2016 final rule.
3-227
-------
Technologies Considered in the Agencies' Analysis
Cost data thus varies widely in the literature. Of the various studies reviewed by the
agencies, not all are equal in their original intent, rigor, transparency, or applicability to this
regulatory purpose. The individual studies range from complete vehicle redesign to advanced
optimization of individual components, and were conducted by researchers with a wide range
of experience and background. Some of the studies were literature reviews, while others
developed new designs for lighter components or complete lighter vehicles, while yet others
built physical components or systems, and conducted testing on those components and
systems. Some of the studies focused only on a certain sub-system (which is a building block
for the overall vehicle design), while some of them took a systematical approach and re-
designed the whole vehicle to achieve the maximum mass reduction and cost reduction. The
latter studies typically identified a specific baseline vehicle, and then utilized different
engineering approaches and investigated a variety of mass-reduction concepts that could be
applied to that vehicle. Some of the differences between studies emanate from the
characteristics of the baseline vehicle and its adaptability to the new technology or method,
and the cost assumptions relating to the original components and the redesigned components.
Assumptions regarding the degree and cost of any associated mass de-compounding can also
confound comparisons. Despite this variation in the literature, in actual practice, we
believe manufacturers will choose a target mass reduction for a whole vehicle and for each
sub-system, and work to find the lowest total cost method to achieve those targets. Such a
process would consider numerous primary and secondary cost factors (including engineering,
facilities, equipment, tooling, and retraining costs) as well as technological and manufacturing
risks.ggg
Regardless of the confidence in specific estimates, the agencies must select a curve
that will be applied to the whole fleet that will define the average cost per kg of mass
reduction as a function of total percentage of mass reduction. There are many significant
challenges that make it difficult for the agencies to establish an estimated cost curve based on
the literature, such as the differences in the baselines used in the studies, whether the studies
considered platform sharing and powertrain sharing, and other considerations.
fff The concept of secondary weight savings or mass compounding (also called mass decompounding) derives
from the qualitative understanding that as vehicle weight decreases, other vehicle systems can also decrease in
mass while maintaining the original vehicle level of performance and function. For instance, following a primary
weight reduction in the vehicle (e.g. Body in White), the designs of some of the other dependant vehicle
subsystems (tires, suspensions, brakes, powertrain, body structure) may be redesigned and reduced in mass to
account for the overall lighter vehicle. The lighter vehicle is also associated with lighter loads, less friction and
drag, and may require less power to be accelerated, and the powertrain may therefore be scaled down in size with
a potential for reduced mass, even while maintaining equivalent acceleration performance and functionality. The
compounded or secondary mass savings from these additional systems may then drive further mass reductions in
the original primary weight reduction (e.g. Body in White). Mass compounding factors found in literature are
rough estimates of the secondary mass reduction amount.
888 We also note that the cost of mass reduction in the CAFE model is quantified on a per pound basis that is a
function of the percentage decrease in vehicle mass. We assume that OEMs would find the most cost-effective
approach to achieve such a mass reduction. Realistically, this would depend heavily on the baseline vehicle as
well as the size and adaptability of the initial design to the new technology. Thus, the CAFE model strives to be
realistic in the aggregate while recognizing that the figures proposed for any specific model may be debatable.
3-228
-------
Technologies Considered in the Agencies' Analysis
The costs for mass reduction employed for the main analysis for this final rule are the
same as those in the NPRM. The agencies considered updating cost estimates based on the
studies that were underway when the NPRM was issued. Those studies included the
EPA/ICCT funded Phase 2 Toyota Venza Low Development project and the NHTSA funded
Honda Accord mass reduction project, which are described in the section titled "What
additional studies are the agencies conducting to inform our estimates of mass reduction
amounts, cost, and effectiveness? " However, these studies were in the middle of the peer
review process and had not yet been finalized at the time when the inputs for the main
analysis for this final rule were required. For the final rule, the agencies decided to continue
to use the same costs for mass reduction that were used in the NRPM.
The agencies examined all the studies in Table 3-128 including information supplied
by manufacturers (during meetings held subsequent to the TAR) when deciding the mass
reduction cost estimate used for the proposal, which has been carried forward for this FRM.
The agencies considered three major factors in examining these studies. First, whether a
study was rigorous in terms of how it evaluates and validates mass reduction from
technological and design perspectives. This includes consideration of a study's
comprehensiveness, the technical rigor of its methodology, the validation methods employed,
and the relevance of the technologies evaluated in the study given our rulemaking time frame.
Second, whether a study was rigorous in terms of its estimation of costs, including the
completeness and rigor of the methodology, such as whether the study includes data for all
categories of direct manufacturing costs, and whether the study presents detailed cost
information for both the baseline and the light-weighted design. And third, the degree of peer
review, including if the study is peer-reviewed, and whether it has effectively addressed any
critical technical, methodological, and cost issues raised by the peer-review, if this
information is available.
Some of the variation may be attributed to the complexity of mass reduction as it is
not one single discrete technology and can have direct as well as indirect effects on other
systems and components. The 2011 NAS study speaks to this point when it states on page 7-1
that "[t]he term material substitution oversimplifies the complexity of introducing advanced
materials, because seldom does one part change without changing others around it." These
variations underscore that there is not a unique mass reduction solution as there are many
different methods with varying costs for taking mass out of vehicles, and every manufacturer,
even every vehicle, could have a different approach depending on the specific vehicle,
assembly plant and model year of implementation. The agencies recognize that there are
challenges to characterizing the mass reduction plans for the entire future fleet due to the
complexity and variety of methods available. So far the agencies have not found any study
that addresses how to generalize the mass reduction that is achievable on a single vehicle to
^ The agencies considered confidential cost information provided by OEMs that covered a range of
components, systems, designs and materials. Some of these cost estimates are higher than some of the literature
studies, and manufacturers provided varying levels of detail on the basis for the costs such as whether mass
compounding is included, or whether the costs include markup factors.
3-229
-------
Technologies Considered in the Agencies' Analysis
the whole fleet. Table 3-128 contains a summary of the data contained in the studies, and the
OEM CBI data, which the agencies reviewed. There is a degree of uncertainty associated
with comparing the costs from the range of studies in the literature when trying to summarize
them in a single table, and we encourage interested stakeholders to carefully review the
information in the literature. For some of the cost estimates presented in the papers there are
unknowns such as: what year the costs are estimated for, whether mass decompounding (and
potential resultant cost savings) was taken into account, and whether mark-ups or indirect
costs were included. The agencies tried to normalize the cost estimations from all these
studies by converting them to 2009 year dollar, applying mass compounding factor of 1.35 for
mass reduction amount more than 10 percent if it has not been applied in the study and
factoring out the RPE specified in the study to derive direct manufacture costs for
comparison. There are some papers that give cost for only component mass reduction, others
that have more general subsystem costs and others yet that estimate total vehicle mass
reduction costs (which often include and present data at the subsystem level). Other studies
have multiple scenarios for different materials, different vehicle structures and mass reduction
strategies. Thus, a single study which contains more than one vehicle can be broken down
into a range of vehicle types, or at the subsystem level, or even at the component level. While
Table 3-128 is inclusive of all of the information reviewed by the agencies for the NPRM, for
the reasons described above the technical staff for the two agencies applied various
approaches in evaluating the information. The linear mass-cost relationship developed for the
proposal is carried forward to this final rule and presented below is the consensus assessment
from the two agencies of the appropriate mass cost for this final rule.
Table 3-128 Mass Reduction Studies Considered for Estimating Mass Reduction Cost for this FRM
studies
ra
01
tt
8
Cost Information from Studies
S1
c
.0
«
3
-a
oi •— '
1 £
"oi .SP
U) 'at
ra -S
nn s
HO °^,
= BO
'c c
•??
^ B
0) Q.
DC £
ui o
i "§"
•in-
tt
8
LLJ
Q_
ce.
O
Ol
"5.
4-*
^J
i_
= S
O o
O (N
tt
8
bo
c
X =
L. W
5 -2
R 1
a §
vt
VI i — i
(U -Q
MS"
to O
-S ts
D Di
Individual Cost Data Points
AISI, 1998(ULSAB)
AISI, 2000 (ULSAC)
Austin et al, 2008 (Sierra Research) -
ULS Unibody
Austin et al, 2008 (Sierra Research) - AL
Unibody
Austin et al, 2008 (Sierra Research) -
ULS BoF
Austin et al, 2008 (Sierra Research) - AL
BoF
Bull et al, 2008 (Alum Assoc.) - AL BIW
Bull et al, 2008 (Alum Assoc.) - AL
Closure
Bull et al, 2008 (Alum Assoc.) - Whole
Vehicle
1998
2000
2008
2008
2008
2008
2008
2008
2008
103
6
320
573
176
298
279
70
573
1
1
1
1
1
1
1
1
1
103
6
320
573
176
298
279
70
573
2977
2977
3200
3200
4500
4500
3378
3378
3378
3.5%
0.2%
10.0%
17.9%
3.9%
6.6%
8.3%
2.1%
17.0%
-$32
$15
$209
$1,805
$171
$1,411
$455
$151
$122
1.0
1.0
1.61
1.61
1.61
1.61
1.0
1.0
1.0
1.28
1.24
1.01
1.01
1.01
1.01
1.01
1.01
1.03
-$41
$18
$131
$1,134
$107
$887
$460
$153
$126
-$0.40
$2.99
$0.41
$1.98
$0.61
$2.98
$1.65
$2.17
$0.22
3-230
-------
Technologies Considered in the Agencies' Analysis
Cheah et al, 2007 (MIT) - 20%
Das, 2008 (ORNL) - AL Body & Panel
Das, 2008 (ORNL)-FRPMC
Das, 2009 (ORNL) - CF Body & Panel, AL
Chassis
Das, 2010 (ORNL) - CF Body & Panel,
Mg Chassis
EEA, 2007 - Midsize Car - Adv Steel
EEA, 2007- Midsize Car- Plast/Comp
EEA, 2007- Midsize Car-AI
EEA, 2007 - Midsize Car- Mg
EEA, 2007 - Light Truck - Adv Steel
EEA, 2007 - Light Truck - Plast/Comp
EEA, 2007 -Light Truck - Al
EEA, 2007 -Light Truck - Mg
Geek etal, 2008 (Ford)
Lotus, 2010 - LD
Lotus, 2010- HD
Montalbo et al, 2008 (GM/MIT) -
Closure-HSS
Montalbo et al, 2008 (GM/MIT) -
Closure- AL
Montalbo et al, 2008 (GM/MIT) -
Closure- Mg/AL
Plotkin et al, 2009 (Argonne)
2007
2008
2008
2009
2010
2007
2007
2007
2007
2007
2007
2007
2007
2008
2010
2010
2008
2008
2008
2009
712
637
536
933
1173
236
254
586
712
422
456
873
1026
1310
660
1217
25
120
139
683
1
1
1.0
1
1
1
1
1.35
1.35
1
1
1.35
1.35
1
1
1
1
1
1
1
712
637
536
933
1173
236
254
791
961
422
456
1179
1385
1310
660
1217
25
120
139
683
3560
3363
3363
3363
3363
3350
3350
3350
3350
4750
4750
4750
4750
5250
3740
3740
4000
4000
4000
3250
20.0%
19.0%
15.9%
27.7%
34.9%
7.0%
7.6%
23.6%
28.7%
8.9%
9.6%
24.8%
29.2%
25.0%
17.6%
32.5%
0.6%
3.0%
3.5%
21.0%
$646
$180
-$280
$1,490
$373
$179
$239
$1,388
$1,508
$291
$398
$1,830
$1,976
$500
-$121
$362
$10
$110
$110
$1,300
1.0
1.5
1.5
1.5
1.5
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
1.03
1.01
1.01
1.00
1.00
1.03
1.03
1.03
1.03
1.03
1.03
1.03
1.03
1.01
1.00
1.00
1.01
1.01
1.01
1.00
$667
$121
-$189
$993
$248
$185
$247
$1,434
$1,558
$301
$411
$1,891
$2,042
$506
-$120
$360
$10
$111
$111
$1,300
$0.94
$0.19
-$0.35
$1.06
$0.21
$0.78
$0.97
$1.81
$1.62
$0.71
$0.90
$1.60
$1.47
$0.39
-$0.18
$0.30
$0.41
$0.92
$0.80
$1.90
(... Continued) Mass Reduction Studies Considered for Estimating Mass Reduction Cost for this FRM
Studies
(5
01
to
8
Cost Information from Studies
5"
o
'r,
3
•D
Ol
CC
U)
0
as
u.
no
c
O
Q.
F
,s
c
K
^ -Q
2 u>
H .=
= -a
0 C
Ol 3
£ O
V) °-
V, C
3,s
£
M
'01
|u
!c
01
_c
!X
5
W) 2,
•— W)
l!
la
K £
i "5"
^i
to
3
LLJ
Q.
rf
at
8
fM
2
^
Ol
"a.
"3
>.
J5
O
Q
8
t>0
1
3
C
as
£
a
§
o ^
O
•g
3
01
ce.
in
ra
M-
o
3 „
.ti —
D S
Cost Curves
NAS, 2010
2010
2010
2010
2010
2010
1.0%
2.0%
5.0%
10.0%
20.0%
$ 1.41
$ 1.46
$ 1.65
$ 1.52
$ 1.88
3-231
-------
Technologies Considered in the Agencies' Analysis
OEMl
OEM2
OEMS
OEM4
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2010
2011
2011
2011
8.0%
9.0%
9.5%
10.0%
11.0%
0.4%
0.9%
1.9%
2.3%
2.4%
3.1%
3.6%
4.0%
4.1%
4.5%
4.8%
5.0%
4.0%
7.5%
10.0%
6.9%
8.1%
16.4%
S 6.00
S 7.00
S 8.00
S 12.00
S 25.00
S
S o.io
S 0.20
S 0.33
S 0.38
S 0.60
S 0.76
S 0.85
S 0.88
S 0.98
S 1.09
S 1.17
S 0.57
S 1.01
S 1.51
S 0.97
S 1.02
S 1.95
EPA and NHTSA scrutinized the various available studies in the literature as well as
confidential information provided by several auto firms based on the kinds of factors
described above for purposes of estimating the cost of mass-reduction in the 2017-2025
timeframe. We determined that there was wide variation across the studies with respect to
costs estimates, applicability to the 2017-2025 time frame, and technical rigor. The mass cost
curve that was developed is defined by the following equation and is shown in
Figure 3-33:
Mass Reduction Direct Manufacturing Cost (DMC) ($/lb) = $4.36/(%-lb) x Percentage of
Mass Reduction Level (%) (2010$)
3-232
-------
Technologies Considered in the Agencies' Analysis
$1.00
$0.90
$0.80
S- $0.70
£> $0.60
£ $0.50
~ $0.40
= $0.30
$0.20
$0.10
$0.00
0
Mass Reduction Cost
jrf
X
^S\
^
J^
XL
Slope = 4.S36 ^X*
_x
X
^
^T
x^
% 5% 10% 15% 20% 25%
Percent of Mass Reduction
Figure 3-33 NPRM and FRM Mass Reduction Direct Manufacturing Cost
For example, this results in an estimated $175 cost increase for a 10% mass reduction of a
4,0001b vehicle (or $0.44/lb), and a $394 cost increase for 15% reduction on the same vehicle
(or $0.66/lb).
As mentioned in the NPRM, due to the wide variation in data used to select this
estimated cost curve, the agencies have also conducted cost sensitivity studies in their
respective RIAs in both the proposal and final rule using values of+/-40%. The wide
variability in the applicability and rigor of the studies also provides justification for continued
research in this field.
The agencies consider this DMC to be applicable to the MY2017 and consider mass
reduction technology to be on the flat portion of the learning curve in the MY2017-2025
timeframe. To estimate indirect costs for applied mass reduction of up to 15%, the agencies
have applied a low complexity ICM of 1.24 through 2018 and 1.19 thereafter. To estimate
indirect costs for applied mass reduction of 15% to 25%, the agencies have applied a medium
complexity ICM of 1.39 through 2024 and 1.29 thereafter. To estimate indirect costs for
applied mass reduction greater than 25%, the agencies believe it is appropriate to apply a
highl complexity ICM of 1.56 through 2024 and 1.35 thereafter.
The agencies sought comment in the draft Joint TSD for the NPRM (p. 210) regarding
options for realistically and appropriately assessing the degree of feasible mass reduction for
vehicles in the rulemaking timeframe and the total costs to achieve that mass reduction, but
got no specific response. The agencies also sought comments on what practical limiting
factors need to be considered when considering maximum feasible amount of mass reduction;
the degree to which these limiting factors will impact the amount of feasible mass reduction
3-233
-------
Technologies Considered in the Agencies' Analysis
(in terms of the percent of mass reduction); the best method(s) to assess an appropriate and
feasible fleet-wide amount mass reduction amount (because each study mainly focuses on a
single vehicle); etc. In its comments, VW stated that it "projects full vehicle weight reductions
during the time period of this regulation on average in the order of 7-10%." VW noted that
this was lower than the agencies' estimates in the NPRM of upwards of 20% mass reduction
for large cars and some trucks, which VW stated may exceed cost effective limits. As stated
later in this section, the detailed studies sponsored by the agencies suggest that 20% mass
reduction is likely feasible for the rulemaking period without using exotic materials or highly
advanced technologies. The accompanying detailed cost analysis indicates that the cost of
reducing mass by 20% can potentially be economical. The agencies also noted in the NPRM
that we expected to refine our estimate of both the amount and the cost of mass reduction
between the NPRM and the final rule based on the agencies' ongoing work described a later
section, below. As stated before, due to the limited time and the extensive scope of these
studies, the agencies did not finish them in time for inclusion in the final rule analysis.
How effective do the agencies estimate that mass reduction will be?
A rule of thumb used by researchers and industry, based on testing and simulation, is
that 10 percent reduction in vehicle mass can be expected to generate a 6 to 8 percent increase
in fuel economy if the vehicle powertrain and other components are also downsized
accordingly.102 In the analysis for the MYs 2012-2016 final rule, NHTSA and EPA estimated
that a 10 percent mass reduction with engine downsizing would result in a 6.5 percent
reduction in fuel consumption while maintaining equivalent vehicle performance (i.e., 0-60
mph time, towing capacity, etc.), consistent with estimates in the 2002 NAS report. For small
amounts of mass reduction, such as the 1.5 percent used at vehicle refresh in NHTSA's
modeling, no engine downsizing was used, so a 10 percent mass reduction without engine
downsizing was assumed to result in a 3.5 percent reduction in fuel consumption. In this
FRM, both agencies have chosen to use the effectiveness value for mass reduction from
EPA's lumped parameter model to maintain consistency. EPA's lumped parameter model-
estimated mass reduction effectiveness is based on a simulation model developed by Ricardo,
Inc. under contract to EPA. The 2011 Ricardo simulation results show an effectiveness of 5.1
percent for every 10 percent reduction in mass. NHTSA has assumed that for mass reduction
amounts less than 10 percent, the effectiveness is 3.5 percent. For mass reduction greater than
10 percent, NHTSA estimates the effectiveness is 5.1 percent in order to avoid double
counting benefits - because the effectiveness of engine downsizing is included in the
effectiveness of the engine decision tree when applying engine downsizing, it should
appropriately be removed from the mass reduction effectiveness value in the mass reduction
decision tree. EPA applies an effectiveness of 5.1 percent for every 10 percent mass
reduction, and this scales linearly from 0 percent mass reduction, up to the maximum applied
mass reduction for any given vehicle, which in this final rule is never larger than 20 percent.
What additional studies are the agencies conducting to inform our estimates of mass
reduction amounts, cost, and effectiveness?
In the MYs 2012-2016 final rule, the agencies stated that there are several areas
concerning vehicle mass reduction and vehicle safety on which the agencies would focus their
research efforts and undertake further study. The following vehicle level projects focus on the
3-234
-------
Technologies Considered in the Agencies' Analysis
goals stated in the MYs 2012-2016 final rule, which include determining the maximum
potential for mass reduction in the MY 2017-2025 timeframe by using advanced materials and
improved designs while continuing to meeting safety regulations and voluntary guidelines and
while maintaining all aspects of vehicle functionality. The fourth study investigates the
effects of resultant study designs on fleet safety by evaluating crash performance with objects
and other vehicles of different size and mass.
1. NHTSA sponsored mass reduction study on a Honda Accord
2. EPA sponsored mass reduction study on a Toyota Venza (Phase 2 Low
Development)
3. California Air Resources Board mass reduction study on a Toyota Venza (Phase 2
High Development)
4. NHTSA fleet-wide simulation study - crash analysis using the resultant designs
from the studies 1-3 with objects and the design models of other vehicles with
different size and mass.
Due to the extensive scope of work for these studies and tight time schedule, some of the
studies were finished, but peer reviews and response to peer reviews were not completed in
time to enable the results to inform the final rule. We note, however, that the intermediate
results from the mass reduction studies would corroborate the level of feasible amount of
mass reduction the agencies chose to apply in the NPRM and FRM analyses. Rulemaking
modeling results show that the costs for mass reduction are not sensitive to the cost curve of
the rulemaking. In the NPRM, EPA found that a +/- 40% change in the cost of mass
reduction had very little impact on the cost of the program. This is largely because of safety
restraints imposed in the amount of mass reduction selected for the various vehicle classes
primarily drive the penetration rates of the technology, rather than the relative cost-
effectiveness of the technology itself.
The following sections describe the status and results of the studies sponsored by the
agencies.
NHTSA Sponsored Mass Reduction Study
BACKGROUND: NHTSA awarded a contract in December 2010 to Electricore, with
EDAG and George Washington University (GWU) as subcontractors, to study the maximum
feasible amount of mass reduction for a mid-size car - specifically, a Honda Accord - while
keeping the vehicle functionality the same as the baseline vehicle. The
Electricore/EDAG/GWU project team was charged with maximizing the amount of mass
reduction using technologies that are considered feasible for production of 200,000 units per
year during the time frame of this rulemaking while maintaining retail price in parity (within
±10%) with the baseline vehicle. In addition, all designs, materials, technologies and
manufacturing processes must be realistically projected to be feasible for industry-wide
application in MYs 2017-2025. The project focused on mass reduction and allowed
powertrain downsizing, however alternative powertrains, such as diesels, HEVs and EVs,
were not to be considered.
3-235
-------
Technologies Considered in the Agencies' Analysis
MATERIAL AND TECHNOLOGY SELECTION: For vehicle redesigns, OEMs
normally select technologies, materials and manufacturing processes that are currently in use
on existing vehicle platforms or planned to be in use on future vehicle platforms. The use of
the same or similar technologies, materials and manufacturing processes helps maintain or
improve component and vehicle reliability, manufacturability and cost. New materials,
technologies and processes are often introduced in low-volume, high price vehicles first and
then migrate to high production volume vehicle lines over time. This significantly reduces the
risk to OEMs associated with implementing new technologies. Recognizing this 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 (MY 2012-2015) on low-volume
production vehicles. The recommended materials (Advanced High Strength Steels,
Aluminum, Magnesium and Plastics) manufacturing processes (Stamping, Hot Stamping, Die
Casting, Extrusions, Roll Forming) and assembly methods (Spot welding, Laser welding 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 process parameters for manufacturing with Advanced
High Strength Steels can be supported by computer simulation. This approach minimized
those material and technology options which would likely be overly aggressive or unrealistic
to implement in mass production in model years 2017-2025.
ENGINEERING APPROACH: 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 torn down and benchmarked 2011
Honda Accord and then undertook a series of baseline, noting the designs, materials,
technologies and overall design optimization level of the baseline vehicle. Vehicle
performance, safety simulation and cost analyses were run in parallel to the design study to
help ensure that the design decisions for the concept vehicle would be informed by a well-
documented baseline, thus enabling the resultant design to meet the defined project criteria.
While working within the constraint of maintaining the baseline Honda Accord's
exterior size and shape, the body structure was first redesigned using topology optimization
with six load cases including bending stiffness, torsion stiffness, IIHS frontal impact, IIHS
side impact, FMVSS pole impact, FMVSS rear impact and FMVSS roof crush cases. The
load paths from topology optimization were analyzed and interpreted by technical experts and
the results were then fed into low fidelity 3G (Gauge, Grade and Geometry) optimization
programs to further optimize for material properties, material thicknesses and cross-sectional
shapes while trying to achieve the maximum amount of mass reduction. The
Electricore/EDAG/GWU team carefully reviewed the optimization results and built detailed
CAD/CAE models for the body structure, closures, bumpers, suspension, and instrumentation
panel. The vehicle designs were also carefully reviewed by manufacturing technical experts to
ensure that they could be manufactured at high volume production rates. Detailed
manufacturing layouts were created and were later used to estimate costs.
Multiple materials were used for this study. The body structure was redesigned using a
significant amount of advanced high strength steel (AHSS). The closure and suspension were
3-236
-------
Technologies Considered in the Agencies' Analysis
designed using a significant amount of aluminum. Magnesium was used for the
instrumentation cross-car beam. A limited amount of composite material was used for the seat
structure. Electricore and its sub-contractors consulted industry leaders and experts for each
component and sub-system when deciding which mass reduction technologies were feasible.
DESIGN AND FUNCTION VALIDATION: In order to ensure that the light weighted
vehicle had the same functionality as the baseline vehicle, Electricore and its sub-contractors
used the CAD/CAE/powertrain models and conducted simulation modeling. This is the first
mass reduction study that has been released publicly that includes such a broad array of
vehicle simulation modeling analyses to assess vehicle functionality and performance relative
to these critical attributes. These significant additional analyses provide greater confidence
that the designs employed in this study are more feasible for production implementation than
a study without these analyses, although the agency notes that significantly more testing and
validation work is required to refine and finalize a design for production.
• Safety: Safety performance of the light-weighted design is compared to the safety
rating of the baseline MY2011 Honda Accord for seven consumer information and
federal safety crash tests using LS-DYNA111. These seven tests are NCAP frontal test,
NCAP lateral MDB test, NCAP lateral pole test, IfflS roof crush, IfflS lateral MDB,
IMS front offset test, and FMVSS No. 301 rear impact tests. All tests achieved safety
performance equivalent to MY 2011 Honda Accord when comparing crash pulse and
passenger compartment intrusion levels, with no damage to the fuel tank. This study
does not include restraint systems and dummy which would be part of NHTSA's fleet
simulation study.
• Body Stiffness/ Ride and Handling/NVH: Vehicle body torsional and bending
stiffness are signatures for the vehicle structure performance. Higher stiffness is
generally associated with a refined ride and handling qualities. The baseline vehicle
body structure underwent testing for normal modes of vibration, and torsion and
bending stiffness. A detailed FEA model of the light-weighted structure was created
and analyzed using the MSC/NASTRAN simulation. The torsional stiffness of the
light-weighted design is 30% higher than the baseline vehicle while the bending
stiffness is 40% higher. The normal mode frequency test results for the light-weighted
body structure, which represents vehicle dynamic stiffness, also are within 2.3% of the
targets. These stiffness and modes results show that the light-weighted design will
have improved ride and handling and improved NVH performance comparing to a
vehicle with lower stiffness.
• Vehicle Ride and Handling: In the light-weighted design, the front suspension is
redesigned using a MacPherson strut instead of the heavier double wishbone used in
111 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.
3-237
-------
Technologies Considered in the Agencies' Analysis
the baseline vehicle. Vehicle ride and handling is evaluated using MSC/ADAMSJJJ
modeling on five maneuvers, fish-hook test, double lane change maneuver, pothole
test, 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 according
to the ISO standard show that the chosen suspension geometry and vehicle parameter
of the light-weighted design are within acceptable range for safe high speed
maneuvers. These simulations are performed to further validate the chosen light
weighted front suspension design.
• Durability: There are two types of durability, stress related and corrosion related.
Stress related durability for the light-weighted vehicle is evaluated using strain-based
analysis based on pot hole, 0.8G forward braking and 0.7G cornering road load cases
using ADAMS model. Results from the simulation show that the life of the light-
weighted vehicle body structure exceeds the targets. Although timing and funding did
not allow corrosion testing to be conducted, the Electricore/EDAG/GWU team
considered the properties of materials used, and the location and the functionality of
the components to avoid potential issues with corrosion.
• Powertrain Performance: The powertrain of the light-weighted vehicle is downsized
from 2.4L naturally aspirated engine to 1.8L naturally aspirated engine to maintain the
same vehicle acceleration and towing compared to the baseline 2011 Honda Accord. A
powertrain simulation tool PSAT1^ is used to verify and validate the light-weighted
vehicle for fuel economy and powertrain performance. The light-weighted vehicle
with 1.8L NA engine will have 32 mpg fuel economy with comparable 0-30 mph time,
0-60 mph time, quarter mile time, gradability and maximum speed at grade. The only
metrics that the light-weighted vehicle performs less than the baseline vehicle is
vehicle maximum speed (127 mph for the baseline Accord and 112 mph for the light-
weighted design) which the Electricore/EDAG/GWU team and NHTSA believe is
acceptable. As a result of the improved fuel economy, the fuel tank for the light-
weighted vehicle can be reduced from 18.5 gallon to 15.8 gallon with the same driving
range, which further reduced vehicle weight both by reducing fuel tank mass and the
mass of fuel carried by the vehicle.
• Manufacturability: The manufacturability of all proposed body structure panels were
then assessed using simulation tools, which included HYPER-FORM for stamping
parts, and other single step process simulation tools for parts manufactured using other
methods, such as hot stamping for B-pillar.
JJJ MSC/ ADAMS: Macneal-Schwendler Corporation/Automatic Dynamic Analysis of Mechanical Systems.
kkk PSAT is a plug-and-play architecture software that allows the user to build and evaluate a vehicle's fuel
economy and powertrain performance under varying load conditions and drive cycles. It uses MATLAB in a
Simulink environment to record data, calculate and input powertrain requirements based on driver demand and
current powertrain values. The software is sponsored by the U.S. Department of Energy and developed by
Argonne National Laboratory (ANL). http://www.transportation.anl.gov/modeling_simulation/PSAT/index.html
3-238
-------
Technologies Considered in the Agencies' Analysis
COST ANALYSIS; A detailed cost analysis for the light weighted design and cost
estimates for alternative design options were also conducted. For OEM-manufactured parts, a
detailed cost model was built based on a Technical Cost Modeling (TCM) approach
developed by the Massachusetts Institute of Technology (MIT) Materials Systems
Laboratory's research103 for estimating the manufacturing costs of OEM parts. The costs
were broken down into each of the operations involved in the manufacturing, such as for a
sheet metal part production by starting from blanking the steel coil, until the final operation to
fabricate the component. Total costs were then categorized into fixed cost, such as tooling,
equipment, and facilities; and variable costs such as labor, material, energy, and maintenance.
These costs were assessed through an interactive process between the product designer,
manufacturing engineers and cost analysts. For OEM-purchased parts, the costs were
estimated by consultation with experienced cost analysts and Tier 1 suppliers. Forty-one
concise spreadsheets are created for both the baseline vehicle and the light-weighted design in
the cost model to calculate both the manufacturing and assembly costs.
FINAL RESULTS: To achieve the same vehicle performance as the baseline vehicle,
the size of the engine for the light-weighted vehicle was proportionally reduced from 2.4L-
177 HP to 1.8L-140HP. Overall the complete light weight vehicle achieved a total weight
savings of 22 percent (332kg) relative to the baseline vehicle (1480 kg) at an incremental cost
increase of $319 or $0.96 per kg. Without the mass and cost reduction allowance for the
powertrain (including engine, transmission, fuel system, exhaust system and fuel) the mass
saving for the 'glider' is 24 percent (264 kg) at mass saving cost premium of $1.63 per kg of
mass saving. The Electricore/EDAG/GWU team also developed a cost curve to cover a range
of mass reduction levels from 0% to 28% for both the full vehicle with engine downsizing and
for the glider only. When developing the cost curves, the project team used data that were
developed in the study to derive a mass compounding factor (secondary mass reduction/total
mass reduction), which was determined to be 0.7. The cost curves are shown in Figure 3-34
and Figure 3-35.
3-239
-------
Technologies Considered in the Agencies' Analysis
Composite BIW &
Mag/Alum Closures
Aluminum BIW &
Closures & Chassis
Frames
Non Structural &
Aluminum Closures
IWV- Engineered Solution
AHSSBIW & Aluminum
Closures & Chassis Frames
with 1.8 L-140 HP Engine
Non structural
masses
Mass Compounding
PT, Chassis, BIW
AHSSBIW & closures
5.05S 10.0% 15.0% 20.0%
Mass Reduction {% of Curb Vehicle Weight)
Baseline Vehicle V $-
2.41-177 HP 0.0%
Figure 3-34 Mass Reduction Cost with Allowance for Powertrain Downsizing
I
's
00
*J
u
Ol
i_
5
n
01
01
9.00
8.00
7.00
0.00
Composite BIW &
Mag/Alum Closures
- Engineered So
AHSSBIW & Aluminum
Closures & Chassis Frames
Baseline Vehicle - '
2.4 L-177 HP
5.0% 10.0% 15.0% 20.0% 25.0% 30.0%
Mass Reduction (% of Curb Vehicle Weight w/o PT)
Figure 3-35. Mass Reduction Cost for the Glider Only
3-240
-------
Technologies Considered in the Agencies' Analysis
PEER REVIEW: The study has been peer reviewed by three technical experts from
the industry, academia and a DOE national lab. In the peer reviewer charge letter, the agency
asked the peer reviewers to comment on the following five specific items as well as any other
potential areas for comments.
• Assumptions and data sources
• Vehicle design and optimization methodology and its rigorousness
• Vehicle functionality and crashworthiness testing methodological rigor
• Vehicle manufacturing cost methodology and its rigorousness
• Conclusions and findings
Comments from peer reviewers were generally positive. The peer reviewers concurred
with the methodologies employed in the study and the technologies applied to the light-
weighted design, although one peer reviewer commented that not enough composite materials
were used in the design. One peer reviewer stated in his comments that "the main findings
appear to be based on sound economic and engineering principles." The peer reviewers stated
that the cost estimates developed in the study, particularly based on the TCM model, seem to
be reasonable, with one peer reviewer commenting the final cost is on the lower side and
another commenting it is on the higher side. All three peer reviewers looked into the details of
the CAE and cost modeling. One significant concern identified in the peer review was
whether the light-weighted vehicle maintained the same performance level in the NCAP side
MDB test. In response to that concern, the Electricore/EDAG/GWU team conducted
simulation testing and revised the B-pillar design, increasing the gauge for the steel for better
performance. Because NCAP only measures injuries to dummies and the crash performance
of the light-weighted design is based on the vehicle center of gravity crash pulse level, B-
pillar velocity and passenger compartment and intrusion, to assess correlation of the model
performance to the baseline vehicle, NHTSA asked a contractor who performs NHTSA's
NCAP testing to take additional measurements of the interior intrusion for the 2011 baseline
Honda Accord. The updated design and the Honda Accord test data showed similar intrusion
results for both NCAP and IIHS side impact tests, and those results support that the light-
weighted design could possibly achieve similar NCAP and IIHS ratings, especially when the
structure design is fine tuned with the restraint system design which NHTSA will study in the
fleet simulation study described later on in this section. For other peer review comments, the
Electricore/EDAG/GWU team addressed the comments fully in the report and also composed
a response to peer review comment document, which is included at the end of the report. The
final report104, CAE model and cost model, and peer review comments105 are available in
Docket No. NHTSA-2010-0131 and can also be found on NHTSA's website111.
EPA Sponsored Mass Reduction 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
3-241
-------
Technologies Considered in the Agencies' Analysis
feasibility, safety and cost of 20% mass reduction on a 2017-2020 production ready mid-size
crossover utility vehicle (CUV) specifically, a Toyota Venza while maintaining cost parity or
reduction. The EPA report is entitled "Light-Duty Vehicle Mass-Reduction and Cost Analysis
-Midsize Crossover Utility Vehicle".106 This study is a Phase 2 study of the low development
design in the 2010 Lotus Engineering study "An Assessment of Mass Reduction
Opportunities for a 2017-2020 Model Year Vehicle Program"107, herein described as "Phase
1".
Results for the EPA Phase 2 study of the 1710kg 2010 Toyota Venza include an 18%
mass reduction (with powertrain), 312kg, at -$0.43/kg cost (cost savings), including tooling.
While the results for $/kg appear similar between the Phase 1 Lotus study (without
powertrain, 19% mass reduction, 246kg, at -$0.44/kg), it should be noted that each study took
slightly different approaches. The Phase 1 study included mass reduction of every system
except the powertrain. The EPA Phase 2 study focused on the vehicle as a whole (including
all systems), but also included the powertrain.
LOTUS PHASE 1 STUDY: 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% MR and 2017 production with technology readiness of
2014) and high development (40% MR 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 original
powertrain was changed to a hybrid configuration. The Phase 1 project achieved 19%
(without the powertrain) at 99% of original cost at full phase-in after peer review comments
taken into consideration.mmm This was calculated to be -$0.45/kg utilizing information from
Lotus.
The 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
10R
scope of the Phase 1 study) as noted by the peer reviewer recommendations. The study was
peer reviewed. Mr. Sujit Das, of ORNL and an author of several reports on mass reduction,
reported that the mass reduction opportunities were reasonable and likely to meet the stated
objectives. Mr. Das also recommended using a consistent cost methodology. Dr. Mai en, a
professor at the University of Michigan, reported the mass reduction opportunities were
mmm 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.
3-242
-------
Technologies Considered in the Agencies' Analysis
reasonable and likely to meet the stated objectives and also recommended a data driven
methodology that can be examined at each step of the analysis.109
OBJECTIVES OF EPA PHASE 2 STUDY: The study works to maximize the amount
of mass reduction with technologies and techniques that are considered feasible in
manufacturability and cost effective for a MY 2017 high volume production vehicle. The
EPA Phase 2 study includes the creation of several CAE body in white (BIW) models which
could be used to analyze body stiffness, NVH modal characteristics (overall torsion mode,
overall lateral bending mode, rear end match boxing mode and overall vertical bending rear
end mode in addition to overall and bending and torsional stiffness) and crash (FMVSS and
NCAP) performance. The study also includes a rigorous cost analysis including tooling and
piece cost. The in-depth cost analysis utilizes several cost models including the one described
in the NHTSA project above. In addition, EPA expanded the scope of the work to include an
updated look (2012) at all of the mass reduction technologies and techniques so that FEV was
not limited to only the ideas originally generated by Lotus which were determined in 2009.
As part of this EPA Phase 2 study, FEV/EDAG analyzed the BIW ideas from Lotus's Phase 1
study through CAE modeling and FEV included the technologies for mass reduction with the
information provided in the Phase 1 Lotus Engineering report for the low development
scenario.
VERIFICATION OF THE LOTUS BIW DESIGN FOR NVH: Similar to Lotus Phase
1 study, the EPA Phase 2 study begins with vehicle tear down and BOM development. FEV
and its subcontractors tore down a MY 2010 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 (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 HEEDS 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% 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%
mass reduction from the body structure. All other systems 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
3-243
-------
Technologies Considered in the Agencies' Analysis
cost effective technologies and techniques that are considered feasible and manufacturable in
high volumes in MY2017. FEV included a full discussion of the chosen mass reduction
options for each component and subsystem.
UPDATE RESEARCH ON MASS REDUCTION TECHNOLOGIES: FEV and
Munro created a BOM based on the teardown analysis. Mass reduction technology review
was conducted at the system and sub-system level. The staff at FEV and Munro consists of
experts from the automotive industry and discussion also included outside venders of mass
reduction technologies. Forty of the 150 Lotus Phase 1 concepts were included in the final
mass reduction technology selection. SAFETY FEASIBILITY: 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 (which had been compared qualitatively to the available actual NHTSA
crash results of the Venza). Potential compliance with safety and performance of the light
weighted CAE model in FMVSS and NCAP tests was inferred using quantitative
measurements of vehicle delta velocity and intrusion. 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 BRV 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.
COST ANALYSIS: The development of a bill of materials (BOM), on systems and
sub-systems by FEV and Munro, was the basis for the cost analysis. This methodology is
consistent with the peer reviewed approach described earlier in this chapter. The cost for the
mass reduced technologies were developed by determining the difference in cost for those
new components compared to the old, and under the assumption of production scales of
200,000 units (appropriate for the Venza global production). FEV and Munro developed
several thousand cost spreadsheets as the basis for the cost analyses for the mass reduction
technologies and the BIW and closures. Costs include manufacturing (material, labor,
burden) and markup (end item scrap, Sales, General and Administrative (SG&A), Profit,
Engineering, Development and Testing (ED&T) and Research and Development (R&D)). A
separate tooling cost analysis was also performed and at 18% mass reduction calculated a
$0.05/kg for tooling. The cost analysis of the BIW and closures were done by EDAG and
were based on a Technical Cost Modeling (TCM) approach developed by the Massachusetts
Institute of Technology (MIT) Materials Systems Laboratory's research110.
3-244
-------
Technologies Considered in the Agencies' Analysis
RESULTS: The light-weighting effort achieved an 18% mass reduction (with
downsized powertrainnnn) on the base 1710 kg Toyota Venza at a cost of $-0.43/kg (a cost
savings) which includes tooling (cost increase of $0.04/kg). A cost curve was developed to
show the estimated $/kg over a variety of mass reduction levels utilizing the subset of
technologies and techniques developed throughout the study (see Figure 3-36). The two
curves represent non-compounded mass reduction technologies ("primary") and compounded
mass reduction scenario (a total of "primary" and "secondary"). These curves were
determined by reviewing the BOM part by part and identifying the parts within systems that
would benefit from mass reduction and be able to utilize mass compounding. It is important
to note that the potential for secondary mass reduction was evaluated at many points along the
whole cost curve. The cost curve was used to determine a value for the average cost per
kilogram of cumulative mass reduction (in terms of $/kg for mass reduction at a specific mass
reduction level).
Vehicle Level Cost Curve
(Updated 07/25/2012)
25% -*-w/ Compounding
• w/o Compounding
)l( Optimized Vehicle Solution
(-50.47/kg, IB.26%)
% Vehicle Mass-Reduction
Figure 3-36 Cost Curve for the 2010 Toyota Venza - EPA Study (FEV/EDAG/Munro)
PEER REVIEW: The peer review comments for this study were generally positive
and concurred with the ideas and methodology of the EPA study. The documents for the peer
review can be found in EPA docket EPA-HQ-OAR-2010-0799. After accounting for peer
111111 The engine was downsized and downweighted, however the number of cylinders remained the same and it
remained naturally aspirated.
3-245
-------
Technologies Considered in the Agencies' Analysis
review comments to the draft report, mass reduction decreased by 0.5% and though some of
the adjustments resulted in a cost savings, the overall cost increased slightly. Changes to the
BIW CAE models resulted in minimal differences.
There were many positive comments about the report. While the report included mass
reduction and cost analyses for several hundred items, there were some concerns identified in
the peer review comments that influenced the overall amount of mass reduction and the cost.
These included 1) engine magnesium block cost, 2) the (brake) rotor design, 3) aluminum
hollow suspension stabilizer bar, and 4) the closure aluminum material cost.
There were several areas where peer reviewers suggested changes that did not impact percent
mass reduction or cost. First, more information was included to better describe the wheel
mass technology. Second the BIW models were updated to eliminate the inconsistencies in
material assignments - revising the number of through thickness integration points for the
shell elements and correcting the asymmetrical thickness assignments. Finally, the baseline
and optimized BIW models were further refined to include definitions of welding properties,
transverse shear scale factor, element type, element formulation and material failure criteria.
Based on these updates the crash models were rerun (resulting in statistically insignificant
change and the results included in the final report.
California Air Resources Board Sponsored Mass Reduction 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 findings of the safety validation work, Lotus made revisions to strengthen the vehicle
structure through the use of a more aluminum-intensive BIW (and 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 greater than 250 to less than 170. 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
baseline vehicle and comparable to more performance-oriented models. The analysis
included validation of the lightweight vehicle design for standard FMVSS/IIHS front, side,
rear, offset, roof, intrusion, and seatbelt safety tests. Crash tests simulated in CAE showed
results that were 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 210
seatbelt anchorages is 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 214 side impact
3-246
-------
Technologies Considered in the Agencies' Analysis
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 115mm 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%, which is
less than the 10% strain allowed in the test. The pressure change in the fuel tank is less than
2% 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 cost analysis for the Phase 2 lightweight design involved new piece, tooling, and
assembly work on the BIW and closures, and the technologies and costs for the non-BIW
components were carried over from the Phase 1 work. The Lotus design achieved a 37% (141
kg) mass reduction in the body structure, a 38% (484kg) mass reduction in the vehicle
excluding the powertrain, and a 32% (537 kg) mass reduction in the entire vehicle including
the powertrain. The Phase 2 report included an investigation into the manufacturing and
assembly processes to assess whether the low mass aluminum BIW design can feasibly and
cost-effectively be constructed for 60,000 units. Lotus found that the assembly and tooling
cost savings, due to the lower number of BIW parts, relative to the base Venza partially offset
the 60% increase in piece costs for the BIW for a resulting BIW cost increase of $239.
Accounting for all of the other systems (excluding the powertrain) using the results from
Phase I study, the impact is a cost savings of $476 for 484 kg reduced, or -$0.98/kg. For the
complete vehicle with powertrain (hybrid powertrain), the overall cost savings for the whole
vehicle including powertrain is $318 for 537 kg reduced, or -$0.59/kg. The hybrid engine
was downsized from 120hp to lOOhp and the corresponding hybrid system related
components were removed or exchanged for a minimal change in overall mass. The report
was peer reviewed by a cross section of experts, from academia, a DOE lab, DOE and an
aluminum industry representative. The peer review comments were addressed in the peer
review document and were incorporated in the final Phase 2 report. The documents will be
found on EPA's website
http://www.epa.gov/otaq/climate/publications.htm#vehicletechnologies.
NHTSA Fleet Simulation Study
NHTSA has contracted with GWU to build a fleet simulation model to study the
impact and relationship of light-weighted vehicle design with injuries and fatalities. This
study will also include an evaluation of potential countermeasures to reduce any safety
concerns associated with lightweight vehicles in the second phase. NHTSA has included
three light-weighted vehicle designs in this study: the one from Electricore/EDAG/GWU
mentioned above, one from Lotus Engineering funded by California Air Resource Board for
the second phase of the study, evaluating mass reduction levels around 35 percent of total
vehicle mass, and one funded by EPA and the International Council on Clean Transportation
(ICCT). In addition to the lightweight vehicle models, these projects also created CAE
3-247
-------
Technologies Considered in the Agencies' Analysis
models of the baseline vehicles. To estimate the fleet safety implications of light-weighting,
CAE crash simulation modeling was conducted to generate crash pulse and intrusion data for
the baseline and three light-weighted vehicles when they crash with objects (barriers and
poles) and with four other vehicle models (Chevy Silverado, Ford Taurus, Toyota Yaris and
Ford Explorer) that represent a range of current vehicles. The simulated acceleration and
intrusion data were used as inputs to MADYMO occupant models to estimate driver injury.
The crashes were conducted at a range of speeds and the occupant injury risks were combined
based on the frequency of the crash occurring in real world data. The change in driver injury
risk between the baseline and light-weighed vehicles will provide insight into the safety
performance these light-weighting design concepts. This is a large and ambitious project
involves several stages over several years. NHTSA and GWU have completed the first stage
of this study. The frontal crash simulation part of the study is being finished and will be peer
reviewed. The report for this study will be available in NHTSA-2010-0131. Information for
this study can also be found at NHTSA's website000.
The countermeasures section of the study is expected to be finished in early 2013. This
phase of the study is expected to provide information about the relationship of light-weighted
vehicle design with injuries and fatalities and to provide the capability to evaluate the
potential countermeasures to safety concerns associated with light-weighted vehicles. NHTSA
plans to include the following items in future phases of the study to help better understanding
the impact of mass reduction on safety.
• Simulation of crashes between two light-weighted concept vehicles;
• Additional crash configurations, such as side impact, oblique and rear impact
tests;
• Risk analysis for elderly and vulnerable occupants;
• Safety of light-weighted concept vehicles for different size occupants.
• Partner vehicle protection in crashes with other light-weighted concept
vehicles;
While this study is expected to provide information about the relationship of light-weighted
vehicle design with injuries and fatalities and to provide meaningful information to NHTSA
on potential countermeasures to reduce any safety concerns associated with lightweight
vehicles, because this study cannot incorporate all of the variations in vehicle crashes that
occur in the real world, it is expected to provide trend information on the effect of potential
future designs on highway safety, but is not expected to provide information that can be used
to modify the coefficients derived by Kahane that relate mass reduction to highway crash
fatalities. Because the coefficients from the Kahane study are used in the agencies'
assessment of the amount of mass reduction that may be implemented with a neutral effect on
highway safety, the fact that the fleet simulation modeling study is not complete does not
affect the agencies' assessment of the amount of mass reduction that may be implemented
with a neutral effect on safety.
1 Website for fleet study can be found at http://www.nhtsa.gov/fuel-economy.
3-248
-------
Technologies Considered in the Agencies' Analysis
Safety considerations in establishing CAFE/GHG standards along with discussion of
NHTSA's February 25, 2011, mass-size-safety workshop at DOT headquarters, can be found
in Section II.G of the preamble for this final rule. NHTSA intends to host additional
workshops when the studies have reached a sufficient level of completion, to share the results
with the public and continue the fruitful ongoing public dialogue on these issues.
3.5 How did the agencies consider real-world limits when defining the rate at which
technologies can be deployed?
3.5.1 Refresh and redesign schedules
During MYs 2017-2025 manufacturers are expected to go through the normal
automotive business cycle of redesigning and upgrading their light-duty vehicle products, and
in some cases introducing entirely new vehicles not in the market today. The MY 2017-2025
standards timeframe allows manufacturers the time needed to incorporate GHG reduction and
fuel-saving technologies into their normal business cycle while considering the requirements
of the MY 2012-2016 standards. This is important because it has the potential to avoid the
much higher costs that could occur if manufacturers need to add or change technology at
times other than their scheduled vehicle redesigns. This time period also provides
manufacturers the opportunity to plan for compliance using a multi-year time frame, again
consistent with normal business practice. Over these 9 model years, and the 5 prior model
years that make up the 2012-2016 standards, there will be an opportunity for manufacturers to
evaluate, presumably, every one of their vehicle platforms and models and add technology in
a cost effective way to control GHG emissions and improve fuel economy. This includes all
the technologies considered here and the redesign of the air conditioner systems in ways that
will further reduce GHG emissions and improve fuel economy.
Because of the complexities of the automobile manufacturing process, manufacturers
are generally only able to add new technologies to vehicles on a specific schedule; just
because a technology exists in the marketplace or is made available, does not mean that it can
be immediately applied on all of a manufacturer's vehicles. In the automobile industry there
are two terms that describe when technology changes to vehicles occur: redesign and refresh
(i.e., freshening). Vehicle redesign usually refers to significant changes to a vehicle's
appearance, shape, dimensions, and powertrain. Redesign is traditionally associated with the
introduction of "new" vehicles into the market, often characterized as the "next generation" of
a vehicle, or a new platform. Across the industry, redesign of models generally takes place
about every 5 years. However, while 5 years is a typical design period, there are many
instances where redesign cycles can be longer or shorter. For example, it has generally been
the case that pickup trucks and full size vans have longer redesign cycles (e.g., 6 to 7 years),
while high-volume cars have shorter redesign cycles in order to remain competitive in the
market. There are many other factors that can also affect redesign such as availability of
capital and engineering resources and the extent of platform and component sharing between
models, or even manufacturers.
Vehicle refresh usually refers to less extensive vehicle modifications, such as minor
changes to a vehicle's appearance, a moderate upgrade to a powertrain system, or small
changes to the vehicle's feature or safety equipment content. Refresh is traditionally
3-249
-------
Technologies Considered in the Agencies' Analysis
associated with mid-cycle cosmetic changes to a vehicle, within its current generation, to
make it appear "fresh." Vehicle refresh generally occurs no earlier than two years after a
vehicle redesign or at least two years before a scheduled redesign. For the majority of
technologies discussed today, manufacturers will only be able to apply them at a refresh or
redesign, because their application would be significant enough to involve some level of
engineering, testing, and calibration work.
Most vehicles would likely undergo two redesigns during the 2017-2025 period. Even
with the potential of multiple refresh and redesign cycles, it is still likely that some of the
more advanced and costly technologies (such as cooled boosted EGR engines, or advanced
(P)HEVs) may not be able to be fully implemented within the timeframe of this rule. These
limitations are captured in "phase-in caps," discussed in the next section, and "maximum
technology penetration rates" within the modeling analysis.
The broad technology classes evaluated for purposes of this analysis are defined below
and a brief discussion of the limiting factors considered are presented.
• Conventional Spark Ignition (SI) - This technology category includes all
technologies, such as gasoline direct injection engines, cylinder deactivation, six
and eight speed automatic and dual clutch transmissions, and start-stop micro-
hybrid technology that are not contained in other categories. Many of these
technologies were anticipated as being available in the MYs 2012-2016 time frame
in the recent NHTSA and EPA final rule, and it is expected manufacturers could
expand production to all models by model year 2025. Conventional SI also
includes turbocharged and downsized engines and turbocharged and downsized
engines that include cooled EGR with additional levels of boost and a larger
degree of engine downsizing than seen in the current light-duty gasoline fleet.
These latter technologies are similar to the technologies that many OEMs indicated
were under development and which they anticipate will be introduced into the
market in the 2017-2025 time frame.
• Hybrid - While the agencies recognize there are many types of full-hybrids either
in production or under development, for the purposes of this analysis we have
specifically modeled the P2 type hybrid, as explained in section 3.4.3.6.3. While
the agencies expect the proliferation of these vehicles to increase in this timeframe,
the maximum technology penetration rate and phase-in caps are set at less than
100% in MY 2025 due to industry-wide engineering and capacity constraints for
converting the entire new vehicle fleet to strong hybrids (like P2 and others) in this
time frame. As described above, these technologies (along with PHEVs and EVs)
require a significant cost and complexity, and thus are not expected to be able to
be fully phased into the 2017-2025 fleet like other more conventional (but
advanced) engines.
• Plug-in Hybrid (PHEV) - In the agencies' analyses, this technology includes
PHEVs with a range of 20, 30 and 40 miles. The maximum technology
penetration rates and phase-in caps are set at less than 100% in MY 2025 due to
the same general potential constraints as listed for the HEVs, but are lower for
3-250
-------
Technologies Considered in the Agencies' Analysis
PHEVs due to the current status of the development of these advanced vehicles
and the higher cost relative to HEVs. In addition, some consumers may have
limited or no access to charging infrastructure and, for those consumers, the PHEV
offers little benefit over an HEV at a higher cost. Further, we project (based on
what we know today) that PHEV technology is not available to some vehicle
types, such as large pickups, which are expected to provide a true "work" function.
While it is technically possible to electrify such vehicles, there are tradeoffs in
terms of cost, electric range, and utility (e.g., loss of towing and/or payload
capacity) that may limit the appeal of the vehicle to a narrower market. However,
the agencies are interested in promoting innovation to overcome these potential
obstacles and are thus incentivizing more HEV and PHEV pickup trucks with
credit flexibilities as described in the preamble for this final rule.
• Electric Vehicle (EV) - In our analyses, this technology includes vehicles with
actual on-road ranges of 75, 100, and 150 miles. The actual on-road range was
calculated using a projected 30% gap between two-cycle and on-road range.
These vehicles are powered solely by electricity and are not powered by any liquid
fuels. The maximum technology penetration rates and phase-in caps are set at less
than 100% in MY 2025 due to the same general potential constraints as discussed
for PHEVs. EVs have additional constraints due to limited infrastructure and
range as well. Further, as with PHEVs, we assume that EV technology is not
available to some vehicle types, such as large pickups. While it is possible to
electrify such vehicles, there are tradeoffs in terms of cost, range, and utility that
would limit the appeal of the vehicle to a narrower market. These trade-offs are
expected to reduce the market for other vehicle types as well, and for this analysis
we have considered this in the development of the maximum technology
penetration rates.
• Mass Reduction - This technology includes changing the design to use less
material or substituting lighter materials for heavier materials. Mass reduction
compounding after significant primary mass reduction is achieved can also make
significant contribution to the overall vehicle mass reduction. NHTSA and EPA
have conducted a thorough assessment of the levels of mass reduction that could
be achieved which is both technologically feasible and which can be implemented
in a safe manner for this joint federal FRM (as described earlier in this
Chapter).Vehicle phase-in caps
GHG-reducing and fuel-saving technologies for vehicle applications vary widely in
function, cost, effectiveness and availability. Some of these attributes, like cost and
availability vary from year to year. New technologies often take several years to become
available across the entire market. The agencies use phase-in caps to manage the maximum
rate that the CAFE and OMEGA models can apply new technologies.
Phase-in caps are intended to function as a proxy for a number of real-world
limitations in deploying new technologies in the auto industry. These limitations can include
but are not limited to, engineering resources at the OEM or supplier level, restrictions on
intellectual property that limit deployment, and/or limitations in material or component supply
3-251
-------
Technologies Considered in the Agencies' Analysis
as a market for a new technology develops. Without phase-in caps, the models may apply
technologies at rates that are not representative of what the industry is actually capable of
producing, which would suggest that more stringent standards might be feasible than actually
would be.
EPA applies the caps on an OEM vehicle platform basis for most technologies. For a
given technology with a cap of x%, this means that x% of a vehicle platform can receive that
technology. On a fleet average basis, since all vehicle platforms can receive x% of this
technology, x% of a manufacturer's fleet can also receive that technology. EVs and PHEVs
are an exception to this rule as the agencies limit the availability of these technologies to some
subclasses. Unlike other technologies, in order to maintain utility, EPA only allows non-
towing vehicle types to be electrified in the OMEGA model. As a result, the PHEV and EV
cap was applied so that the average manufacturer could produce to the cap levels. As would
be expected, manufacturers that make more non-towing vehicles can have a higher fraction of
their fleet converted to EVs and PHEVs, while those that make fewer non-towing vehicles
have a lower potential maximum limit on EV and PHEV production.
NHTSA applies phase-in caps in addition to refresh/redesign cycles used in the CAFE
model, which constrain the rate of technology application at the vehicle level so as to ensure a
period of stability following any modeled technology applications. Unlike vehicle-level cycle
settings, phase-in caps, defined on a percent per year basis, constrain technology application
at the OEM level. As discussed above phase-in caps are intended to reflect a manufacturer's
overall resource capacity available for implementing new technologies (such as engineering
and development personnel and financial resources) thereby ensuring that resource capacity is
accounted for in the modeling process. At a high level, phase-in caps and refresh/redesign
cycles work in conjunction with one another to avoid the CAFE modeling process out-pacing
an OEM's limited pool of available resources during the rulemaking time frame, especially in
years where many models may be scheduled for refresh or redesign. This helps to ensure
technological feasibility and economic practicability in determining the stringency of the
standards.
Phase-in caps do not necessarily define market penetration rates and they do not
necessarily define the rate at which a particular technology will be applied, rather they simply
present an upper limit, or ceiling at which the agencies' computer models (OMEGA and
CAFE) can apply new technologies to vehicles to raise their fuel economy and reduce their
CO2 emissions. Ultimately, phase-in caps are determined by the agencies using engineering
judgment. However, there are several sources of information on technology penetration that
the agencies consider in assigning phase-in caps to various technologies:
• Confidential OEM submissions indicate the rate at which an individual
manufacturer can deploy a particular technology. Manufacturer information is
especially helpful if multiple manufacturers indicate similar technology
penetration rates. The agencies consider these CBI submissions along with
other sources of information.
• Historical data from EPA's annual Carbon Dioxide Emissions and Fuel
Economy Trends Report111 database are used to inform the agencies about
typical historical rates of adoption of technologies. Relevant data include both
3-252
-------
Technologies Considered in the Agencies' Analysis
the industry-wide technology penetration data that are included in the annual
Trends report, as well as individual manufacturer-specific technology
penetration data that have not been published in the Trends report, but which
are presented below.
• Trade press articles, company publications, press releases, and other reports
often discuss new technologies, how quickly they will be deployed and
manufacturing strategies that enable faster penetration rates. These articles
provide a useful glimpse into how manufacturers are changing in order to
become more competitive.
• The relative complexity of a technology as well as the availability from
suppliers. Some technologies can be implemented rather easily—like tires.
Other technologies are much more sophisticated—like hybridization.
3.5.1.1 Trends Report and Industry Data
For over 30 years, EPA's Fuel Economy Trends report has tracked the fuel economy
of light duty vehicles and the technology used by automakers to improve fuel economy. A
particularly interesting aspect of the Trends data is how technology is adopted by the industry
and how this changes over time. Trends data shows that industry-wide, it has typically taken
up to 15-20 years for a technology to penetrate the entire fleet. Some technologies such as
port fuel injection and variable valve timing start slowly and then rapidly progress. Others,
like torque converter lockup and front wheel drive penetrate rapidly after their first
appearance on the market. Figure 3-37, below shows these trends.
e
CO
100%
80%
60%
40%
20%
0%
PortF
IX
/ —_S_
-•Multi-Valve
•^
Front Wheel Drive_ . - ~Lss^\r
* ^1^ ***
Lockup /'„-'"' / / variable Valve Timing
"' / /
J'
CVT
5 10 15 20 25 30 35
Years After First Use
112
Figure 3-37 Technology Penetration After First Significant Use
There are several cases where technologies have penetrated the fleet rapidly,
sometimes beginning with significant market penetration, sometimes beginning with
3-253
-------
Technologies Considered in the Agencies' Analysis
relatively small market penetration. For example, six speed automatic transmissions were in
7% of the industry-wide fleet in 2006 and by 2010, they were in 36% of the fleet, for an
increase of 29% in 4 yearsppp. Port fuel injection went from about 12% of the fleet in 1984 to
88% in 1994. Front wheel drive, a technology that requires a complete change in vehicle
architecture, increased from 9% in 1979 to 60% by 1988
113
Recent academic literature has also used deployment rate data from the EPA Fuel
Economy Trends Report, Wards Factory Installed data, and other sources to report to describe
historical deployment rates of a variety of technologies (DeCicco, 2010 and Zoepf, 2011).
DeCicco, for example, cites conversion to fuel injection and front wheel drive in passenger
cars as having seen maximum growth in adoption of 17% and 11% per year respectively.114
Zoepf examines a broader array of automotive technologies and notes a span of maximum
growth rates in passenger cars from 4% to nearly 24% per year with variance based on feature
type.115
While these examples show that the industry is capable of adopting certain new
technologies rapidly industry-wide, considering the rate of introduction of technology by
individual OEMs shows that the pace of technology introduction can in some cases be even
faster. Table 3-129 below shows how individual manufacturers can apply technologies rapidly
to a large fraction of their fleet. Although not typical for most manufacturers and
technologies, the data below shows that manufacturers have chosen to deploy some
technologies very rapidly.
Table 3-129: Historical Phase-In Rates of Selected Technologies
Manufacturer
General Motors
Ford
Honda
Chrysler
Toyota-cars only
Nissan-cars only
Toyota-cars only
Ford
Nissan
Volkswagen
Hyundai
General Motors
General Motors
Technology
Lockup Transmission
Fuel Injection
Fuel Injection
Fuel Injection
Multi-Valve
Multi-Valve
Variable Valve Timing
Multi-Valve
Continuously Variable
Transmission
Gasoline Direct Injection
Variable Valve Timing
Variable Valve Timing
Gasoline Direct Injection
Technology Market Share Increase
1980-1982: 83% in 3 years
1983-1987: 91% in 5 years
1986-1990: 91% in 5 years
1988: 37% in 1 year
1987-1989: 85% in 3 years
1989-1990: 71% in 2 years
2000-2003: 87% in 4 years
2004-2005: 36% in 2 years
2007: 45% in 1 year
2008: 52% in 1 year
2009: 48% in 1 year
2006-2010: 75% in 5 years
2010: 27% in 1 year
Often, a rapid application of technology is helped by having similar vehicle
architecture, or by sharing major components such as engines or transmissions across multiple
products. As discussed below, platform sharing combined with improvements in platform
ppp EPA staff calculated the penetration rate of 6-speed automatic transmissions from 2010 Trends data.
Aggregated source data can be seen on page 54 of the 2010 Fuel Economy Trends Report.
3-254
-------
Technologies Considered in the Agencies' Analysis
and manufacturing flexibility is expected to further enable faster implementation of new
technologies.
3.5.1.2 The rate of technology adoption is increasing
The agencies recognize that new technologies may not achieve rapid deployment
immediately and that small-scale production is a part of the technology learning process. To
this end the phase-in caps distinguish between technologies that have been successfully
applied in existing vehicles and those that under development but are anticipated on
production vehicles in the near future.
The rate of technology adoption appears to be increasing as manufacturers increase
model turnover and decrease the numbers of unique vehicle platforms. This facilitates a
steady stream of new products, increased sales and optimized vehicle redesigns allowing and
fuel consumption-reducing technologies to be applied to as many vehicles as possible. In
today's globally competitive market, and certainly for the U.S., market share and
competitiveness is strongly influenced by a manufacturer's ability to turn over their product
line-up. Merrill Lynch's Car Wars Report116 shows that replacement rate is speeding up and
showroom age is dropping as manufacturers are striving to be more competitive in the market.
Increased model turn-over creates more opportunity for manufacturers to deploy new
technologies faster than in the past.
Zoepf, cited above, reports that the developmental time, from first production
application to maximum growth rate, has been declining exponentially as manufacturers bring
innovations to market progressively faster. Ellison et al. (1995)117 indicate that U.S. and
European automakers reduced overall product development time by more than a year in the
1990s. Ellison et al. point to the increased role that suppliers have had in product
development process during the same time, potentially commoditizing innovations more
quickly.
Vehicle platforms are the basic underpinnings of vehicles and are often shared across
several unique products. By reducing the number of platforms, and making these platforms
flexible, manufacturers can better deploy resources to serve a wider market with more
products. Utilizing a modern, flexible platform architecture, a manufacturer can produce a
sedan, wagon, minivan, and a crossover, or SUV on a single platform and all of these
products can be assembled in a single vehicle assembly plant. Basic components can be
developed and purchased at high volumes, while enabling the manufacturer to exploit what
would otherwise be niche markets. This commonization of platforms does have the potential
to increase the mass for lighter vehicle models within the platform because the platform needs
to be designed for the more severe duty cycle of the SUV and/or larger engine. Volkswagen
has recently launched a new platform called MQB, which will be used world-wide by up to 60
unique models from VW, Audi, Seat, and Skoda. This structure will replace 18 "engine
mounting architectures" with just two.
It gives us the possibility to produce models from different segments and in
varying sizes using the same basic front-end architecture, " .... "We can go from a
typical hatchback to a saloon, cabriolet and SUV with only detailed changes to the
3-255
-------
Technologies Considered in the Agencies' Analysis
size of the wheel carriers. " ... it will be used on every model from the new Lupo all the
way through to the next-generation Sharon.118
One of the key enablers of this drive to reduce platforms and increase model turn-over
is increased manufacturing flexibility.119 For example, in 2004, Ford invested in flexible
manufacturing technology for their Cleveland No. 1 engine plant. Although the plant was
shut down for two years after this investment, Ford was able to retool and reopen the plant at
a low cost to produce their new 3.5L EcoBoost turbocharged, direct injection engine as well
as their 3.7L ^
further stated,
190 191
as their 3.7L V6. In their December, 2008 business plan submitted to Congress, Ford
...nearly all of our U.S. assembly plants will have flexible body shops by 2012
to enable quick response to changing consumer demands and nearly half of our
transmission and engine plants will be flexible, capable of manufacturing various
combinations of transmission and engine families.
Like VW, Ford is also striving to reduce their platforms and complexity. In Ford's
2008 business plan submitted to Congress, they stated that in addition to divesting themselves
from certain luxury brands like Jaguar, Land Rover, Volvo, and Aston Martin, they were
working to consolidate their vehicle platforms from 25 in 2005 to 9 by 2012. Having more
vehicles per platform frees up resources to deploy new technologies across a greater number
of vehicles more quickly and increases the rate at which new technologies can be introduced.
We believe GM's recent restructuring will also enable faster vehicle redesigns and more rapid
penetration rates in the 2010-plus time frame compared to the 1990s and 2000s. In the past
seven years, GM has eliminated five brands (Saturn, Hummer, Saab, Pontiac, and
Oldsmobile), significantly reducing the number of unique products and platforms the
company needed to devote engineering resources to. GM has set a goal to halve its number of
1 99
vehicle platforms by 2018 and boost manufacturing efficiency by 40%.
3.5.1.3 Phase-in Rates Used in the Analysis
Table 3-130 below shows phase-in rates for the technologies used in the OMEGA
model. OMEGA calculations are based on five year intervals, so phase-in caps are derived
for model years 2016, 2021 and 2025. Table 3-131 shows phase-in rates for the technologies
used in the CAFE model. The CAFE model calculations are annual, so phase in rates are
derived for every year of the program. Where possible, phase-in rates for OMEGA and
CAFE were harmonized, but there are some differences mainly where technologies differ
between the agencies.
Most technologies are available at a rate of either 85% or 100% beginning in 2016.
Some advanced technologies expected to enter the market in the near future such as EGR
Boost follow a 3% annual cap increase from 2016 to 2021, then, approximately 10% from
2021 to 2025. Diesels follow an annual 3% increase in phase-in cap through 2025. Hybrids
follow a 3% annual increase from 2016 to 2012, then 5% from 2021 to 2015. PHEVs and EVs
follow a 1% annual cap increase.
3-256
-------
Technologies Considered in the Agencies' Analysis
Lower phase-in caps for Alternate Fueled Vehicles (AFVs) reflect additional
investment in infrastructure that is required to achieve high levels of conversion to a new fuel
type. These limited phase-in caps also reflect as yet unknown consumer responses to HEVs,
PHEVs and BEVs.
Table 3-130 Phase-In Caps used in the OMEGA model
Technology
Low Friction Lubricants
Engine Friction Reduction - level 1
Early Torque Converter lockup
Aggressive Shift Logic - Level 1
Improved Accessories - Level 1
Low Rolling Resistance Tires - Level 1
Low Drag Brakes
WT - Intake Cam Phasing
WT - Coupled Cam Phasing
WT - Dual Cam Phasing
Cylinder Deactivation
Variable Valve Lift - Discrete
Variable Valve Lift - Continuous
Conversion to DOHC
Stoichiometric Gasoline Direct Injection (GDI)
Turbocharging (18 bar BMEP) and Downsizing
Continuously Variable Transmission
6-speed Automatic Transmission
6-speed Dual Clutch Transmission - dry & wet clutch
Electric & Electric/Hydraulic Power Steering
12V Stop-Start
Secondary Axle Disconnect
Aero Drag Reduction - Level 1
Aggressive Shift logic - Level 2 (Shift Optimizer)
8-speed Automatic Transmission
8-speed Dual Clutch Transmission - dry & wet clutch
Improved Accessories - Level 2
Aero Drag Reduction - Level 2
Low Rolling Resistance Tires - Level 2
Engine Friction Reduction - level 2 (inc. low friction lubes - level 2)
High Effiency Gearbox
Turbocharging (24 bar BMEP) and Downsizing
Cooled EGR
P2 Hybrid Electric Vehicle (HEY) and Mild Hybrid (MHEV)
Turbocharging (27 bar BMEP) and Downsizing
2016
100%
100%
100%
100%
100%
100%
100%
85%
85%
85%
85%
85%
85%
85%
85%
85%
85%
85%
85%
85%
85%
85%
85%
0%
30%
30%
30%
30%
0%
0%
0%
15%
15%
15%
0%
2021
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
80%
80%
80%
80%
75%
60%
60%
30%
30%
30%
15%
2025
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
75%
75%
50%
50%
3-257
-------
Technologies Considered in the Agencies' Analysis
Conversion to Advanced Diesel
Full Electric Vehicle (EV)
Plug-in HEV
15%
6%
5%
30%
11%
10%
42%
15%
14%
Table 3-131 Phase-In Caps used in the CAFE Model
Technology
Abbr.
MY
2009
MY
2010
Low Friction Lubricants -Level 1 LUB1 30% 40%
Engine Friction Reduction - Level 1 EFR1 30% 40%
Low Friction Lubricants and Engine Friction Reduction - Level 2
Variable Valve Timing (WT) - Coupled CamPhasing (CCP) on SOHC
Discrete Variable Valve Lift pWL) on SOHC
Cylinder Deactivation on SOHC
Variable Valve Timing (WT) - Intake CamPhasing (ICP)
Variable Valve Timing (WT) - Dual CamPhasing pCP)
Discrete Variable Valve Lift pWL) on DOHC
Continuously Variable Valve Lift (CWL)
Cylinder Deactivation on DOHC
Stoichiometric Gasoline Direct Injection (GDI)
Cylinder Deactivation on OHV
Variable Valve Actuation - CCP and DWL on OHV
Stoichiometric Gasoline Direct Injection (GDI) on OHV
Turbocharging and Downsizing -Level 1 (18 bar BMEP)
Turbocharging and Downsizing - Level 2 (24 bar BMEP)
Cooled Exhaust Gas Re circulation (EGR)- Level 1 (24 bar BMEP)
Cooled Exhaust Gas Re circulation (EGR) - Level 2 (27 bar BMEP)
AdvancedDiesel
6-SpeedManual/Improved Internals
High Efficiency Gearbox (Manual)
Improved Auto. Trans. Controls/Externals
6-Speed Trans with Improved Internals (Auto)
6-speed DCT
8-Speed Trans (Auto or DCT)
High Efficiency Gearbox(Auto or DCT)
Shift Optimizer
Electric Power Steering
Improved Accessories -Level 1
Improved Accessories -Level 2
12V Micro-Hybrid (Stop-Start)
Integrated Starter Generator
Strong Hybrid - Level 1
Conversion fromSHEVl to SHEV2
Strong Hybrid - Level 2
Plug -in Hybrid - 30 mi range
Plug-in Hybrid
Electric Vehicle (Early Adopter) - 75 mile range
Electric Vehicle (Broad Market) - 150 mile range
Fuel Cell Vehicle
Mass Reduction - Level 1
Mass Reduction - Level 2
Mass Reduction - Level 3
Mass Reduction - Level 4
Mass Reduction - Level 5
Low Rolling Resistance Tires - Level 1
Low Drag Brakes
Secondary Axle Disconnect
Aero Drag Reduction, Level 1
Aero Drag Reduction, Level 2
LUB2 EFR2
CCPS
DWLS
DEACS
ICP
DCP
DWLD
CWL
DEACD
SGDI
DEACO
WA
SGDIO
TRBDS1 SD
TRBDS2 SD
CEGR1 SD
CEGR2 LD
ADSL LD
6MAN
HETRANSM
IATC
NAUTO
DCT
8SPD
HETRANS
SHFTOPT
EPS
IACC1
IACC2
MHEV
ISO
SHEV1
SHEV1 2
SHEV2
PHEV1
PHEV2
EV1
EV4
FCV
MR1
MR2
MR3
MR4
MR5
ROLL1
ROLL2
LDB
SAX
AERO1
AERO2
0%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
15%
0%
0%
0%
0%
15%
0%
15%
15%
15%
0%
0%
0%
5%
5%
0%
15%
0%
0%
0%
0%
0%
0%
0%
0%
0%
15%
15%
15%
0%
0%
20%
0%
20%
15%
30%
0%
0%
25%
25%
25%
25%
25%
25%
25%
25%
25%
0%
0%
0%
0%
25%
0%
25%
25%
25%
0%
0%
0%
20%
20%
0%
25%
0%
0%
0%
0%
0%
0%
0%
0%
25%
25%
25%
0%
0%
35%
0%
35%
25%
40%
0%
MY
2011
50%
50%
0%
35%
35%
35%
35%
35%
35%
35%
0%
0%
0%
0%
0%
35%
35%
0%
0%
0%
35%
35%
0%
35%
0%
0%
0%
0%
0%
0%
0%
0%
35%
35%
10%
0%
50%
0%
50%
35%
50%
10%
MY
2012
60%
60%
0%
45%
45%
45%
45%
45%
45%
45%
45%
45%
45%
45%
45%
45%
3%
3%
0%
0%
45%
0%
45%
45%
45%
0%
0%
0%
50%
50%
0%
45%
3%
3%
3%
3%
1%
1%
1%
0%
45%
45%
45%
20%
0%
65%
0%
65%
45%
60%
20%
MY
2013
MY
2014
70% | 80%
70% 80%
0% 0%
55% | 65%
55% | 65%
55% | 65%
55% | 65%
55% | 65%
55%
55%
55%
55%
55%
55%
55%
55%
6%
0%
3%
55%
0%
55%
55%
55%
0%
0%
0%
65%
65%
0%
55%
6%
6%
6%
6%
2%
2%
2%
0%
55%
55%
55%
30%
0%
80%
0%
80%
55%
70%
30%
65%
65%
65%
65%
9%
9%
0%
6%
0%
65%
65%
10%
0%
0%
80%
80%
10%
65%
9%
9%
9%
9%
3%
3%
3%
0%
65%
40%
0%
95%
0%
95%
65%
80%
40%
MY
2015
MY
2016
MY
2017
90% 100%! 100%
90% 1 100% 100%
0% 0% 12%
75% 85% 95%
75% 85% 95%
75% 85% 95%
75% | 85% 95%
75% 85% 95%
75% 85% 95%
12% 15% 18%
12%
0%
6%
75%
0%
75%
75%
75%
20%
0%
0%
95%
95%
20%
75%
12%
12%
12%
12%
4%
4%
4%
0%
75%
75%
75%
50%
0%
100%
0%
100%
75%
90%
50%
15%
0%
6%
85%
0%
85%
30%
0%
0%
100%
100%
30%
15%
15%
15%
5%
5%
5%
1%
18%
3%
6%
95%
12%
95%
95%
95%
40%
12%
20%
100%
100%
40%
95%
18%
18%
18%
18%
6%
6%
5%
2%
85% 95%
85% 95%
85% 95%
60% | 70%
10% 20%
100%
0%
100%
85%
100%
60%
100%
15%
100%
95%
MY
2018
100%
100%
24%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
21%
6%
6%
100%
24%
100%
100%
100%
50%
24%
40%
100%
100%
50%
100%
21%
21%
21%
7%
7%
5%
3%
100%
100%
100%
80%
30%
100%
30%
100%
100%
100% 100%
70% 80%
MY
2019
100%
100%
36%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
24%
9%
6%
100%
36%
100%
100%
100%
60%
36%
60%
100%
100%
60%
100%
MY
2«0_
100%
100%
48%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
27%
12%
6%
100%
48%
100%
100%
100%
70%
48%
80%
100%
100%
70%
100%
24% 27%
24% 27%
24%
8%
8%
5%
4%
100%
100%
100%
90%
40%
100%
45%
100%
100%
100%
90%
27%
9%
9%
5%
5%
100%
100%
100%
100%
50%
100%
60%
100%
100%
100%
100%
MY
2021_
100%
100%
60%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
30%
30%
15%
6%
100%
60%
100%
100%
100%
80%
60%
100%
100%
100%
80%
100%
30%
30%
30%
30%
10%
10%
5%
6%
100%
100%
100%
100%
60%
100%
75%
100%
100%
100%
100%
MY MY
2022 J2023
100% 100%
100%| 100%
72% 84%
100% 100%
100% 100%
100%| 100%
100%| 100%
100% 100%
100% 100%
100% 100%
100%
100%
100%
100%
100%
100%
45%
45%
25%
6%
100%
72%
100%
100%
100%
90%
72%
100%
100%
100%
90%
100%
35%
35%
11%
11%
5%
7%
100%
100%
100%
100%
70%
100%
90%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
60%
60%
6%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
40%
40%
40%
40%
12%
12%
5%
8%
100%
100%
100%
100%
80%
100%
100%
100%
100%
100%
100%
MY
2024
100%
100%
96%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
75%
75%
45%
6%
100%
96%
100%
100%
100%
100%
96%
100%
100%
100%
100%
100%
45%
MY
2025
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
75%
50%
6%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
50%
45% | 50%
45% 50%
45% 50%
13% 14%
13% 14%
5%
9%
9%
100%
100%
100%
100%
90%
100%
100%
100%
100%
100%
100%
5%
10%
10%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
3.6 How are the technologies applied in the agencies' respective models?
To estimate potential technology application in response to potential CAFE standards,
and accompanying costs, effects, and benefits of potential CAFE standards, NHTSA uses the
CAFE Compliance and Effects Modeling System, which was developed specifically for that
3-258
-------
Technologies Considered in the Agencies' Analysis
purpose by DOT's Volpe National Transportation Systems Center (Volpe Center). To
estimate potential technology application in response to potential CAFE standards, and
accompanying costs, EPA uses the OMEGA model, which EPA staff developed specifically
for that purpose. The models apply different but related methods to estimate and account for
potential applications of technology. The models and methods are discussed in the agencies'
respective RIAs and preamble sections, and in detail in documentation. The agencies have
each developed modeling system inputs reflecting estimates that have been agreed to and
presented above.
3.7 Maintenance and Repair Costs Associated with New Technologies
In the proposal, we requested comment on maintenance, repair, and other operating-
costs and whether these might increase or decrease with the new technologies (See 76 FR
74925) We received comments on this topic from NADA. These comments stated that the
agencies should include maintenance and repair costs in estimates of total cost of ownership
(i.e., in our payback analyses).qqq NADA proffered their website
(http://www.nadaguides.com/Cars/Cost-to-Own) as a place to find useful information on
operating costs that might be used in our final analyses. This website tool is meant to help
consumers quantify the cost of ownership of a new vehicle. The tool includes estimates for
depreciation, fees, financing, insurance, fuel maintenance, opportunity costs and repairs for
the first five years of ownership. The agencies acknowledge that the tool may be useful for
consumers; however, there is no information provided on how these estimates were
determined. Without documentation of the basis for estimates, the website information is of
limited use in this rulemaking where the agencies document the source and basis for each
factual assertion. Also, the costs do not extend beyond five years, which the agencies require
for purposes of estimating social costs and costs of ownership throughout vehicles' useful
lives. There are also evident substantive anomalies in the website information.1^ For these
reasons, the agencies have performed an independent analysis to quantify maintenance costs.
Here we summarize what we have done for the final rule with respect to maintenance
and repair costs. 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 to occur by auto makers on a regular schedule. Examples of
maintenance are oil and air filter changes, tire replacements, brake pad replacement, etc.
Repair costs are those costs that are unexpected and, as such, occur randomly and uniquely for
every vehicle owner, if at all. Examples of repair would be parts replacement following an
qqq See NADA (EPA-HQ-OAR-2010-0799-9575, p. 10).
1X1 For example, comparing the 2012 Hyundai Sonata showed the same cost for fuel ($11,024) regardless of
whether it is a hybrid option or not. The HEV fuel economy rating is 35/40 mpg City/Highway for the HEV and
2.4L non HEV rating is 24/35. Another example is the 2012 Ford Fusion SEL: the front wheel drive and the all-
wheel drive versions have identical fuel cost despite having different fuel economies.
3-259
-------
Technologies Considered in the Agencies' Analysis
accident, light bulb replacement, turbocharger replacement following a mechanical failure,
etc.
How each agency has folded the costs presented here into their respective final
analyses is presented in each agency's respective preamble sections (section III for EPA,
section IV for NHTSA) and RIAs.
The agencies have also evaluated ownership costs that include financing, sales tax,
and insurance costs, and discuss those costs in TSD 4 and in each agency's respective
preamble sections (section III for EPA, section IV for NHTSA) and RIAs.
3.7.1
Maintenance Costs
To estimate maintenance costs that could reasonably be attributed to these rules, the
agencies have looked 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 are attributing to this rulemaking. 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, we have looked at
specific parts rather than specific vehicle models.
By comparing the manufacturer recommended maintenance schedule of the items
being compared, we were able to estimate the differences in maintenance intervals for the
two. With estimates of the costs per maintenance event, we are able to put together a picture
of the maintenance cost differences associated with the "new" technology.
The technologies considered, maintenance interval comparisons, costs per
maintenance event are shown in Table 3-132.
Table 3-132 Maintenance Interval and Maintenance Cost Differences for 2017-2025 Enabling
Technologies (dollar values in 2010$)a
2017-2025
Technology
Low Rolling
Resistance Tires -
Level 1
Low Rolling
Resistance Tires -
Level 2
Stoichiometric
Gasoline Direct
Injection (GDI)
Turbocharging (18
bar BMEP) and
Downsizing
Reference
Case
Michelin Harmony
Michelin Energy Saver
A/S
20 10 Hyundai Sonata
2.4L
2011F-1505.0L
Control
Case
Michelin Energy
Saver A/S
does not exist
20 11 Hyundai Sonata
2.4L
201 IF- 150 EcoBoost
Mainte
nance Interval
Difference
Identical
Identical
Identical
Identical
Main
tenance
Event Cost
Difference
+$6.44 every
40K miles
+$43.52 every
40K miles
+$0.00
+$0.00
3-260
-------
Technologies Considered in the Agencies' Analysis
6-speed Dual Clutch
Transmission - dry &
wet clutch
Electric &
Electric/Hydraulic
Power Steering
12V Stop-Start
8-speed Automatic
Transmission
Cooled EGR
Conversion to
Advanced Diesel
Hybrid Electric
Vehicle (HEV)
P2HEV
Plug-in HEV
Full Electric Vehicle
(EV) - oil change
EV - air filter change
EV - spark plugs
EV- brake fluid
EV - engine coolant
EV/PHEV0- battery
coolant
EV - battery health
check
2005 VW Jetta Auto (6-
spd '09G')
2009 Ford Fusion
~
2010 BMW 750i6-spd
~
2011 VW Jetta SE 2. 5L
20 12 Ford Escape V6
2012 Hyundai Sonata 14
2012 Toyota Camry V6
2012 Chevy Silverado
5.3L
20 12 Sonata V6
20 12 Chevrolet Craze
20 11 Nissan Versa
20 11 Nissan Versa
20 11 Nissan Versa
20 11 Nissan Versa
20 12 Ford Focus
20 11 Ford Focus
2011 Nissan Versa
2005 VW Jetta DSG
(6-spd '02E'
20 10 Ford Fusion
2013 Volvo V40
2010BMW760Li8-
spd
2013 Volvo V40
2011 VW Jetta TDI
20 12 Ford Escape
Hybrid 2012
Hyundai Sonata
Hybrid 2012
Toyota Camry Hybrid
20 12 Silverado 2-
Mode Hybrid
2013 Sonata Hybrid
20 12 Chevrolet Volt
20 11 Nissan Leaf
20 11 Nissan Leaf
20 11 Nissan Leaf
20 11 Nissan Leaf
20 12 Ford Focus EV
20 11 Ford Focus EV
20 11 Nissan Leaf
Identical
Identical
N/A
Identical
N/A
Identical
Identical
N/A
Identical (for
common service
items)
No interval for
EV
No interval for
EV
No interval for
EV
Identical
No interval for
EV
No interval for
Focus (gasoline)
No interval for
Versa
+$0.00
+$0.00
~
+$0.00
~
+$49.25 every
20K miles
+$0.00
~
+$0.00
-$38.67 every
7. 5K miles
-$28.60 every
30K miles
-$83.00 every
105K miles
+$0.00
-$59.00 every
100Kmilesb
+$117.00
every 150K
milesb
+$38.67 every
15K miles
a All maintenance interval, hours required, and part(s) cost differentials between reference and control cases
were sourced from the ALLDATA subscription database (www.alldatapro.com) in January through February of
2012, unless noted otherwise in the text.
b These are the values the agencies used when conducting analyses. However, as newer information became
available the agencies concluded these revised values (cost and interval) resided within an appropriate range
given the uncertainty in how future systems will be designed. Additional information is available in bulleted text
below.
0 EPA also applied this maintenance cost adjustment to PHEVs; NHTSA did not.
Further comments and details with respect to Table 3-132:
• Low Rolling Resistance Tires - Level 1: Current Uniform Tire Quality Grading
ratings (treadwear, traction, temperature) for "LRR" tires do not give a clear indication
of tire life vs. conventional tires (e.g. 225/50R17 Michelin Harmony = 740 A B and
Michelin Energy Saver A/S = 480 A A; whereas Goodyear Assurance Fuel Max = 580
A A and Goodyear Assurance TripleTred = 540 A A). The $6 value per maintenance
3-261
-------
Technologies Considered in the Agencies' Analysis
event in based on the 2025MY LRRT1 incremental cost presented in Table 3-100 of
this Joint TSD.
• Low Rolling Resistance Tires - Level 2: The $44 value per maintenance event in
based on the 2025MY LRRT2 incremental cost presented in Table 3-100 of this Joint
TSD.
• Stoichiometric GDI: 36,000 mile fuel filter interval for reference case (filter cost =
$25.99), 37,500 mile interval for control case (filter cost = $34.68, or +$8.69 for every
36,000 miles). However, BMW does not require any fuel filter changes for their
turbocharged GDI engines.
• Turbocharging (18 bar BMEP) and Downsizing: Oil change interval, oil type, and
labor hours are identical between the reference and control cases. The reference case
takes more oil - 7.7 quarts vs 6.2 quarts - but they take the exact same type of oil (5W-
20 synthetic blend, Ford specification WSS-M2C9302-A). The control case uses a
larger oil filter with higher filtration efficiency compared to the reference case and the
cost is $13.89 vs. $9.76 or +$4.13 for the control case every 5,000 miles. However,
BMW uses the exact same oil filter and oil specification in naturally aspirated and
turbocharged GDI applications.
• 6 speed Dual Clutch Transmission (dry & wet): Control case requires fluid & filter
change every 40,000 miles; reference case is "fill-for-life". However, the 2012 Ford
Fiesta dual clutch transmission requires fluid and filter replacement at 150,000 miles;
the dual clutch transmission requires 2.2 quarts of fluid and the automatic transmission
requires 6.9 quarts.
• Electric & Electric/Hydraulic Power Steering: No power steering fluid changes are
required in either the reference or control cases - both are "fill-for-life".
• 12 Volt Stop-Start: No information available.
• 8 speed Automatic Transmission: Replace fluid and filter at 150,000 miles for both
reference and control cases.
• Cooled EGR: No information available.
• Conversion to Advanced Diesel (from gasoline): Identical oil change and air filter
maintenance intervals, but different oil capacity (4.8 liters for reference case and 4.3
liters for control case), oil type (VW 502 00 for reference case and VW 507 00 for
control case), and oil filter (06D115562 for reference case @ $14.00 and 071115562C
for control case @ $9.00). According to VW service bulletin, "502" and "507" oil
specs have converged into a single list of approved oils for the North American
market, so oil change cost is assumed to be equal between gasoline and diesel engines.
However, the control case requires a fuel filter change every 20,000 miles: fuel filter
part number 1K0127434A @ $37.25 + 0.4 hrs labor @ $30/hr, or +$49.25 every
20,000 miles.
• Hybrid Electric Vehicle (HEV): Ford Escape in the control case has longer oil change
interval (10,000 vs. 7,500 miles) compared to its reference case. However, Sonata,
Camry, and Silverado reference and control cases have identical engine oil change
3-262
-------
Technologies Considered in the Agencies' Analysis
intervals, oil types, and labor hours, as well as identical transmission fluid change
intervals and fluid types.
• P2 HEV: No information available.
• Plug-in HEV: Oil change interval, oil type, and labor hours are identical for the
reference and control cases.
• Full Electric Vehicle (EV) - Oil Change: Reference case has an oil change interval
every 7,500 miles: the cost is 4.4 quarts of 5W-30 @ $3.50/quart + $8.27 oil filter
(part number 1520865FOC) + 0.5 hours labor @ $30/hr = -$38.67 (i.e., savings) for
control case every 7,500 miles.
• Full Electric Vehicle (EV) - Air Filter Change: Reference case replacement interval is
every 30,000 miles; cost is $19.60 for the part + 0.3 hours labor @ $30/hr = -$28.60
(savings) for control case every 30,000 miles.
• Full Electric Vehicle (EV) - Spark Plug Replacement: Reference case replacement
interval is every 105,000 miles; cost is $32.00 for parts (4 spark plugs @ estimated
$8/plug), and 1.7 hours labor @ $30/hr = -$83.00 (savings) in control case every
105,000 miles.
• Full Electric Vehicle (EV) - Brake Fluid Replacement: Interval and fluid
specifications are identical in reference and control cases.
• Full Electric Vehicle (EV) - Engine Coolant Replacement: Reference case
replacement interval is every 100,000 miles; cost is $29.00 for parts (7.25 quarts @
estimated $4/quart), and 1.0 hour labor (estimated) @ $30/hr = -$59.00 (savings) for
control case every 100,000 miles. More recent information suggests a reference case
replacement interval every 100,000 miles; cost is $21.20 for parts (5.3 quarts @
estimated $4/quart), and 1.0 hour labor (estimated) @ $30/hr = -$51.20 (savings) for
control case every 100,000 miles.
• Full Electric Vehicle (EV) - Battery Coolant Replacement: Control case has a
recommended battery coolant replacement 150,000 miles; uses same coolant as a
gasoline engine but approximately three times the amount ($29.00 x 3 = $87.00);
assume labor is the same as the gasoline engine coolant changes ($30.00) for a total
cost of+$117.00 for control case every 150,000 miles. More recent information
suggests that perhaps the control case should have used a recommended battery
coolant replacement at 150,000 miles; uses same coolant as a gasoline engine but
approximately three times the amount for a parts cost of $63.20 (15.8 quarts @ est.
$4/quart); assume labor is the same as the gasoline engine coolant changes ($30.00)
for a total cost of+$93.20 for control case every 150,000 miles. Also, EPA applied
this cost to PHEVs as well, whereas NHTSA did not.
• Full Electric Vehicle (EV) - Battery Health Check: Two auto makers recommend
periodic battery/electrical checks to run in-depth diagnostics and visual inspection; no
information available on costs or interval; assume cost = cost of oil change and
interval is double that for oil change; +$38.67 every 15,000 miles in control case.
3-263
-------
Technologies Considered in the Agencies' Analysis
There is evidence supporting that brake maintenance costs are lower in hybrid electric
vehicles that are equipped with regenerative braking. The electric regeneration reduces the
amount of energy that the brake system dissipates which causes less wear on the brake pads
and rotors. However, the maintenance schedules do not reflect a lower frequency of
maintenance; therefore, the agencies have attempted to remain consistent with the current
methodology and have assumed no difference in cost (i.e., no savings).
For the first time in CAFE and GHG rulemaking, both agencies now include
maintenance costs in their benefit-cost analyses and in their respective payback analyses. As
noted above, please refer to each agency's preamble sections and final RIAs (Chapter 5 of
EPA's RIA, Chapter VIII of NHTSA's RIA) for details of how the maintenance costs
presented above are accounted for in each agency's respective analysis.
3.7.2 Repair Costs
Although NADA is correct that the agencies' NPRM analyses did not account for
repair costs to equipment added as a result of these rules and incurred throughout a vehicle's
useful life, the agencies' NPRM analysis did account for the costs of repairs covered by
manufacturers' warranties. (See 76 FR 74925 and 74927) 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 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. Therefore, the
agencies' central analyses supporting the final rule does not include these potential costs.
However, as at proposal, our analyses do include estimated warranty costs and, therefore, the
costs of repairs covered by OEMs' warranties. Repair costs are discussed further in each
agency's Regulatory Impact Analysis and preamble sections.
3-264
-------
Technologies Considered in the Agencies' Analysis
References:
1 U.S. EPA, "Project Report: Computer Simulation of Light-Duty Vehicle Technologies for
Greenhouse Gas Emission Reduction in the 2020-2025 Timeframe", Contract No. EP-C-11-
007, Work Assignment 0-12,
2 U.S. EPA, "Light-Duty Technology Cost Analysis Pilot Study," Contract No. EP-C-07-069,
Work Assignment 1-3, December 2009, EPA-420-R-09-020, Docket EPA-HQ-OAR-2009-
0472-11282
3 U.S. EPA, "Light-Duty Technology Cost Analysis Pilot Study Peer Review Report —
Response to Comments Document", December 21, 2009, EPA-HQ-OAR-2009-0472-11285
4 U.S. EPA, "Light-duty Technology Cost Analysis - Report on Additional Case Studies,"
EPA-HQ-OAR-2009-0472-11604
5 FEV, Inc., "Light-Duty Technology Cost Analysis, Report on Additional Transmission,
Mild Hybrid, and Valvetrain Technology Case Studies", Contract No. EP-C-07-069, Work
Assignment 3-3. November 2011.
6 FEV, Inc., "Light-Duty Technology Cost Analysis, Power-Split and P2 HEV Case Studies",
Contract No. EP-C-07-069, Work Assignment 3-3, EPA-420-R-11-015, November 2011.
Docket ID: EPA-HQ-OAR-2010-0799; Document ID: EPA-HQ-OAR-2010-0799-1101.
7 ANL BatPaC model can be found in Docket ID EPA-HQ-OAR-2010-0799.
8 ANL BatPaC model peer review report can be found in Docket ID EPA-HQ-OAR-2010-
0799
9 ANL BatPaC model can be found in Docket ID EPA-HQ-OAR-2010-0799.
10 EPA-420-R-10-901, April 2010.
11 "Interim Joint Technical Assessment Report: Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards for Model Years 2017-
2025," September 2010.
12 75 FR 76337.
13 RTI International. Automobile Industry Retail Price Equivalent and Indirect Cost
Multipliers. February 2009. http://www.epa.gov/otaq/ld-hwy/420r09003.pdf; 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. The peer review for the RTI report is at
http://www.epa.gov/otaq/ld-hwy/420r09004.pdf.
3-265
-------
Technologies Considered in the Agencies' Analysis
14 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.
15 Helfand, Gloria, and Todd Sherwood, "Documentation of the Development of Indirect Cost
Multipliers for Three Automotive Technologies," August 2009.
16 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.
17 Rogozhin, A., Gallaher, M., McManus, W., February 2009. Automobile industry retail
price equivalent and indirect cost multipliers. EPA-420-R-09-003. Available at:
/http ://www. epa.gov/otaq/ld-hwy/420r09003 .pdf
18 Rogozhin, A., Gallaher, M., McManus, W., February 2009. Automobile industry retail
price equivalent and indirect cost multipliers. EPA-420-R-09-003, Table 4-3.
19 U.S. Environmental Protection Agency and National Highway Traffic Safety
Administration (November 2011). Draft Joint Technical Support Document: Proposed
Rulemakingfor 2017-2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and
Corporate Average Fuel Economy Standards." EPA-420-D-11_901, available at
http://www.epa.gov/otaq/climate/documents/420dll901.pdf, p. 3-12.
90
National Research Council. Assessment of Fuel Economy Technologies for Light-Duty
Vehicles. Washington, D.C.: National Academies Press, 2010.
21 NRC, ibid, pp. 3-22, 6-16.
99 _
FEV, Inc., "Potential Stranded Capital Analysis on EPA Light-Duty Technology Cost
Analysis", Contract No. EP-C-07-069 Work Assignment 3-3, November 2011.
23 76 FR 57106 (September 15, 2011).
24 U.S. EPA, "A Study of Potential Effectiveness of Carbon Dioxide Reducing Vehicle
Technologies", Contract No. EP-C-06-003, Work Assignment 1-14,, June 2008, Report*
EPA420-R-08-004, available in the EPA docket EPA-HQ-OAR-2009-0472 and on the
internet at http://www.epa.gov/otaq/technology/420r08004a.pdf
25
76 FR 57106 (September 15, 2011)
26 Woldring, D., Landenfeld, T., Christie, M.J., 2007, "DI Boost: Application of a High
Performance Gasoline Direct Injection Concept." SAE Technical Paper Series No. 2007-01-
1410.
27 Kapus, P.E., Fraidl, G.K., Prevedel, K., Fuerhapter, A., 2007, "GDI Turbo - The Next
Steps." JSAE Technical Paper No. 20075355.
3-266
-------
Technologies Considered in the Agencies' Analysis
Hancock, D., Fraser, N., Jeremy, M., Sykes, R., Blaxill, H., 2008, "A New 3 Cylinder 1.21
Advanced Downsizing Technology Demonstrator Engine." SAE Technical Paper Series No.
2008-01-0611.
29 Lumsden, G., OudeNijeweme, D., Fraser, N. Blaxill, H., 2009, "Development of a
Turbocharged Direct Injection Downsizing Demonstrator Engine." SAE Technical Paper
Series No. 2009-01-1503.
30 Cruff, L., Kaiser, M., Krause, S., Harris, R., Krueger, U., Williams, M., 2010, "EBDI® -
Application of a Fully Flexible High Bmep Downsized Spark Ignited Engine." SAE Technical
Paper Series No. 2010-01-0587.
31 Turner, J.W.G., Pearson, R.J., Curtis, R., Holland, B., 2009, "Sabre: A Cost Effective
Engine Technology Combination for High Efficiency, High Performance and Low CO2
Emissions." Low Carbon Vehicles 2009: EVIechE Conference Proceedings.
32 Taylor, J., Fraser, N., Wieske, P., 2010, "Water Cooled Exhaust Manifold and Full Load
EGR Technology Applied to a Downsized Direct Injection Spark Ignition Engine." SAE
Technical Paper Series No. 2010-01-0356.
33 Roth, D.B., Keller, P, Becker, M., 2010, "Requirements of External EGRSystems for Dual
Cam Phaser Turbo GDI Engines." SAE Technical Paper Series No. 2010-01-0588.
34 Burress, T.A., C.L. Coomer, S.L. Campbell, L.E. Seiber, R.H. Staunton, and J.P.
Cunningham, 2008, "Evaluation of the 2007 Toyota Camry Hybrid Synergy Drive System."
ORNL technical report TM-2007/190.
35ICF International, 2011. "Peer Review of Ricardo, Inc. Draft Report, 'Computer
Simulation of Light-Duty Vehicle Technologies for Greenhouse Gas Emission Reduction in
the 2020-2025 Timeframe'". Contract No. EP-C-06-094, Work Assignment 4-04, Docket
EPA-HQ-OAR-2010-0799, September 30, 2011.
36 Ricardo Inc., "Project Report Computer Simulation of Light-duty Vehicle Technologies for
Greenhouse Gas Emission Reduction in the 2020-2025 Timeframe," 2011, Contract No. EP-
C-l 1-007, Work Assignment 0-12, Docket EPA-HQ-OAR-2010-0799, November, 2011.
37 Systems Research and Applications Corporation (SRA), "Peer Draft Response to Peer
Review of: Ricardo, Inc. Draft Report, 'Computer Simulation of Light-Duty Vehicle
Technologies for Greenhouse Gas Emission Reduction in the 2020-2025 Timeframe'" EPA
Contract No. EP-C-11-007, Work Assignment 0-12, Docket ID EPA-HQ-OAR-2010-0799,
November, 2011.
TO
Moawad, A. and Rousseau, A., "Impact of Electric Drive Vehicle Technologies on Fuel
Efficiency," Energy Systems Division, Argonne National Laboratory, ANL/ESD/12-7,
August 2012.
3-267
-------
Technologies Considered in the Agencies' Analysis
39 "Impact of Friction Reduction Technologies on Fuel Economy," Fenske, G. Presented at
the March 2009 Chicago Chapter Meeting of the 'Society of Tribologists and Lubricated
Engineers' Meeting, March 18th, 2009. Available at:
http://www.chicagostle.org/program/2008-
2009/Impact%20of%20Friction%20Reduction%20Technologies%20on%20Fuel%20Econom
y%20-%20with%20VGs%20removed.pdf (last accessed July 9, 2009).
40 "Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy
Trends: 1975 Through 2011", EPA-420-R-12-00la, March 2012,. Available at
http://www.epa.gov/otaq/cert/mpg/fetrends/2012/420rl2001a.pdf (last accessed August 7,
2012).
41 Paul Whitaker, Ricardo, Inc., "Gasoline Engine Performance And Emissions - Future
Technologies and Optimization," ERC Symposium, Low Emission Combustion Technologies
for Future 1C Engines, Madison, WI, June 8-9, 2005, Docket EPA-HQ-OAR-2009-0472-
0155. Available at
http://www.erc. wisc.edu/symposiums/2005_Symposium/June%208%20PM/Whitaker_Ricard
o.pdf (last accessed Nov. 9, 2008).
42 "Development and Optimization of the Ford 3.5L V6 EcoBoost Combustion System,"
Yi,J., Wooldridge, S., Coulson, G., Hilditch, J. Iyer, C.O., Moilanen, P., Papaioannou, G.,
Reiche, D. Shelby, M., VanDerWege, B., Weaver, C. Xu, Z., Davis, G., Hinds, B. Schamel,
A. SAE Technical Paper No. 2009-01-1494, 2009, Docket EPA-HQ-OAR-2009-0472-2860.
43 David Woldring and Tilo Landenfeld of Bosch, and Mark J. Christie of Ricardo, "DI Boost:
Application of a High Performance Gasoline Direct Injection Concept," SAE 2007-01-1410.
Available at http://www.sae.org/technical/papers/2007-01-1410 (last accessed Nov. 9, 2008)
44 Yves Boccadoro, Loi'c Kermanac'h, Laurent Siauve, and Jean-Michel Vincent, Renault
Powertrain Division, "The New Renault TCE 1.2L Turbocharged Gasoline Engine," 28th
Vienna Motor Symposium, April 2007.
45 Tobias Heiter, Matthias Philipp, Robert Bosch, "Gasoline Direct Injection: Is There a
Simplified, Cost-Optimal System Approach for an Attractive Future of Gasoline Engines?"
AVL Engine & Environment Conference, September 2005.
46 U.S. Environmental Protection Agency, "Draft Report - Light-Duty Technology Cost
Analysis Pilot Study," Contract No. EP-C-07-069, Work Assignment 1-3, September 3, 2009,
Docket EPA-HQ-OAR-2009-0472-0149.
47 Kaiser, M., Krueger, U., Harris, R., Cruff, L. "Doing More with Less - The Fuel Economy
Benefits of Cooled EGR on a Direct Injected Spark Ignited Boosted Engine," SAE Technical
Paper Series, No. 2010-01-0589.
48 Kapus, P.E., Fraidl, G.K., Prevedel, K., Fuerhapter, A. "GDI Turbo - The Next Steps,"
JSAE Technical Paper No. 20075355, 2007.
3-268
-------
Technologies Considered in the Agencies' Analysis
49 "EPA Staff Technical Report: Cost and Effectiveness Estimates of Technologies Used to
Reduce Light-duty Vehicle Carbon Dioxide Emissions," EPA420-R-08-008, March 2008,
Docket EPA-HQ-OAR-2009-0472-0132.
50 Tim Lake, John Stokes, Richard Murphy, and Richard Osborne of Ricardo and Andreas
Schamel of Ford-Werke, "Turbocharging Concepts for Downsized DI Gasoline Engines,"
VKA/ika Aachen Colloquium 2003. Available at
http://cat.inist.fr/?aModele=afficheN&cpsidt=16973598 (last accessed Nov. 9, 2008).
51http://media.gm.com/content/media/us/en/gm/news.detail.html/content/Pages/news/us/en/20
1 l/Jul/0722_cruze_diesel (last accessed: October 21, 2011)
52 http://www.mazda.com/publicity/release/2010/201010/101020a.html (last accessed:
October 21,2011)
53 U.S. EPA, "Computer Simulation of Light-Duty Vehicle Technologies for Greenhouse Gas
Emission Reduction in the 2020-2025 Timeframe", Contract No. EP-C-11-007, Work
Assignment 0-12, Docket EPA-HQ-OAR-2010-0799, November, 2011.
54 General Motors, news release, "From Hybrids to Six-Speeds, Direct Injection And More,
GM's 2008 Global Powertrain Lineup Provides More Miles with Less Fuel" (released Mar. 6,
2007). Available at
http://www.gm. com/experience/fuel economy/news/2007/adv engines/2008-powertrain-
lineup-082707.jsp (last accessed Sept. 18, 2008).
55 "EPA Staff Technical Report: Cost and Effectiveness Estimates of Technologies Used to
Reduce Light-duty Vehicle Carbon Dioxide Emissions" Environmental Protection Agency,
EPA420-R-08-008, March 2008, at page 17, Docket EPA-HQ-OAR-2009-0472-0132.
56 FEV, Inc., "Light-Duty Technology Cost Analysis, Report on Additional Transmission,
Mild Hybrid, and Valvetrain Technology Case Studies", Contract No. EP-C-07-069, Work
Assignment 3-3, Docket EPA-HQ-OAR-2010-0799, November 2011.
57 FEV, Inc., "Light-Duty Technology Cost Analysis, Report on Additional Transmission,
Mild Hybrid, and Valvetrain Technology Case Studies", Contract No. EP-C-07-069, Work
Assignment 3-3, Docket EPA-HQ-OAR-2010-0799, November 2011.
CO
http://automobiles.honda.com/insight-hybrid/features.aspx?Feature=ima (last accessed on
March 29, 2010; web page printout is contained in Docket EPA-HQ-OAR-2009-0472 as
"Honda IMA webpage.pdf).
59 "Latest Chevrolet Volt Battery Pack and Generator Details and Clarifications." Lyle
Dennis interview of Rob Peterson (GM) regarding the all-electric drive range of the GM Volt,
August 29, 2007. Accessed on the Internet on June 30, 2009 at: http://gm-
volt.com/2007/08/29/latest-Chevrolet-volt-battery-pack-and-generator-details-and-
clarifications/
3-269
-------
Technologies Considered in the Agencies' Analysis
60 "Active Combination of Ultracapacitors and Batteries for PHEV ESS." Bohn, T. U.S.
Department of Energy 2009 Vehicle Technologies Merit Review, May 20, 2009, Docket
EPA-HQ-O AR-2009-0472-0163.
61 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).
rr\
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 DJ.
Santini, Argonne National Laboratory, August 24, 2010.
63 Argonne National Laboratories, "Modeling the Performance and Cost of Lithium-Ion
Batteries for Electric-Drive Vehicles", Docket ID EPA-HQ-OAR-2010-0799 orNHTSA-
2010-0131-0101.
64 Peer review of report: Argonne National Laboratories, "Modeling the Performance and
Cost of Lithium-Ion Batteries for Electric-Drive Vehicles", 2011 can be found in Docket ID
EPA-HQ-OAR-2010-0799 or NHTSA-2010-0131-0101.
65 Argonne National Laboratories BatPac model can be found in Docket ID EPA-HQ-OAR-
2010-0799 or NHTSA-2010-0131-0102.
66 Nelson, P., Gallagher, K., Bloom, I. Dees, D.W. "BatPac Model Beta." Microsoft Excel™-
based spreadsheet model with example generic inputs and outputs. The model is available in
docket EPA-HQ-OAR-2010-0799 or NHTSA-2010-0131-0102.
67 Anderman, M. (2010) Feedback on ARB's Zero-Emission Vehicle Staff Technical Report
of 11/25/2009 including attachment A: Status of EV Technology Commercialization,
Advanced Automotive Batteries, January 6, 2010
68 Frost & Sullivan (2009b) World Hybrid Electric and Electric Vehicle Lithium-ion Battery
Market, N6BF-27, Sep 2009
69 Barnett, B. (2009) "PHEV Battery Cost Assessment" TIAX LLC presentation at U.S.
DOE/EERE 2009 Vehicle Technologies Program Annual Merit Review, May 19, 2009.
Accessed on the Internet on November 14, 2011 at:
http://wwwl.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2009/energy_storage/es_02
_barnett.pdf
70 Boston Consulting Group (2010) Batteries for Electric Cars - Challenges, Opportunities,
and the Outlook to 2020
71 National Research Council (2010) Transitions to Alternative Transportation Technologies—
Plug-in Hybrid Electric Vehicles.
3-270
-------
Technologies Considered in the Agencies' Analysis
72 K. G. Gallagher, P. A. Nelson, (2010) "An Initial BatPac Variation Study" Docket ID
EPA-HQ-O AR-2010-0799
73 "Hyundai ups tech ante with Sonata Hybrid," Automotive News, August 2, 2010.
74 "Chevrolet Stands Behind Volt With Standard Eight-Year, 100,000-Mile Battery
Warranty," GM Press release
(http://media.gm.com/content/media/us/en/news/news_detail.brand_gm.html/content/Pages/n
ews/us/en/2010/July/0714_volt_b artery)
75 "Nissan's new 2012 hybrid system aims for 1.8-L efficiency with a 3.5-L V6," SAE
Automotive Engineering Online, February 15, 2010.
76 "Lithium-ion Battery," Nissan Technological Development Activities (http://www.nissan-
global.com/EN/TECHNOLOGY/INTRODUCTION/DETAILS/LI-ION-EV/), 2009.
77 Plug-in Hybrid Electric Vehicle Charging Infrastructure Review (November 2008)
INL/EXT-08-15058, U.S. Department of Energy Vehicle Technologies Program - Advanced
Vehicle Testing Activity. Last accessed on the Internet on November 14, 2011 at the
following URL: http://avt.inl.gov/pdf/phev/phevInfrastructureReport08.pdf
78 May, J.W., Mattilla, M. "Plugging In: A Stakeholder Investment Guide for Public Electric-
Vehicle Charging Infrastructure." Rocky Mountain Institute - Project Get Ready, July 2009.
Last accessed on the Internet on November 14, 2011 at the following URL:
http://projectgetready.com/docs/Plugging%20In%20-
%20A%20Stakeholder%20Investment%20Guide.pdf
79 "Electric Vehicle Charging Infrastructure Deployment Guidelines - British Columbia."
Report by the Electric Transportation Engineering Corporation for Natural Resources Canada
and BCHydro. Last accessed on the Internet on November 14, 2011 at the following URL:
http://www.bchydro.com/etc/medialib/internet/documents/environment/EVcharging_infrastru
cture_guidelines09.Par.0001.File.EV%20Charging%20Infrastructure%20Guidelines-BC-
Aug09.pdf
80 "Electrification Roadmap - Revolutionizing Transportation and Achieving Energy
Security." Electrification Coalition, November 2009. Accessed on the Internet on November
14, 2011 at: http://www.electrificationcoalition.org/sites/default/files/SAF_1213_EC-
Roadmapjv 12_Onli ne. pdf
81 "The Disappearing Spare Tire" Edmunds.com, May 11, 2011; EPA-HQ-O AR-2010-0799.
http://www.edmunds.com/car-buying/the-disappearing-spare-tire.html (last accessed
9/6/2011)
82 see U.S Patent 5,227,425, Rauline to Michelin, July 13, 1993
3-271
-------
Technologies Considered in the Agencies' Analysis
OQ
"Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy
Trends: 1975 Through 2011", EPA420-R-12-00la, U.S. Environmental Protection Agency
Office of Transportation and Air Quality, March 2012
84 Lutsey, "Review of technical literature and trends related to automobile mass-reduction
technology", UCD-ITS-RR-10-10, May 2010 . EPA-HQ-OAR-2010-0799-0722. Available
at http://pubs.its.ucdavis.edu/publication_detail.php?id=1390 (last accessed Jun. 10, 2012).
85 Adrian Lund, IfflS, "The Relative Safety of Large and Small Passenger Vehicles." EPA-
HQ-OAR-2010-0799. Available at
http://www.nhtsa.gov/staticfiles/rulemaking/pdf/MSS/MSSworkshop-Lund.pdf (last accessed
Jun. 10, 2012).
86 Lutsey, "Review of technical literature and trends related to automobile mass-reduction
technology", UCD-ITS-RR-10-10, May 2010 . EPA-HQ-OAR-2010-2009-0722. Available
at http://pubs.its.ucdavis.edu/publication_detail.php?id=1390 (last accessed Jun. 10, 2012).
Q"J
Lotus Engineering, Inc. "An Assessment of Mass Reduction Opportunities for a 2017-2020
Model Year Vehicle program", March 2010.
88 Committee on the Assessment of Technologies for Improving Light-Duty Vehicle Fuel
Economy; National Research Council, "Assessment of Fuel Economy Technologies for Light-
Duty Vehicles", 2011. Available at http://www.nap.edu/catalog.php?record_id=12924 (last
accessed Jun 27, 2012).
89 Ford Sustainability Report 2010/11, http://corporate.ford.com/microsites/sustainability-
report-2010-ll/issues-climate-plan-economy (last accessed Aug. 26, 2011) EPA-HQ-OAR-
2010-0799.
90 ETA, US Steel, Tata Steel, "The ACP Process™ as Applied to the Future Steel Vehicle" ,
Docket No. EPA-HQ-OAR-2010-0799 orNHTSA-2010-0131. Available at
http://www.eta.com/index.php/engineering/product-design-development/the-acp-
process/success-stories/181 -the-acp-process-as-applied-to-the-future-steel-vehicle- (last
accessed Aug. 2,2012).
91 SAE World Congress, "Focus B-pillar 'tailor rolled' to 8 different thicknesses," Feb. 24,
2010. EPA-HQ-OAR-2010-0799. Available at http://www.sae.org/mags/AEI/7695 (last
accessed Jun. 10,2012).
92 VW comments, Docket No. EPA-HQ-OAR-2010-0799 and NHTSA-2010-0131-0247, at
18.
93 Committee on the Assessment of Technologies for Improving Light-Duty Vehicle Fuel
Economy; National Research Council, "Assessment of Fuel Economy Technologies for Light-
Duty Vehicles", June 3,2011. Available at http://www.nap.edu/catalog.php?record_id=12924
(last accessed Jun 27, 2012).
3-272
-------
Technologies Considered in the Agencies' Analysis
94 Malen, E. and K. Reddy, "Preliminary Vehicle Mass Estimation Using Empirical
Subsystem Influence Coefficients," Auto-Steel Partnership Report, May 2007. EPA-HQ-
OAR-2010-0799-1193. Available at http://www.a-
sp.org/~/media/Files/Autosteel/Research/Lightweighting/mass_compoundingpdf.ashx (last
accessed Jun. 27,2012).
95 Bull, M., R. Chavali, A. Mascarin, "Benefit Analysis: Use of Aluminum Structures in
Conjunction with Alternative Powertrain Technologies in Automobiles," Aluminum
Association Research Report, May 2008. Docket No. NHTSA-2010-0131-0097. Available at
http://aluminumintransportation.org/downloads/IBIS-Powertrain-Study.pdf (last accessed
Aug. 17,2011).
96 Bjelkengren, C, "The Impact of Mass Decompounding on Assessing the Value of Vehicle
Lightweighting", Docket NHTSA-2010-0131, Available at
http://msl.mit.edu/theses/Bjelkengren C-thesis.pdf (last accessed Aug 3, 2012).
97 American Iron and Steel Institute (AISI), 2009. "New Study Finds Increased Use of
Advanced High-Strength Steels Helps Decrease Overall Vehicle Weight." EPA-HQ-OAR-
2010-0799-1149. Also Available at
http://www.steel.org/en/sitecore/content/Global/Document%20Types/News/2009/Auto%20-
%20New%20Study%20Finds%20Increased%20Use%20of%20Advanced%20High-
Strength%20Steels.aspx (last accessed on June 11, 2012). .
98 Ford, 2010. "The 5.0 Liter is Back: 2011 Ford Mustang GT Leads Class with 412 HP, Fuel
Efficiency, Chassis Dynamics." EPA-HQ-OAR-2010-0799. Available at
http://media.ford.com/article_display.cfm?article_id=31645 (last accessed Jun. 10, 2012).
99 Information copied from http://www.mazda.com/csr/environment/making_car/
weight_reduction.html and docketed. EPA Docket: EPA-HQ-OAR-2010-0799.
100 The New York Times, "Automakers Resolve to Drop a Few pounds", Sept 2011.
http://www.nytimes.com/2011/09/18/automobiles/autoshow/in-frankfurt-automakers-vow-to-
drop-a-few-pounds.html?_r=l&smid=tw-nytimeswheels&seid=auto EPA Docket EPA-HQ-
OAR-2010-0799.
101 "Survey of vehicle mass-reduction technology trends and prospects" May 182010. EPA-
HQ-OAR-2010-0799. Also available at
http ://www. arb. ca.gov/msprog/levprog/leviii/meetings/051810/lutsey_its_may 18_final .pdf
109
NAS 2010, "Assessment of Fuel Economy Technologies for Light-Duty Vehicles". June
2010, page 7-14
103 Frank Field, Randolph Kirchain and Richard Roth, Process cost modeling: Strategic
engineering and economic evaluation of materials technologies, JOM Journal of the Minerals,
Metals and Materials Society, Volume 59, Number 10, 21-32. Available at
3-273
-------
Technologies Considered in the Agencies' Analysis
http://msl.mit.edu/publications/Field_KirchainCM_StratEvalMatls.pdf (last accessed Jun. 10,
2012).
104 Electricore/EDAG/GWU, "Mass Reduction for Light-Duty Vehicles for Model Years 2017-
2025", NHTS A Docket NHTS A-20 10-0131.
105 "Peer Review for 'Mass Reduction for Light-Duty Vehicles for Model Year 2017-2025 "',
NHTSA Docket NHTSA-2010-0131.
106 FEV, " Light-Duty Vehicle Mass-Reduction and Cost Analysis - Midsize Crossover Utility
Vehicle ". August 2012, EPA Docket: EPA-HQ-OAR-20 10-0799.
107 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-20 10-0799.
108 j^pj international, "peer Review of Lotus Engineering Vehicle Mass Reduction Study"
EPA-HQ-OAR-20 10-0799-07 10, November 2010.
109
international, "Peer Review of Lotus Engineering Vehicle Mass Reduction Study"
EPA-HQ-OAR-20 10-0799-07 10, November 2010.
110 Frank Field, Randolph Kirchain and Richard Roth, Process cost modeling: Strategic
engineering and economic evaluation of materials technologies, JOM Journal of the Minerals,
Metals and Materials Society, Volume 59, Number 10, 21-32. Available at
http://msl.mit.edu/pubs/docs/Field_KirchainCM_StratEvalMatls.pdf (last accessed Aug. 22,
2011). EPA-HQ-OAR-20 10-0799.
111 U.S. EPA, Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel
Economy Trends: 1975 Through 2011, http://epa.gov/otaq/fetrends.htm
112 EPA; Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy
Trends: 1975 Through 2010; Figure 28, pg 69 EPA-HQ-OAR-2010-0799-0813.
113 EPA; Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy
Trends: 1975 Through 2010; Table 13, pg 50. EPA-HQ-OAR-2010-0799-0813.
114 DeCicco, J. M. (2010). A fuel efficiency horizon for U.S. automo- biles. Technical report,
University of Michigan, School of Natural Resources and Environment. Report prepared for
The Energy Foundation. EPA-HQ-OAR-20 10-0799-08 16. Available online at
http://energy.umich.edu/info/pdfs/Fuel%20Efficiency%20Horizon%20FINAL.pdf.
115 Zoepf, Stephen. (201 1) "Automotive Features: Mass Impact and Deployment
Characterization" Masters thesis. Massachusetts Institute of Technology, Technology and
Policy Program, Engineering Systems Division. EPA-HQ-OAR-20 10-0799. Also available
online: http://web.mit.edu/sloan-auto-lab/research/beforeh2/files/Zoepf MS Thesis.pdf
3-274
-------
Technologies Considered in the Agencies' Analysis
116 Murphy, John, Bank of America Merrill Lynch, "Car Wars 2010-2013," July 15, 2009.
EPA-HQ-O AR-2010-0799-0818
117 Ellison, D. J., Clark, K. B., Fujimoto, T., and suk Hyun, Y. (1995). Product development
performance in the auto industry: 1990s update. Technical report, International Motor Vehicle
Program.
118 Autocar, "VW's New Platform for 60 Models," November 12, 2009,
http://www.autocar.co.uk/News/NewsArticle.aspx?AR=244881 EPA Docket EPA-HQ-
O AR-2010-070
119 Mayne, Eric, "Aligning Capacity, Demand Poses Ultimate Brain Teaser," Wards Auto,
July 29, 2008. EPA Docket EPA-HQ-O AR-2010-0799-0811
120 Pope, Byron, "Ford's Cleveland Engine No. 1 to Build 3.7L V-6," Wards Auto, March 6,
2009. EPA Docket EPA-HQ-O AR-2010-0799-0815
191
Ford Motor Company, Ford Motor Company Business Plan Submitted to the House
Financial Services Committee, December 2, 2008. EPA Docket EPA-HQ-O AR-2010-0799-
0814
122 Rogers, Christina "GM to halve number of platforms globally," Detroit News, August 10,
2011. EPA Docket EPA-HQ-O AR-2010-0799-0817.
3-275
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Chapter 4: Economic and Other Assumptions Used in the
Agencies' Analysis
4.1 How the Agencies use the economic and other assumptions in their analyses
Improving new vehicles' fuel efficiency and reducing greenhouse gas (GHG)
emissions provides direct benefits to their buyers and users by reducing fuel consumption and
fuel costs throughout those vehicles' lifetimes, stimulating increased vehicle miles traveled
(VMT) through the fuel economy rebound effect, and often increasing vehicles' driving range
leading to less frequent refueling. At the same time, the reduction in fuel use that results
from requiring higher fuel economy and reducing GHG emissions also produces wider
benefits to the U.S. economy by lowering the cost of economic externalities that result from
U.S. petroleum consumption and imports. This occurs because reducing U.S. oil consumption
and imports lowers the potential costs from disruptions in the flow of oil imports, reduces the
sensitivity of the U.S. economy to oil price shocks, and has the potential to reduce the global
price of petroleum. Reducing fuel consumption and GHGs also lowers the economic costs of
environmental externalities resulting from fuel production and use, including reducing
potential future human health and economic damages from changes in the global climate
caused by greenhouse gas emissions, and reducing the impacts on human health from
emissions of criteria air pollutants.
These social benefits are partly offset by the increase in fuel use that results from
added vehicle use due to the fuel economy rebound effect, as well as by added costs from the
increased congestion, crashes, and noise caused by increased vehicle use. They would also be
offset by any loss in the utility that new vehicles provide to their buyers (and subsequent
owners) if manufacturers include reductions in vehicles' performance, carrying capacity, or
comfort as part of their strategies to comply with higher fuel economy requirements and GHG
standards. However, the agencies' analyses supporting the final standards do not anticipate
any such reductions in utility as being necessary, and the analysis includes the costs to
manufacturers of preserving vehicle capabilities.51 (For example, the costs of engine
downsizing include the costs of turbocharging the engine to maintain its performance). The
total economic benefits from requiring higher fuel economy and reducing GHG emissions are
likely to be substantial, and EPA and NHTSA have developed detailed estimates of the
economic benefits from adopting the final standards.
This chapter discusses the common economic and other values used by both NHTSA
and EPA in their rulemaking analyses. These inputs incorporate a range of forecast
information, economic estimates, and input parameters. This chapter describes the sources
that EPA and NHTSA have relied upon for this information, the rationale underlying each
assumption, and the agencies' estimates of specific parameter values. These common values
are then used as inputs into each agency's respective modeling and other analyses of the
a Two exceptions - hybrid vehicles that may have some limited towing capacity, and electric vehicles - are
discussed elsewhere.
4-1
-------
Economic and Other Assumptions Used in the Agencies' Analysis
economic benefits and costs of the EPA and NHTSA programs. While the underlying input
values are common to both agencies, programmatic differences, and differences in the way
each agency assesses its program result in differing benefit and cost estimates. This issue is
discussed further in Section 1C of the preamble to the joint rulemaking. Unless otherwise
noted, a summary of the public comments received on the topics described in this chapter and
the agencies responses are included in the preamble Section II.E, Section III.H, and Section
IV.X.
4.2 What assumptions do the agencies use in the impact analyses?
4.2.1 The on-road fuel economy "gap"
4.2.1.1 Definition and past use by EPA and NHTSA
In aggregate, actual fuel economy levels achieved by vehicles in on-road driving fall
significantly short of their levels measured in the laboratory-like test conditions and two-cycle
tests used under the CAFE program to determine the fuel economy ratings for different
models for purposes of compliance with the CAFE and CC>2 standards. The test procedure
used to determine compliance is highly controlled, and does not reflect real-world driving in a
variety of ways - real-world driving tends to be more aggressive than the Federal Test
Procedure (FTP) and Highway Fuel Economy Test (FIFET) test cycles used to establish
compliance with the GHG and CAFE regulations. Real world driving tends to include more
stops and starts and more rapid acceleration/deceleration, and may include the use of
technologies like air-conditioning that reduce fuel economy but that are not exercised on these
test cycles.1 There are also a number of environmental elements that affect real-world
achieved fuel economy which are not measured on the two cycle compliance test, such as
wind resistance, road roughness, grade, temperature, and fuel energy content. The agencies'
analyses for this final rulemaking recognize this gap between compliance results and real
world performance, and account for it by adjusting the fuel economy downward from its rated
value.
In December 2006, EPA adopted changes to its regulations on fuel economy labeling,
which were intended to bring vehicles' label fuel economy levels seen by consumers shopping
for new vehicles closer to their actual on-road fuel economy levels. Comparisons of on-road
and CAFE fuel economy levels developed by EPA as part of its 2006 Final Rule
implementing new fuel economy labeling requirements for new vehicles indicated that actual
on-road fuel economy for light-duty vehicles average about 20 percent lower than compliance
fuel economy ratings.3 While there is great heterogeneity among individual drivers, as
discussed in the referenced material, the 20 percent figure appears to represent an accurate
average for modeling a fleet. For example, if the overall EPA fuel economy rating of a light
truck is 20 MPG, the on-road fuel economy actually achieved by a typical driver of that
vehicle is expected to be 16 mpg (20*.80). In its analysis supporting the Final Rule
establishing CAFE standards for MY 2011, NHTSA employed EPA's revised estimate of this
on-road fuel economy gap in its analysis of the fuel savings resulting from alternative fuel
efficiency standards. EPA and NHTSA likewise employed this fuel economy gap for
estimating fuel savings in the MYs 2012-2016 rulemaking in the Interim Joint Technical
Assessment Report (TAR) analysis for MYs 2017 and later, the proposal, and in this
4-2
-------
Economic and Other Assumptions Used in the Agencies' Analysis
rulemaking establishing GHG and fuel economy standards for MY 2017 and later light duty
vehicles.
An analysis conducted by NHTSA confirmed that EPA's estimate of a 20 percent gap
between test and on-road fuel economy for the majority of vehicles is well-founded. NHTSA
used data on the number of passenger cars and light trucks of each model year that were in
service (registered for use) during each calendar year from 2000 through 2006; average fuel
economy for passenger cars and light trucks produced during each model year; and estimates
of average miles driven per year by cars and light trucks of different ages during each
calendar year over that period. These data were combined to develop estimates of the usage-
weighted average fuel economy that the U.S. passenger car and light truck fleets would have
achieved during each year from 2000 through 2006 under test conditions.
Table 4-1 compares NHTSA's estimates of fleet-wide average fuel economy under
test conditions for 2000 through 2006 to the Federal Highway Administration's (FHWA)
published estimates of on-road fuel economy achieved by passenger cars and light trucks
during each of those years. As it shows, FHWA's estimates of fuel economy for passenger
cars ranged from 21-23 percent lower than NHTSA's estimates of its fleet-wide average value
under test conditions over this period, and FHWA's estimates of fuel economy for light trucks
ranged from 16-18 percent lower than NHTSA's estimates of its fleet-wide average value
under test conditions. Thus, these results appear to confirm that the 20 percent on-road fuel
economy gap represents a reasonable estimate for use in evaluating the fuel savings likely to
result from more stringent fuel economy and CC>2 standards in MYs 2017-2025.
Table 4-1 Estimated Fleet-Wide Fuel Economy of Passenger Cars and Light Trucks
Compared to Reported Fuel Economy
YEAR
2000
2001
2002
2003
2004
2005
PASSENGER CARS
NHTSA
Estimated
Test MPG
28.2
28.2
28.3
28.4
28.5
28.6
FHWA
Reported
MPG
21.9
22.1
22.0
22.2
22.5
22.1
Percent
Difference
-22.2%
-21.7%
-22.3%
-21.9%
-21.1%
-22.8%
LIGHT-DUTY TRUCKS
NHTSA
Estimated
Test MPG
20.8
20.8
20.9
21.0
21.0
21.1
FHWA
Reported
MPG
17.4
17.6
17.5
17.2
17.2
17.7
Percent
Difference
-16.3%
-15.5%
-16.2%
-18.0%
-18.3%
-16.3%
The agencies did not adopt this approach in assessing benefits of the GHG emission standards and fuel
consumption standards for heavy duty vehicles (76 FR 57106 (Sept. 15, 2011)) since compliance with those
rules is assessed using test procedures that necessitate different modeling assumptions.
4-3
-------
Economic and Other Assumptions Used in the Agencies' Analysis
2006
Avg.,
2000-
2006
28.8
28.4
22.5
22.2
-21.8%
-22.0%
21.2
21.0
17.8
17.5
-16.2%
-16.7%
We are aware of two potential issues involved in these estimates. One, the estimates
of total annual car and truck VMT are developed by the states and submitted to FHWA. Each
state uses its own definition of a car and a truck. For example, some states classify minivans
as cars and some as trucks. Thus, there are known inconsistencies with these estimates when
evaluated separately for cars and trucks. Also, total gasoline consumption can be reasonably
estimated from excise tax receipts, but separate estimates for cars and trucks are not available.
For this reason, NHTSA and EPA are not using distinct on-road fuel economy gaps for cars
and trucks, but one common value of 20 percent for both vehicle classes for purposes of
estimating the fuel savings of the standards. This figure lies between the separate estimated
for cars and light trucks reported in Table 4-1.
Consistent with the MYs 2012-2016 rulemaking, the TAR, and the proposal, in this
final rulemaking the agencies are assuming that the on-road fuel economy gap for liquid fuel
is 20 percent. As in the TAR and proposal, the agencies assume that the overall energy
shortfall for the electric drivetrain (for vehicles that have those instead of or in addition to
gasoline engines) is 30 percent when driven on wall electricity. The 30 percent value was
derived from the agencies' engineering judgment based on several data points. Foremost
among these, during the stakeholder meetings conducted prior to the Interim Joint TAR,
confidential business information (CBI) was supplied by several manufacturers which
indicated that electrically powered vehicles had greater variability in their on-road energy
consumption than vehicles powered by internal combustion engines. Second, data from
EPA's 2006 analysis of the "five cycle" fuel economy label as part of the rulemaking
discussed above potentially supported a larger on-road shortfall for vehicles with hybrid-
electric drivetrains4 And third, heavy accessory load, extreme (both high and low)
temperatures, and aggressive driving have deleterious impacts of unknown magnitudes on
battery performance. As a counterpoint, CBI provided by several other manufacturers
suggested that the on-road/laboratory differential attributable to electric operation should
approach that of liquid fuel operation in the future. Consequently, 30 percent was judged by
the agencies to be a reasonable estimate for the interim Joint TAR, and was carried into the
analysis supporting the final rules.
The recent 2011 Fuel Economy labeling rule similarly employs a 30% on-road
shortfall for electric vehicles.5 Under the labeling program, for gasoline vehicles, there are
two methods for getting label values: full 5-cycle or derived 5-cycle. Full 5-cycle means all
five cycles are tested, and bag MPG results are used in a set of formulae to determine label
MPG. Derived 5-cycle involves testing on the FTP and Highway tests and adjusting those
values using regression-based formulae, to get label MPG values. The derived 5-cycle
adjustment results in an ever-increasing adjustment in percentage terms. However, the data
on which the derived 5-cycle formulae are based ends at roughly 70 MPG, where the
adjustment is about 70% or an on-road gap of 30% (assuming that the five cycle formula
represents the real world). For labeling purposes, lacking any EVs or PHEVs (or any vehicles
4-4
-------
Economic and Other Assumptions Used in the Agencies' Analysis
beyond 70 MPG) in the database at the time this adjustment was derived, the adjustment was
set at 70% for MPG values beyond 70 MPG.
Electric vehicles are allowed and expected to use the derived 5-cycle method, which
suggests that their on-road gap will be approximately 30% during the near future. Individual
EVs may vary, and as additional data becomes available the agencies will consider whether
the 30% average gap remains appropriate.
The U.S. Coalition for Advanced Diesel Cars suggested that the on-road gap used in
the proposal was overly conservative, and that advanced technology vehicles may have on-
road gaps larger than 20%. The agencies recognize this potential issue - future changes in
driver behavior or vehicle technology may change the on-road gap. The Coalition states that
the EPA 2012 Trends Report shows that the gap for gasoline vehicles grew from 20% in 2005
to 20.5% in 2010, and that therefore the 20% value used by the agencies is understated. We
note that in recognition of the potentially greater gap for electrification technologies, the
agencies are using a 30% adjustment for wall electricity; but more broadly, to the extent that
the Coalition is suggesting that the agencies extrapolate the growth trend in the gap into the
future, the agencies do not agree that the estimate of the future on road gap would be
appropriately estimated by extrapolating the historical relationship between the test procedure
and real world fuel consumption and emissions. That historical rate of change occurred as a
result of the specific technological changes in vehicles over that timeframe. In the future,
different technologies will be employed, that are likely to affect the gap differently. As an
example, while some technologies such as electrification may increase the on-road gap, other
off-cycle technologies such as tire pressure management systems, air conditioning
improvements and aerodynamic improvements may decrease it. Thus, the agencies are
continuing to use the same on-road gap methodology as in the proposal for this final
rulemaking, but will monitor the EPA fuel economy database as these vehicles enter the fleet.
Considerations in Future Years
Looking forward to MYs 2017-2025, while the agencies do not forecast changes in
most of the factors discussed above that contribute to the on-road gap in ways that would
change our estimates, the agencies expect that two specific factors will change somewhat that
could affect this analysis. Specifically, we anticipate changes in the energy content of fuels
sold at retail as a result of the recent EPA Renewable Fuel Standard 2 (RFS2) rulemaking and
El5 waiver decision,6 as well as a change in reference air conditioning efficiency as a result
of the recent MYs 2012-2016 EPA Light Duty Greenhouse Gas rulemaking.
4.2.1.2 Air Conditioning
Air conditioning is a significant contributor to the on-road efficiency gap. While the
air conditioner is turned off during the FTP and HFET tests, in real world use drivers often
use air conditioning in warm, humid conditions. The air conditioning compressor can also be
engaged during "defrost" operation of the heating system.7 In the MYs 2012-2016
rulemaking, the agencies estimated the average impact of an air conditioning system at
approximately 14.3 grams over an SCO3 test for an average vehicle without any of the
improved air conditioning technologies discussed in that rulemaking. For a 27 MPG (330 g
4-5
-------
Economic and Other Assumptions Used in the Agencies' Analysis
CO2/mile) vehicle, this is approximately 20 percent of the total estimated on-road gap, or
about 4 percent of total fuel consumption.
In the MYs 2012-2016 rule, EPA estimated that 85 percent of MY 2016 vehicles
would reduce their air conditioning-related CC>2 emissions by 40 percent through the use of
advanced air conditioning efficiency technologies.8 Incorporating this change would reduce
the average on-road gap by about 2 percent in the reference case.0 However, as shown in
Chapter 5 of the joint TSD air conditioning-related fuel consumption does not proportionally
decrease as overall engine efficiency improves. Unlike most technologies in this rulemaking,
which have a multiplicative reduction on fuel consumption and CC>2 emissions, the load due
to air conditioning operation is relatively constant across engine efficiency and technology.
As a consequence, as engine efficiency increases, air conditioning operation represents an
increasing percentage of vehicular fuel consumption.11 To some extent, these factors are
expected to counterbalance, so the agencies therefore chose not to make an air conditioning-
related adjustment to the on-road gap for this rule.
4.2.1.3 Fuel Energy Content
Differences in fuel energy content between test conditions and real-world driving is
another contributor to the on-road fuel economy gap. Two-cycle testing for CAFE and CC>2
compliance is based on "certification fuel" which contains no ethanol (also known as EO). The
on-road fuel economy gap is estimated with reference to the difference in fuel energy content
between certification fuel and 2004 retail gallons,6 but this rule produces a reduction in
petroleum based fuel consumption only. Volumes of renewable fuels are statutorily fixed
by the Renewable Fuel Standard, so the entirety of the energy savings will take place as
reduced oil consumption. To estimate the petroleum fuel savings, we modify the on-road gap
by the average difference in energy content between CY 2004 retail fuel used in the five cycle
analysis and certification fuel. This results in an approximately 1% higher fuel economy than
if no additional adjustment was made for fuel energy content, and corresponds to the greater
energy content of certification gasoline as compared to 2004 retail gasoline.
Where:
Gap= 20%
c A 330 gram (27 MPG) vehicle has an estimated gap of about 80 grams/mile (330/0.8). Under the EPA MY
2016 rulemaking we assume a reduction of about 5 grams/mile from indirect air conditioning improvements,
This A/C improvement is about 1-2% of total fuel consumption
d As an example, the air conditioning load of 14.3 g/mile of CO2 is a smaller percentage (4.3%) of 330 g/mile
than of 260 g/mile (5.4%).
e The five cycle formula analysis is based on CY 2004 data.
f Ethanol contains approximately 76,000 British Thermal Units (Btu) per gallon as compared to petroleum
gasoline (Indolene), which contains approximately 115,000 Btu. Thus, a 10 percent ethanol (E10) blend
contains approximately 3.3 percent less energy than a gallon of EO, andanE15 blend contains approximately 5.1
percent less energy than a gallon of EO.
4-6
-------
Economic and Other Assumptions Used in the Agencies' Analysis
EOBTU/Gallon= 115,000
2004 BTU/Gallon = 113,912 (3.14% ethanol, 96.86% petroleum gasoline)
A related adjustment in fuel energy was made in order to "match" fuel savings to the
fuel prices used in this analysis. As discussed below, the agencies use liquid fuel prices from
the Energy Information Administration's (EIA) Annual Energy Outlook (AEO) 2012 Early
Release reference case.g By contrast, and as shown above, the gasoline savings from this rule
are calculated as gallons of certification fuel, which is more energy dense than ethanol
blended market fuel. To appropriately apply the AEO prices on a dollar per btu basis, we
adjust our certification fuel savings upwards by approximately 5% (the difference between the
energy content of E15 retail fuel and certification) when monetizing the fuel savings. This
adjustment more appropriately reflects AEO projections of motor gasoline energy prices.
4.2.2 Fuel prices and the value of saving fuel
Projected future fuel prices are a critical input into the preliminary economic analysis
of alternative fuel economy and GHG standards, because they determine the value of fuel
savings both to new vehicle buyers and to society. For the final rule, EPA and NHTSA relied
on the most recent fuel price projections from the U.S. Energy Information Administration's
(EIA) Annual Energy Outlook (AEO) for this analysis, the AEO 2012 Early Release
Reference Case. The Reference Case forecasts inflation-adjusted (constant-dollar) retail
gasoline and diesel fuel prices and represents the EIA's most up-to-date estimate of future
prices for petroleum products. In the Executive Summary to AEO 2012 Early Release, the
Energy Information Administration describes the reference case. They state that:
"Projections in the Annual Energy Outlook 2012 (AEO2012) Reference case focus on the
factors that shape U.S. energy markets in the long term, under the assumption that current
laws and regulations remain generally unchanged throughout the projection period. The
AEO2012 Reference case provides the basis for examination and discussion of energy
market trends and serves as a starting point for analysis of potential changes in U.S.
energy policies, rules, or regulations or potential technology breakthroughs."11
The Reference case projection is a business-as-usual trend estimate, given known
technology and technological and demographic trends. The agency has published annual
projections of energy prices and consumption levels for the U.S. economy since 1982 in its
Annual Energy Outlooks. 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.
8 EIA projects that the average gallon of retail motor gasoline contains 5.040 mmbtu/barrel (Higher Heating
Value), as compared to 5.253 mmbtu/barrel for pure motor gasoline, which is a difference of approximately
4.5% (AEO 2012 Early Release table 147).
h AEO 2012 ER overview - http://www.eia.gov/forecasts/aeo/er/pdf/0383er(2012).pdf
4-7
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Several commenters (Volkwagen, Consumer Federation of America, Environmental
Defense Fund, Consumer's Union, National Resources Defense Council, Union of Concerned
Scientists) noted that the EIA future fuel price projections used in the proposal were similar to
current prices, "modest," or lower than expected. Other commenters noted the uncertainty
projecting during this extended time period (National Automobile Dealers' Association). No
commenters offered alternative sources for fuel price projection, and in this final rulemaking,
the agencies continue to rely upon EIA projections of future gasoline and diesel prices.
As compared to the gasoline prices used in the proposal-, which relied on projections
from AEO 2011, the AEO 2012 Early Release Reference Case fuel prices are somewhat
higher. A comparison is presented below in Table 4-2.
Table 4-2 Gasoline Prices for Selected Years in AEO 2011 and 2012
(Presented in constant 2010$ and including all taxes)
AEO 2011
AEO 2012 (ER)
2015
$3.17
$3.53
2020
$3.42
$3.76
2030
$3.68
$4.04
The retail fuel price forecasts presented in AEO 2012 Early Release span the period
from 2009 through 2035. Measured in constant 2010 dollars, the AEO 2012 Early Release
Reference Case projections of retail gasoline prices during calendar year 2017 is $3.63 per
gallon, rising gradually to $4.09 by the year 2035 (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 2017-25 requires fuel price forecasts that extend
through approximately 2060, approximately the last year during which a significant number
of MY 2025 vehicles will remain in service.1 Due to the difficulty in accurately projecting
fuel prices over this long time span, the agencies have used a simple method for extrapolation
over the out years. To obtain fuel price forecasts for the years 2036 and later, the agencies
assume that retail fuel prices will continue to increase after 2035 at the average annual rate
(0.8%) projected for 2017-2035 in the AEO 2012 Early Release Reference Case. The years
between 2008 and 2016 were not included in the extrapolation due to the high volatility in the
AEO projection for those years This assumption results in a projected retail price of gasoline
that reaches $4.57 in 2050.
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 $0.43 per gallon during 2008, while those
1NHTSA defines the maximum lifetime of vehicles as the highest age at which more than 2 percent of those
originally produced during a model year remain in service. In the case of light trucks, for example, this age has
been assumed at 37 years for this analysis, see section 4.2.3.
4-8
-------
Economic and Other Assumptions Used in the Agencies' Analysis
levied on diesel averaged $0.46. 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.9 When calculating the value of fuel saved by an individual driver, however,
these taxes are included as 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 $0.36 (in constant 2010 dollars) more from the perspective of an individual
vehicle buyer than from the overall perspective of the U.S. economy .J
In the estimates of costs and benefits presented in the preamble and in the agencies'
RIAs, the agencies have included the full fuel savings over vehicles' expected lifetimes,
discounted to their present values using both 3 and 7 percent discount rates. Additional
discussion of this approach can be found in preamble Sections III.H and IV.C.3.
4.2.3 Vehicle Lifetimes and Survival Rates
The agencies' analyses of fuel savings and related benefits from adopting more
stringent fuel economy and GHG standards for MYs 2017-2025 passenger cars and light
trucks 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.
The first step in estimating lifetime fuel consumption by vehicles produced during a
model year is to calculate the number of those vehicles expected to remain in service during
each future calendar year after they are produced and sold.k This number is calculated by
multiplying the number of vehicles originally produced during a model year by the proportion
expected to remain in service at the age they will have reached during each subsequent
calendar year, often referred to as a "survival rate."
J For society, the fuel taxes represent a transfer payment. By contrast, an individual realizes savings from not
paying the additional money.
k Vehicles are defined to be of age 1 during the calendar year corresponding to the model year in which they are
produced; thus for example, model year 2000 vehicles are considered to be of age 1 during calendar year 2000,
age 2 during calendar year 2001, and to reach their maximum age of 30 years during calendar year 2029.
NHTSA considers the maximum lifetime of vehicles to be the age after which less than 2 percent of the vehicles
originally produced during a model year remain in service. Applying these conventions to vehicle registration
data indicates that passenger cars have a maximum age of 30 years, while light trucks have a maximum lifetime
of 37 years. See Lu, S., NHTSA, Regulatory Analysis and Evaluation Division, "Vehicle Survivability and
Travel Mileage Schedules," DOT HS 809 952, 8-11 (January 2006). Available at http://www-
nrd.nhtsa.dot.gov/Pubs/809952.pdf (last accessed Sept. 9, 2011). For the Final Rule, the survivability schedules
developed by Lu were updated using national vehicle registration data collected by R.L. Polk for calendar years
2006-2010.
4-9
-------
Economic and Other Assumptions Used in the Agencies' Analysis
The proportions of passenger cars and light trucks expected to remain in service at
each age are estimated from R.L. Polk vehicle registration data for calendar years 1970-2010,
and are shown in Table 4-3.10 Note that these survival rates were calculated against the pre-
MY 2011 definitions of cars and light trucks, and are not projected to change over time in the
analysis. The rates are applied to vehicles based on their regulatory class (passenger car or
light truck) regardless of fuel type or level of technology.
The survival and annual mileage estimates reported in this section's tables reflect the
convention that vehicles are defined to be of age 1 during the calendar year that coincides
with their model year. Thus for example, model year 2017 vehicles will be considered to be
of age 1 during calendar year 2017. This convention is used in order to account for the fact
that vehicles produced during a model year typically are first offered for sale in June through
September of the preceding calendar year (for example, sales of a model year typically begin
in June through September of the previous calendar year, depending on manufacturer). Thus,
virtually all of the vehicles produced during a model year will be in use for some or all of the
calendar year coinciding with their model year, and they are considered to be of age 1 during
that year.1
1A slight increase in the fraction of new passenger cars remaining in service beyond age 10 has accounted for a
small share of growth in the U.S. automobile fleet. The fraction of new automobiles remaining in service to
various ages was computed from R.L. Polk vehicle registration data for 1977 through 2005 by the DOT's Center
for Statistical Analysis.
4-10
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Table 4-3 Survival Rates
VEHICLE AGE
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
32
33
34
35
36
37
ESTIMATED
SURVIVAL
FRACTION
CARS
1.0000
0.9878
0.9766
0.9614
0.9450
0.9298
0.9113
0.8912
0.8689
0.8397
0.7999
0.7556
0.7055
0.6527
0.5946
0.5311
0.4585
0.3832
0.3077
0.2414
0.1833
0.1388
0.1066
0.0820
0.0629
0.0514
0.0420
0.0337
0.0281
0.0235
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
ESTIMATED
SURVIVAL
FRACTION
LIGHT TRUCKS
1.0000
0.9776
0.9630
0.9428
0.9311
0.9152
0.8933
0.8700
0.8411
0.7963
0.7423
0.6916
0.6410
0.5833
0.5350
0.4861
0.4422
0.3976
0.3520
0.3092
0.2666
0.2278
0.2019
0.1750
0.1584
0.1452
0.1390
0.1250
0.1112
0.1028
0.0933
0.0835
0.0731
0.0619
0.0502
0.0384
0.0273
4-11
-------
Economic and Other Assumptions Used in the Agencies' Analysis
4.2.4 VMT
The second step in estimating lifetime fuel use by the cars or light trucks produced
during a future model year is to calculate the total number of miles that they will be driven
during each year of their expected lifetimes. To estimate total miles driven, the number of
cars and light trucks projected to remain in use during each future calendar year is multiplied
by the average number of miles a surviving car or light truck is expected to be driven at the
age it will have reached in that year. Estimates of average annual miles driven by cars and
light trucks of various ages were developed by NHTSA from the Federal Highway
Administration's 2009 National Household Travel Survey. This updates the schedules of
annual miles driven that were used in the NPRM, which were based on the previous National
Household Travel Survey, conducted in 2001. Additionally, the agencies have accounted for
the higher usage of fleet vehicles, which include rental vehicles as well as those owned by
corporations and government agencies. These represent about 20% of new vehicle sales, are
not represented in the NHTS, and are driven much more intensively (on average) than
household vehicles for the first several years of their lives before being absorbed into the
household vehicle population.™ The updated mileage schedules are reported in Table 4-4.
These estimates represent the average number of miles driven by a surviving light duty
vehicle at each age over its estimated full lifetime. To determine the number of miles a
typical vehicle produced during a given model year is expected to be driven at a specific age,
the average annual mileage for a vehicle of that model year and age is multiplied by the
corresponding survival rate for vehicles of that age. NHTSA has documented these analyses
in a memo to the docket.
m Using the Annual Energy Outlook 2012, early release version of the National Energy Modeling System,
developed and maintained by the U.S. Energy Information Administration, the proportion of fleet vehicles and
their typical usage were calculated and then averaged into the household mileage accumulation schedules
developed using the 2009 NHTS.
4-12
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Table 4-4 CY 2009 Mileage Schedules based on NHTS Data
VEHICLE AGE
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
32
33
34
35
36
37
ESTIMATED
VEHICLE MILES
TRAVELED
CARS
14,700
14,252
14,025
13,593
13,324
13,064
12,809
11,378
11,087
10,806
10,535
10,273
10,021
9,779
9,547
9,324
9,111
8,908
8,714
8,530
8,356
8,192
8,037
7,892
7,757
7,632
7,516
7,410
7,314
7,227
7,151
7,083
7,026
6,979
6,941
6,912
6,894
ESTIMATED
VEHICLE MILES
TRAVELED
LIGHT TRUCKS
15,974
15,404
14,841
14,435
14,038
13,650
12,590
12,192
11,810
1 1 ,443
11,091
10,755
10,434
10,129
9,839
9,564
9,305
9,061
8,833
8,620
8,423
8,241
8,075
7,923
7,788
7,668
7,563
7,473
7,399
7,341
7,298
7,270
7,258
7,246
7,233
7,221
7,209
4-13
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Projecting Vehicle Use in Future Years
The estimates of annual miles driven by passenger cars and light trucks at each age
were also adjusted to reflect projected future growth in average use for vehicles of all ages.
Increases in the average number of miles cars and trucks are driven each year have been an
important source of historical growth in total car and light truck use, and are expected to be a
continued source of future growth in total light-duty vehicle travel as well. As an illustration
of the importance of growth in average vehicle use, the total number of miles driven by
passenger cars increased 35 percent from 1985 through 2005, equivalent to a compound
annual growth rate of 1.5 percent.11 During that same time, however, the total number of
passenger cars registered in the U.S. grew by only about 0.3 percent annually." Thus growth
in the average number of miles automobiles are driven each year accounted for the remaining
1.2 percent (= 1.5 percent - 0.3 percent) annual growth in total automobile use.0
In the U.S., overall change in VMT is attributable to factors such as employment rate,
vehicle ownership rates, demographic trends, the cost of driving, and other macroeconomic
factors. Rather than independently developing estimates of these factors, the agencies have
used the DOT Volpe Center NEMSP run which considers many of these factors, as a
benchmark of total VMT levels in each future year. The VMT projections produced by this
NEMS run are highly similar to those shown in AEO 2012 Early Release. The AEO 2012
Early Release Reference Case projection of total car and light truck use and of the number of
cars and light trucks in use suggest that their average annual use will continue to increase
from 2010 through 2035, although at a slower rate of increase than shown in AEO 201 l.q In
calendar year 2030, total VMT projected in AEO 2012 Early Release is 10% lower than that
projected in AEO 2011.
In order to develop reasonable estimates of future growth in the average number of
miles driven by cars and light trucks of all ages, the agencies calculated the average rate of
growth in the mileage schedules necessary for total car and light truck travel to closely
correspond to AEO 2012 Early Release Reference Case. The growth rate in average annual
n A slight increase in the fraction of new passenger cars remaining in service beyond age 10 has accounted for a
small share of growth in the U.S. automobile fleet. The fraction of new automobiles remaining in service to
various ages was computed from R.L. Polk vehicle registration data for 1977 through 2005 by the agency's
Center for Statistical Analysis.
0 See supra note k below.
p This is the version of NEMS that is used in AEO 2012 Early Release, and modified by the Volpe center to hold
new vehicle fuel economy constant after 2016. See TSD 1 for additional details. This version produces VMT
estimates that are highly similar to those in the AEO 2012 Early Release
q The agencies note that VMT growth has slowed, and because the impact of VMT is an important element in
our benefit estimates, we will continue to monitor this trend to see whether this is a reversal in trend or
temporary slowdown. See the 2009 National Household Travel Survey (http://nhts.ornl.gov/2009/pub/stt.pdf)
and National Transportation Statistics
(http://www.bts.gov/publications/national transportation statistics/html/table 04 09.html')
4-14
-------
Economic and Other Assumptions Used in the Agencies' Analysis
car and light truck use produced by this calculation is approximately 0.6 percent per year/
When the 0.6% annual growth rate is combined with the MY 2010 base sales projection (TSD
1), as well as the VMT, and survival schedules derived for this rule (previously discussed in
sections 4.2.3 and 4.2.4) the estimated total vehicle usage closely approximates that contained
in AEO 2012 ER (section 4.2.4.2). In the agencies' respective modeling, a growth rate is
applied to the mileage figures reported in Table 4-4 (after adjusting vehicle populations for
expected vehicle survival rates) to estimate average annual mileage during each calendar year
analyzed and during the expected lifetimes of model year 2017-25 cars and light trucks8
While EPA used this aggregate approach, accounting for all factors that influence
reference case VMT in a single annual growth factor of 0.6%, NHTSA separated the changing
cost of driving into a second factor, and therefore used a secular growth rate of 0.5%. We
discuss the agencies' two approaches in more detail below.
In the NHTSA analysis, the elasticity of annual vehicle use with respect to fuel cost
per mile was applied to the percentage change in cost-per-mile travel between each future
year's vehicle and the cost per mile of a vehicle that was the same age in calendar year 2008.
Thus, the NHTSA method of modeling the rebound effect captures changes both in fuel cost
relative to calendar year 2008 and in the fuel consumption rates relative to that year, and
inherently assumes the same response to changes in fuel price and fuel efficiency.
Percent difference in VMT = (rebound effect * (FCPM2008- FCPMCAFE Aitemative)/FCPM20o8)
Where FCPM = fuel cost per mile
EPA developed the reference case VMT using the single growth factor discussed
above; this single growth factor reflects driver responsiveness to changes in projected fuel
prices and fuel efficiency, and other factors consistent with the AEO 2012 ER Reference
Case. To develop EPA's policy case VMT, EPA applied the elasticity of annual vehicle use
with respect to fuel cost per mile corresponding to the 10 percent fuel economy rebound effect
used in this analysis (i.e., an elasticity of annual vehicle use with respect to fuel cost per mile
driven of-0.10; see Section 4.2.5) was applied to the percentage change in cost-per-mile
travel between each future year's vehicle under a policy case and a reference case in the same
year. In other words, if the per mile fuel cost of a MY 2025 vehicle under the policy case was
30% less than its counterpart under the reference case, the change in VMT would be 3%.*
r It was not possible to estimate separate growth rates in average annual use for cars and light trucks, because of
the significant reclassification of light truck models as passenger cars discussed previously.
s As indicated previously, a vehicle's age during any future calendar year is uniquely determined by the
difference between that calendar year and the model year when it was produced.
Under the equation: percent difference in VMT = (rebound effect * (FCPMreference case - FCPMpollcy
case)/FCPMreference case) and the rebound effect = 10%. A 30% change in fuel costs, multiplied by a 10% rebound
effect would result in 3% additional driving.
4-15
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Thus, in the EPA analysis, the rebound effect only captures the impact of the EPA program
relative to the reference case standards. Reference case changes in the cost of fuel or fuel
consumption rates are handled separately. In other words, the change in VMT relative to the
reference case is proportional to
Percent difference in VMT = (rebound effect * (FCPMreference case - FCPMpoilcy caSe)/FCPMreference case)
Where FCPM = fuel cost per mile
As a result of the difference between the two approaches of capturing the future
impact of fuel prices and fuel efficiency on VMT, the agencies also differ on how they
ensured consistency with growth in total vehicle use across the entire fleet (including older
vehicles already in the population that are not impacted by this rule). EPA uses the 0.6%
annual estimate of secular VMT growth directly in the OMEGA model. By contrast, since the
NHTSA model considers the effects of changes in fuel cost per mile since the 2009 NHTS as
the reference point of the fuel economy rebound effect, in order to avoid double counting the
effect of changes in fuel cost per mile, NHTSA uses a growth factor of 0.5%. NHTSA
separated the growth rate because of its need for consistent results among the alternative
scenarios and baselines it considered in this rulemaking. For the primary case, these
approaches yield highly similar estimates of VMT schedules, and consequently of total VMT
(see Table 4-5).
Thus, the agencies each made adjustments to vehicle use to account for projected
changes in future fuel prices, fuel efficiency, and other factors that influence growth in
average vehicle use during each future calendar year. Because the effects of fuel prices and
other factors influencing growth in average vehicle use differ for each year, these adjustments
result in different VMT schedules for each future model year. The net impact resulting from
these adjustments is continued growth over time in the average number of miles that vehicles
of each age are driven, although at slower rates than those observed from 1985 - 2005. u
4.2.4.1 VMT equationv
The following equation summarizes in mathematical form the adjustments that are
made to the values of average miles driven by vehicle age derived from the 2009 NHTS to
derive the estimates of average miles driven by vehicles of each model year during future
u Observed aggregate VMT in recent years has actually declined (about 0.4% per year over the past decade), but
it is unclear if the underlying cause is general shift in behavior or a response to a set of temporary economic
conditions.
v While both agencies applied the VMT calculation described above in the NPRM, for the final rule, in the EPA
baseline calculation, the rebound effect is in effect embedded in the growth rate. Under the regulatory
alternatives, the rebound effect is based solely on the percentage increase in fuel economy over the relevant
baseline model year. NHTSA continued to follow the NPRM approach because of its requirement to produce an
Environmental Impact Statement for the rule, and the need for consistent results among the alternative scenarios
it considers.
4-16
-------
Economic and Other Assumptions Used in the Agencies' Analysis
calendar years that are used in this analysis.
Where:
Vy = Average miles driven in the base calendar year (from NHTSA analysis of 2009 NHTS data) by a
vehicle of age y during the base calendar year
Gl = Growth Rate
YS = Years since the base calendar year
R= Elasticity of VMT with respect to FCPM (-0.10). Note that, for EPA, this value is zero in the
reference case since EPA's Gl already incorporates impacts on VMT due to changes in FCPM.
FCPM^ = Fuel cost per mile of a vehicle of age y in calendar year x
FCPMt y = For NHTSA, the fuel cost per mile of a vehicle of age y in calendar year 2008. For EPA, this
variable is identical to FCPMX y in the reference case, and in the policy case this variable represents the
fuel cost per mile of a reference case vehicle of age y in calendar year t
For NHTSA, the base calendar year is 2008, for EPA 2009.
In turn, fuel cost per mile of an age y vehicle in calendar year x is determined by the
following equation, which can be extended for any number of fuels:
Where:
ECy= Electricity consumption of age y vehicle (in KWh) per mile
EPX = Electricity Price (in $ per KWh) during calendar year x
GCy = Gasoline Consumption of age y vehicle (in gallons) per mile
OP* = Gasoline Price (in $ per gallon) during calendar year x
DCy = Diesel Consumption of age y vehicle (in gallons) per mile
DPX = Diesel Price (in $ per gallon) during calendar year x
The NHTSA and EPA models project slightly different fuel costs per mile for vehicles
affected by the proposed standards, because of the different structures of the respective
agencies' programs and the different technologies projected by each agency's model to be
used by vehicle manufacturers to comply with each program. Over the entire lifetimes of
those vehicles, however, the agencies' estimates of the number of miles they are expected to
be driven differ by about 1% for cars and less than 1% for light trucks. For comparison,
Table 4-5 presents the agencies' estimates of the average number of miles driven by model
year 2021 and 2025 cars and light trucks at over their estimated average lifetimes.
Table 4-5 Survival Weighted Per-Vehicle Reference VMT used in the Agencies' Analyses
EPA
NHTSA
MY 2021
Cars
204,161
206,768
Light
Trucks
218,399
218,812
MY 2025
Cars
209,037
211,795
Light
Trucks
223,688
223,865
4-17
-------
Economic and Other Assumptions Used in the Agencies' Analysis
4.2.4.2 Comparison to other VMT Projections
As a check on their estimates of vehicle use, the agencies compared the forecasts of
aggregate car and light truck VMT derived using the procedure described in preceding
sections to the AEO 2012 Early Release reference case forecast of light duty VMT (see Figure
4-1). Using the MY 2010 baseline, which includes the AEO 2012 ER fleet projection, the
aggregate VMT projected in this analysis is within 1% of the AEO 2012 Light Duty
projections over the time period 2017-2035.12 If AEO VMT is linearly extrapolated at the
average growth rate of the period 2017-2035, the agencies' estimates remain within 2% of this
projection through 2050. EPA's VMT estimates are compared to the AEO projection in the
chart below, but based on the similarity of VMT schedules, is indicative of both agencies'
analysis.w
VMT
CO
4,000
3,800
3,600
3,400
3,200
3,000
2,800
2,600
2,400
2,200
2,000
•Reference AEO
2012
2017
2022 2027
Calendar Year
2032
Figure 4-1 Comparison of AEO and Projected VMT
4.2.5 Accounting for the fuel economy rebound effect
The rebound effect refers to the increase in vehicle use that results when an increase in
fuel efficiency lowers the cost per mile of driving, which can encourage people to drive more.
Because this additional driving results in some fuel consumption and emissions, it results in
smaller fuel savings and emissions reductions than would otherwise have resulted from the
final standards. Thus the magnitude of the rebound effect is one determinant of the actual fuel
savings and emission reductions that are likely to result from adopting stricter fuel economy
"'See note p above.
4-18
-------
Economic and Other Assumptions Used in the Agencies' Analysis
or GHG emissions standards, and is an important parameter affecting EPA's and NHTSA's
evaluation of standards for future model years.x
The fuel economy rebound effect is measured directly by estimating the change in
vehicle use, often expressed in terms of vehicle miles traveled (VMT), that results from a
change in vehicle fuel efficiency/ However, analysts commonly measure the rebound effect
indirectly, by estimating the change in vehicle use that results from a change in fuel cost per
mile driven, which depends on both vehicle fuel efficiency and fuel prices.z When expressed
as positive percentages, the elasticities of vehicle use with respect to fuel efficiency or per-
mile fuel costs give the percentage increase in vehicle use that results from a one percent
increase in fuel efficiency, or a one percent reduction in fuel cost per mile. For example, a 10
percent rebound effect means that a 10 percent increase in fuel efficiency or a 10 percent
decrease in fuel cost per mile is expected to result in a 1 percent increase in vehicle use.
The fuel economy rebound effect for light-duty vehicles has been the subject of a large
number of studies since the early 1980s. Although these studies have reported a wide range
of estimates of its exact magnitude, they generally conclude that a significant rebound effect
occurs when the cost per mile of driving decreases.aa The most common approach to
estimating its magnitude has been to analyze household survey data on vehicle use, fuel
consumption, fuel prices (often obtained from external sources), and other variables that
influence travel demand. Other studies have relied on annual aggregate U.S. data. Finally,
more recent studies have used annual data from individual states.*
x The rebound effect discussed in this section refers solely to the effect of increased fuel efficiency on vehicle
use, which has traditionally been referred to as the "fuel economy rebound effect." More recently, some authors
have referred to the fuel economy rebound effect as the "VMT rebound effect," which helps distinguish it from
other rebound effects that could potentially impact the fuel savings and emissions reductions from our standards
such as the "indirect rebound effect," which occurs when buyers of vehicles with improved fuel economy spend
money they save on fuel purchases to buy other products and services that consume or use energy. The
discussion in this section exclusively addresses the fuel economy rebound effect as traditionally defined, and
uses this term throughout. The agencies received one comment on the proposed rulemaking suggesting that the
agencies should attempt to quantify the indirect rebound effect; see preamble III.H.4 for a discussion of this
topic.
y 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.
z 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.
aa Some studies estimate that the long-run rebound effect is significantly larger than the immediate response to
increased fuel efficiency. Although their estimates of the adjustment period required for the rebound effect to
reach its long-run magnitude vary, this long-run effect could be more appropriate for evaluating the fuel savings
and emissions reductions resulting from stricter standards that would apply throughout the lifetime of future
model year vehicles.
bb In effect, these studies treat U.S. states as a data "panel" by applying appropriate estimation procedures to data
consisting of each year's average values of these variables for the separate states.
4-19
-------
Economic and Other Assumptions Used in the Agencies' Analysis
This section surveys these previous studies, summarizes recent work on the rebound
effect,13 and explains the basis for the 10 percent rebound effect EPA and NHTSA are using
in this rulemaking.
4.2.5.1 Summary of historical literature on 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 this
rule will affect future driving behavior. Therefore, the more recent studies have been
described in more detail in Section 4.2.5.2 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 table.
Table 4-6 Estimates of the Rebound Effect Using U.S. Aggregate Time-Series Data on Vehicle Travel1
AUTHOR
(YEAR)
Mayo & Mathis
(1988)
Gately (1992)
Greene (1992)
Jones (1992)
Schimek (1996)
SHORT-RUN
22%
9%
Linear 5-1 9%
Log -linear 13%
13%
5-7%
LONG-RUN
26%
9%
Linear 5-1 9%
Log -linear 13%
30%
21-29%
TIME PERIOD
1958-84
1966-88
1957-89
1957-89
1950-94
1 Source: Sorrell and Dimitropolous (2007) table 4.6.
Table 4-7 Estimates of the Rebound Effect Using U.S. State Level Data1
AUTHOR
(YEAR)
Haughton & Sarkar
(1996)
Small and Van Bender
(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
1 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) discussed in section 4.2.5.2.
4-20
-------
Economic and Other Assumptions Used in the Agencies' Analysis
While studies using national (Table 4-6) and state level (Table 4-7) data have found
relatively consistent long-run estimates of the rebound effect, household surveys display more
variability (Table 4-8). 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 residential density on vehicle use from that of fuel prices, since households in urban
areas are likely to face higher fuel prices.14
Table 4-8 Estimates of the Rebound Effect Using U.S. Survey Data1
Author
(year)
Goldberg (1996)
Greene, Kahn, and
Gibson (1999a)
Pickrell & Schimek
(1999)
Puller & Greening (1999)
West (2004)
Short-Run
0%
49%
87%
Long-Run
23%
4-34%
Time Period
CES 1984-90
EIA RTECS
1979-1994
NPTS 1995
Single year
CES 1980-90
Single year, cross-sectional
CES 1997
Single year
1 Source: Sorrell and Dimitropolous (2007) table 4.8 and the agencies' addition of Pickrell & Schimek (1999).
It is important to note that some of these studies actually quantify the price elasticity
of gasoline demand (e.g., Puller & Greening15) or the elasticity of VMT with respect to the
price of gasoline (e.g., Pickrell & Schimek), 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). This topic is discussed in more detail in Section
4.2.5.2.
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.cc Finally,
00 Six of the household survey studies evaluated in Table 4-9 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, David L., and Patricia S. Hu, "The Influence of the Price of
4-21
-------
Economic and Other Assumptions Used in the Agencies' Analysis
one recent study using state-level data concludes that the rebound effect varies directly in
response to changes in personal income and the degree of urbanization of U.S. cities, as well
as fuel costs.
In order to provide a more comprehensive overview of previous estimates of the
rebound effect, NHTSA reviewed 22 studies of the rebound effect conducted from 1983
through 2005. The agency then performed a detailed analysis of the 66 separate estimates of
the long-run rebound effect reported in these studies, which is summarized in Table 4-9
below.dd As the table indicates, these 66 estimates of the long-run rebound effect range from
as low as 7 percent to as high as 75 percent, with a mean value of 23 percent. Limiting the
sample to 50 estimates reported in the 17 published studies of the rebound effect yields the
same range, but a slightly higher mean estimate (24 percent).
The type of data used and authors' assumption about whether the rebound effect varies
over time have important effects on its estimated magnitude. The 34 estimates derived from
analysis of U.S. annual time-series data produce a mean estimate of 18 percent for the long-
run rebound effect, while the mean of 23 estimates based on household survey data is
considerably larger (31 percent), and the mean of 9 estimates based on state data (25 percent)
is close to that for the entire sample. The 37 estimates assuming a constant rebound effect
produce a mean of 23 percent, identical to the mean of the 29 estimates reported in studies
that allowed the rebound effect to vary in response to fuel prices, vehicle ownership, or
household income.
Table 4-9 Summary Statistics for Estimates of the Rebound Effect
Gasoline on Vehicle Use in Multivehicle Households," Transportation Research Record 988, pp. 19-24 (Docket
EPA-HQ-OAR-2010-0799); Hensher, David A., Frank W. Milthorpe, and Nariida C. Smith, "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 (Docket EPA-HQ-OAR-2010-0799); Walls, Margaret A, Alan J. Krupnick,
and H. S. Hood, "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, Rachel, and Don Pickrell,
"Factors Affecting Vehicle Use in Multiple-Vehicle Households," 2009 National Household Travel Survey
Workshop, June 2011, http://onlinepubs.trb.org/onlinepubs/conferences/2011/NHTSl/West.pdf (Docket EPA-
HQ-OAR-2010-0799). The two studies with rebound estimates that decrease with higher household vehicle
ownership are Mannering, Fred L. and Clifford Winston, "A Dynamic Empirical Analysis of Household Vehicle
Ownership and Utilization, Rand Journal of Economics 16:2 (1985), pp. 215-236 (Docket EPA-HQ-OAR-2010-
0799), and Greene, David L., James R. Kahn, and Robert C. Gibson, "Fuel Economy Rebound Effect for
Household Vehicles," The Energy Journal, 20:3 (1999), 1-21 (Docket EPA-HQ-OAR-2010-0799) (note that the
latter showed virtually no difference in the rebound effect as households went from 1 to 2, a moderate decline
from 2 to 3 vehicles, and a slight increase from 3 to 4 vehicles; on balance, the rebound estimate for households
with 4 vehicles was slightly lower than for households with 1 or 2 vehicles).
ddln some cases, NHTSA derived estimates of the overall rebound effect from more detailed results reported in
the studies. For example, where studies estimated different rebound effects for households owning different
numbers of vehicles but did not report an overall value, the agency computed a weighted average of the reported
values using the distribution of households among vehicle ownership categories.
4-22
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Category of Estimates
All Estimates
Published Estimates
U.S. Time-Series Data
Household Survey Data
Pooled U.S. State Data
Constant Rebound Effect (1)
Variable Rebound Effect: (1)
Number of
Studies
22
17
7
13
2
15
10
Number of
Estimates
66
50
34
23
9
37
29
Range
Low
7%
7%
7%
9%
8%
7%
10%
High
75%
75%
45%
75%
58%
75%
45%
Distribution
Median
22%
22%
14%
31%
22%
20%
23%
Mean
23%
24%
18%
31%
25%
23%
23%
Std. Dev.
14%
14%
9%
16%
14%
16%
10%
4.2.5.2 Summary of recent studies and analyses of the rebound effect
More recent studies published between 2007 and 2010 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.ee 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 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
vary over time.16 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 most
recent five year period (1997-2001), 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 6 percent.17
Finally, when the Small methodology was used to project the future rebound effect, estimates
of the rebound effect throughout 2010-2030 were below 6 percent given a range of future
gasoline price and income projections.
18
In 2010, Hymel, Small and Van Dender extended the Small and Van Dender model by
adding congestion as an endogenous variable.19 Although controlling for congestion
significantly increased their estimates of the rebound effect, Hymel, Small and Van Dender
ee 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)). Note that two studies discussed in this section, Small and Van
Dender (2007) and Hymel, Small, and Van Dender (2010), find that the rebound effect is more strongly
dependent on income than on fuel costs. A third study, Greene (2012), did not directly test the effect of fuel
costs on the rebound effect, but found evidence supporting the effect of income. Several other studies have
shown that the rebound effect rises with household vehicle ownership (see section 4.2.5.1), which has generally
increased with income.
4-23
-------
Economic and Other Assumptions Used in the Agencies' Analysis
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 in 2008-2009 under contract with EPA further
appears to support the theory that the magnitude of the rebound effect is declining over time
90
and may be as low as zero. 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 Bender's prior finding. 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 showed that the hypothesis of equal elasticities for gasoline prices and fuel efficiency
can be rejected. In spite of this result, Greene also tested Small and Van Bender'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 9 percent in 2030.
Since there has been little variation in fuel efficiency in the data over time, isolating
the impact of fuel efficiency 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 efficiency 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 gasoline prices than to changes in fuel efficiency itself. Recent work conducted
by Kenneth Gillingham included an estimate of the elasticity of VMT with respect to the price
of gasoline of-0.17, while his corresponding estimate of the elasticity of VMT with respect to
fuel economy was only 0.05.21 While this research pertains specifically to California, this
finding suggests that the common assumption that consumers respond similarly to changes in
gasoline prices and changes in fuel efficiency may overstate the magnitude of the rebound
effect. The agencies stated in the draft TSD accompanying the NPRM that more research is
needed in this area, and sought comment on this aspect of the rebound effect. No comments
were received on this specific issue.
Another question discussed by Gillingham is whether consumers actually respond the
same way to an increase in the cost of driving compared to a decrease in the cost of driving.
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 & Gately and Sentenac-
Chemin have shown that demand for transportation fuel is asymmetric.22'23 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 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.24 Furthermore, Dargay & Gately and Sentenac-Chemin find evidence that
consumers respond more to a large shock than a small, gradual change in fuel prices. Since
4-24
-------
Economic and Other Assumptions Used in the Agencies' Analysis
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. Although these types of asymmetric responses have been noted at the aggregate
level on oil and gasoline consumption, little research has been done on these same phenomena
in the context of changes in vehicle fuel efficiency and the resulting rebound effect. The
agencies stated in the draft TSD accompanying the NPRM that more research in this area is
also important, and sought comment on this aspect of the rebound effect. No comments were
received on this specific issue.
Other recent studies came to our attention after we finalized our estimate of the
rebound effect used in the analysis for our final rules.25 We will examine these and other new
studies on this topic for future rulemakings.
4.2.5.3 Basis for rebound effect used by EPA and NHTSA in this rule
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.
Nevertheless, NHTSA requires a single point estimate for the rebound effect as an input to its
analysis, 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 was used for this final
rule (i.e.., we assume a 10 percent decrease in fuel cost per mile from our standards would
result in a 1 percent increase in VMT). EPA uses a range of 0-20 percent for sensitivity
testing, while NHTSA uses 5-20 percent.
As Table 4-6, Table 4-7, Table 4-8, and Table 4-9 indicate, the 10 percent figure is on
the low end of the range reported in previous research. However, some recent research -
particularly that conducted by Hymel, Small and Van Dender, Small and Van Dender, and
Greene - reports evidence that the magnitude of the rebound effect is likely to be declining
over time. Furthermore, for the reasons described in section 4.2.5.2, historical estimates of
the rebound effect may overstate the effect of a gradual decrease in the cost of driving due to
our standards.
As a consequence, the agencies concluded that a value on the low end of the historical
estimates reported in Table 4-6, Table 4-7, Table 4-8, and Table 4-9 is likely to provide a
more reliable estimate of its magnitude during the future period spanned by the agencies'
analyses of the impacts of this rulemaking. The 10 percent estimate lies within the 10-30
percent range of estimates for the historical rebound effect reported in most research, and at
the upper end of the 5-10 percent range of estimates for the future rebound effect reported in
the recent studies by Small and Greene. 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.
In their proposed rules, the agencies sought comment and new data on alternative
methods for estimating the rebound effect over the period that our rulemaking would go into
effect. In particular, the agencies sought comment and data on the potential that the rebound
effect could be lower than the estimates in the literature if drivers respond more to changes in
4-25
-------
Economic and Other Assumptions Used in the Agencies' Analysis
fuel prices than fuel efficiency, price rises than price decreases, and price shocks than gradual
price changes (as discussed in section 4.2.5.2). EPA also sought comment and data on the
rebound effect for consumers driving vehicles powered by grid electricity. We believe more
research on these topics is important. During the public comment period, the agencies did not
receive any comments on these topics and the few comments we did receive on the fuel
economy rebound effect did not provide persuasive new evidence. Hence, EPA and NHTSA
have elected to continue to use the 10 percent estimate of the rebound effect in the analyses
supporting this final rulemaking. The agencies will review the estimate of the rebound effect
again for any future rulemakings based on the best available information at that time.
4.2.6 Benefits from additional driving
The increase in travel associated with the rebound effect produces additional benefits
to vehicle owners, which reflect the value to drivers and other vehicle occupants 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 costs drivers incur plus the consumer surplus they receive from the additional
accessibility it provides. As evidenced by the fact that drivers elect to make more frequent or
longer trips when the cost of driving declines, the benefits from this added travel exceed
drivers' added outlays for the fuel consumed. The amount by which the benefits from this
increased driving travel exceed its increased fuel costs measures the net benefits drivers
receive from the additional travel, usually are referred to as increased consumer surplus.
The agencies' analysis estimates the economic value of the increased consumer
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 consumer surplus from
additional vehicle use represents a small fraction of this benefit.
4.2.7 Added costs from increased vehicle use
While it provides some benefits to drivers, increased vehicle use associated with the
rebound effect also contributes to increased traffic congestion, motor vehicle accidents, and
highway noise. Depending on how the additional travel is distributed over 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. 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.
Increased vehicle use due to the rebound effect may also increase the costs associated
with traffic accidents. Drivers may take account of the potential costs they (and their
passengers) face from the possibility of being involved in an accident when they decide to
4-26
-------
Economic and Other Assumptions Used in the Agencies' Analysis
make additional trips. However, they probably do not consider all of the potential costs they
impose on occupants of other vehicles and on pedestrians when accidents occur. Thus any
increase in these "external" accident costs must be considered as another cost of additional
rebound-effect driving. Like increased delay costs, any increase in these external accident
costs caused by added driving is likely to depend on the traffic conditions under which it takes
place, since accidents are more frequent in heavier traffic (although their severity may be
reduced by the slower speeds at which heavier traffic typically moves).
Finally, added vehicle use from the rebound effect may also increase traffic noise.
Noise generated by vehicles causes inconvenience, irritation, and potentially even discomfort
to occupants of other vehicles, to pedestrians and other bystanders, and to residents or
occupants of surrounding property. Because these effects are unlikely to be taken into
account by the drivers whose vehicles contribute to traffic noise, they represent additional
externalities associated with motor vehicle use. Although there is considerable uncertainty in
measuring their value, any increase in the economic costs of traffic noise resulting from added
vehicle use must be included together with other increased external costs from the rebound
effect. To estimate the increased external costs caused by added driving due to the rebound
effect, EPA and NHTSA rely on estimates of congestion, accident, and noise costs caused by
automobiles and light trucks developed previously by the Federal Highway Administration.26
NHTSA employed these estimates previously in its analysis accompanying the MY 2011 final
rule, and the agencies jointly applied them in the MYs 2012-2016 rulemaking and in the
proposal to this rulemaking, and the agencies continue to find them appropriate for this final
rule. The values are intended to measure the increases in costs (or "marginal" external costs)
from added congestion, property damages and injuries in traffic accidents, and noise levels
caused by automobiles and light trucks that are borne by persons other than their drivers and
occupants.
Updated to 2010 dollars, FHWA's "Middle" estimates for marginal congestion,
accident, and noise costs caused by automobile use amount to 5.6 cents, 2.4 cents, and 0.1
cents per vehicle-mile (for a total of 8.1 cents per mile), while those for pickup trucks and
vans are 5.1 cents, 2.7 cents, and 0.1 cents per vehicle-mile (for a total of 7.8 cents per
97 ff
mile). ' These costs are multiplied by the mileage increases attributable to the rebound
effect to yield the estimated increases in congestion, accident, and noise externality costs
during future years.
ff The Federal Highway Administration's estimates of these costs agree closely with some other recent estimates.
For example, recent published research conducted by Resources for the Future (RFF) estimates marginal
congestion and external accident costs for increased light-duty vehicle use in the U.S. to be 3.5 and 3.0 cents per
vehicle-mile inyear-2002 dollars. See Ian W.H. Parry and Kenneth A. Small, "Does Britain or the U.S. Have
the Right Gasoline Tax?" Discussion Paper 02-12, Resources for the Future, 19 and Table 1 (March 2002).
Available at http://www.rff.org/rff/Documents/RFF-DP-02-12.pdf (last accessed May 30, 2012).
4-27
-------
Economic and Other Assumptions Used in the Agencies' Analysis
4.2.8 Petroleum and energy security impacts
The final standards for MYs 2017-2025 will reduce fuel consumption and GHG
emissions in light-duty vehicles, which will result in improved fuel efficiency and, in turn,
help to reduce U.S. petroleum imports. A reduction of U.S. petroleum imports reduces both
financial and strategic risks caused by potential sudden disruptions in the supply of imported
petroleum to the U.S. This reduction in the expected future economic costs associated with
these risks provides a measure of value of improved U.S. energy security resulting from lower
petroleum imports. This section summarizes the agencies' estimates of U.S. oil import
reductions and energy security benefits of the program. Additional discussion of this issue
can be found in Section III.H and Section IV.C.3 of the preamble.
4.2.8.1 Impact on U.S. petroleum imports
In 2011, U.S. petroleum import expenditures represented 16 percent of total U.S.
imports of all goods and services.2 '29 In 2011, the United States imported 45 percent of the
petroleum it consumed30, while the transportation sector accounted for 70 percent of total U.S.
petroleum consumption.31 These figures compare to approximately 37 percent of U.S.
petroleum supplied by imports and 55 percent of total petroleum consumed by the nation's
transportation sector during 1975.32 Requiring improved fuel economy and lower-GHG
vehicle technology in the U.S. is expected to lower U.S. petroleum imports.
Based on analysis of historical and projected future variation in U.S. petroleum
consumption and imports, EPA and NHTSA estimate that approximately 50 percent of the
reduction in fuel consumption resulting from adopting improved GHG emission and fuel
efficiency standards is likely to be reflected in lower U.S. imports of refined fuel, while the
remaining 50 percent is expected to be reflected in reduced domestic fuel refining.gg Of this
latter figure, 90 percent is anticipated to reduce U.S. imports of crude petroleum for use as a
refinery feedstock, while the remaining 10 percent is expected to reduce U.S. domestic
production of crude petroleum. Thus, on balance, each gallon of fuel saved as a
consequence of our final standards is anticipated to reduce total U.S. imports of petroleum by
0.95 gallons."
88 Differences in forecasted annual U.S. imports of crude petroleum and refined products among the Reference,
High Oil Price, and Low Oil Price scenarios analyzed in EIA's Annual Energy Outlook 2011 range from 35-74
percent of differences in projected annual gasoline and diesel fuel consumption in the U.S. These differences
average 53 percent over the forecast period spanned by AEO 2011.
14 Differences in forecasted annual U.S. imports of crude petroleum among the Reference, High Oil Price, and
Low Oil Price scenarios analyzed in EIA's Annual Energy Outlook 2011 range from 67-104 percent of
differences in total U.S. refining of crude petroleum, and average 90 percent over the forecast period spanned by
AEO 2011.
11 This figure is calculated as 0.50 + 0.50*0.9 = 0.50 + 0.45 = 0.95.
4-28
-------
Economic and Other Assumptions Used in the Agencies' Analysis
4.2.8.2 Background on U.S. energy security
U.S. energy security is broadly defined as protecting the U.S. economy against
circumstances that threaten significant short- and long-term increases in energy costs or
interruptions in energy supplies. Most discussions of U.S. energy security focus on the
economic costs of U.S. dependence on oil imports, and particularly on U.S. reliance on oil
imported from potentially unstable sources. In addition, oil exporters have the ability to raise
the price of oil by exerting monopoly power through the mechanism of a cartel, the
Organization of Petroleum Exporting Countries (OPEC). These factors contribute to the
vulnerability of the U.S. economy to episodic oil supply shocks and price spikes. In 2010,
total U.S. imports of crude oil, including those from OPEC nations as well as other sources,
were $271 billion (in 2010$)33 (see Figure 4-2).
Figure 4-2 U.S. Expenditures on Crude Oil from 1970 through 2010jj
U.S. Expenditures on Crude Oil
700
i—i—i—i—i—i—i—i—r
1970 1975 1980
i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r~
1985 1990 1995 2000 2005
Year
2010
A significant effect of the MY 2017-2025 fuel economy and GHG standards (as well
as the MY 2012-2016 light-duty vehicle standards and the MY 2014-2018 standards for
Source for historical data: EIA Annual Energy Review, various editions. For recent historical and forecasted
data: EIA Annual Energy Outlook (AEO) 2011 Reference Case.
4-29
-------
Economic and Other Assumptions Used in the Agencies' Analysis
medium- and heavy-duty vehicles) will be to reduce consumption of transportation fuels in
the U.S. This will in turn reduce U.S. oil imports, which lowers both financial and strategic
risks associated with potential disruptions in supply or sudden increases in the price of
petroleum. For this rulemaking, an "oil import premium" approach is utilized to estimate
energy security-related costs of importing petroleum into the U.S. Specifically, the oil import
premium measures the expected economic value of costs that are not reflected in the market
price of petroleum, and that are expected to change in response to an incremental change in
the level of U.S. oil imports.
4.2.8.3 Methodology used to estimate U.S. energy security benefits
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 study is included as part of the docket for this rulemaking.34
When conducting the analysis for EPA and NHTSA for purposes of analyzing our
final standards, 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. import demand on the world oil price and on OPEC market power (i.e.., the "demand" or
"monopsony" costs); and (2) the risk of reductions in U.S. economic output and disruption of
the U.S. economy caused by sudden disruptions in the supply of imported oil to the U.S. (i.e.,
"macroeconomic disruption and adjustment costs"). Costs associated with U.S. military
expenditures to help secure stable oil supply from volatile regions of the world were not
included in this analysis, because attributing costs for military operations to specific missions
or activities is difficult and the majority of the literature indicates that it is uncertain if merely
reducing (rather than entirely eliminating) reliance on imported oil would lead to measurable
changes in U.S. military expenditures (as discussed further).
For this analysis, ORNL estimated energy security premiums by incorporating the
AEO 2012 Early Release oil price forecasts and market trends, which was the most recent
data available at the time the analyses for the final rules were conducted. Energy security
premiums for the years 2020, 2025, 2030, and 2035 and beyond are presented in Table 4-11,
as well as a breakdown of the components of the energy security premiums for each of these
years. AEO 2012 Early Release projects energy market trends and values out only to 2035.
The agencies assume that energy security premium estimates post-2035 will remain constant,
consistent with a flat extrapolation of oil prices from AEO 2012 after 2035.
^ AEO 2012 Early Release forecasts energy market trends and values only to 2035. The energy security
premium estimates post-2035 were assumed to be the 2035 estimate.
4-30
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Based on the ORNL analysis, the total oil security premium initially declines slightly
through 2020, and then gradually rises over the future as a result of changing factors such as
the world oil price, global supply/demand balances, U.S. oil imports and consumption, and
U.S. GDP (i.e., the size of economy at risk to oil shocks). The principal factors are steadily
rising world oil prices and a growing economy, but other effects interact. From 2020 to 2030,
the macroeconomic disruption and adjustment component rises by 14 percent. This is over a
period where projected average real world oil prices rise 10 per cent and U.S. GDP, the size
of the economy potentially at risk to oil shocks, grows 28 percent. U.S. oil import quantities
decline through 2020, but are steady thereafter through 2035, while total domestic oil
consumption still rises modestly (by 3 percent) despite higher prices. The value share of oil in
GDP stays fairly high; it is still at 3.9 percent by 2030 (vs. 4.5 percent in 2020).
The components of the energy security premiums and their values are discussed
below, as well as how we generally applied them in our respective analyses of the final
standards. Section III.H and Section IV.C.3 of the preamble contains a detailed discussion of
how the monopsony and macroeconomic disruption/adjustment components were treated in
the analysis of our final standards.
Table 4-10 Energy Security Premiums in Selected Years (lOlOS/Barrel)1
2020
2025
2030
2035+
Monopsony
$10.02
($3. 35 -$17.09)
$9.77
($3.25 -$16.69)
$9.28
($3. 10 -$18. 03)
$9.73
($3.24 -$16.68)
Macroeconomic
Di sruption/Adj ustment
Costs
$7.63
($3.71 -$11.00)
$8.26
($4.03 -$11.92)
$8.77
($4.33 - $12.60)
$9.46
($4.72 -$13.61)
Total
$17.64
($9.83 - $25.00)
$18.03
($10.15 -$25.47)
$18.05
($10.29 - $25.20)
$19.19
($10.94 - $26.78)
1 The main values represent the mid-point of the ranges of the values presented in the parentheses.
The Defour Group commented that there is no relationship between the energy
security benefits of the U.S. and reduced oil consumption by the U.S., since the world
economies are all tied together, thus calling into question estimates of the energy security
benefits of these rules. Moreover, the Defour Group believes there is too much uncertainty in
generating energy security premiums, and asserted that the energy security premiums are not
a credible approach to providing estimates of energy security benefits of the rules.
The EPA sponsored an extensive peer review of the methodology on which the energy
security benefits for these rules is based.35 The methodology of estimating the monopsony
and macroeconomic effects for estimating the energy security benefits of particular actions,
4-31
-------
Economic and Other Assumptions Used in the Agencies' Analysis
policies, and rules has been well documented and is well accepted by the energy security
community.36 Thus, the agencies continue to use the current methodology for estimating the
monopsony and macroeconomic effects for estimating the energy security benefits of our
rules.
4.2.8.4 Monopsony Effect
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. purchases a sufficiently large percentage of foreign oil supplies, it exercises
"monopsony power" in the global petroleum market. This 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 to the U.S. of reducing U.S. oil purchases, due both to reductions in
overall energy consumption in transportation and substitution of transportation fuels derived
from non-petroleum sources is the potential decrease in the total dollar value of U.S. crude oil
purchases. Because lower U.S. oil purchases reduce the price paid for each barrel, the decline
in the total cost of U.S. petroleum purchases exceeds the savings that would result if the
global price for oil remained unchanged. The amount by which it does so - which reflects the
effect of U.S. monopsony power over the world oil price - represents the "monopsony" effect
of reduced U.S. petroleum consumption.
This "monopsony" effect can be readily illustrated with an example. If the U.S.
imports 10 million barrels per day at a world oil price of $50 per barrel, its total daily bill for
oil imports is $500 million. If a decrease in U.S. imports to 9 million barrels per day causes
the world oil price to drop to $49 per barrel, the daily U.S. oil import bill drops to $441
million (9 million barrels times $49 per barrel). While the world oil price declines by only $1,
the resulting decrease in oil purchases equals $59 million per day ($500 million minus $441
million). This is equivalent to an incremental savings of $59 for each barrel by which U.S. oil
imports decline ($59 million per day divided by 1 million barrels per day), or $10 more than
the newly-decreased world price of $49 per barrel.
This additional $10 per barrel reduction in the "monopsony premium" represents the
incremental external benefits to the U.S. associated with the reduction in import payments
beyond the savings that would occur if prices remained unchanged. Of course, this additional
benefit arises only to the extent that reduction in U.S. oil imports actually affects the world oil
price. ORNL estimates this component of the energy security benefit in 2025 to be $9.77
/barrel by which U.S. petroleum imports are reduced, with a range of $3.25 - $16.69/barrel.n
Notwithstanding the discussion above, the agencies do not, in fact, include this component of
the energy security premium as part of the benefit estimates of our final rules, since it is a
11 "Estimating the U.S. Oil Security Premium for the Proposed 2017-2025 Light -Duty Vehicle GHG/Fuel
Economy Rule", Paul N. Leiby, Oak Ridge National Laboratory (ORNL), 2012
4-32
-------
Economic and Other Assumptions Used in the Agencies' Analysis
transfer between the U.S. and oil exporting countries, whose potential climate change
damages are accounted for in the agencies' estimate of the social cost of carbon, as explained
further in section 4.2.8.7 below and in the preamble Sections III.H.8 and IV.C.3.
4.2.8.5 Macroeconomic Disruption and Adjustment Effect
The second component of the oil import premium, the "macroeconomic disruption and
adjustment cost premium", arises from the effect of U.S. oil imports on the expected cost of
disruptions in oil supply and resulting increases in oil prices. 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, further expanding the transfer of U.S. wealth to foreign
producers, and (2) it can lead to macroeconomic contraction, dislocation and losses in Gross
Domestic Product (GDP). ORNL estimates the composite estimate of these two factors that
comprise the macroeconomic disruption/adjustment costs premium to be $8.26 /barrel in
2025, with a range of $4.03-11.92/barrel of imported oil reduced. 3? This component of the
energy security premium is included in the agencies' estimate of the benefits of the final
standards. See more discussion of how the agencies account for the energy security benefits of
the rules in Section III.H.8, and Section IV.C.3.
During oil price shocks, the higher price of imported oil causes increased payments for
imports from the U.S to oil exporters. This increased claim on U.S. economic output is a loss
to the U.S. that is separate from and additional to any reduction in economic output due to the
shock. The increased oil payments during shocks are counted as a loss to the degree that the
expected price increase is not anticipated and internalized by oil consumers.
Secondly, macroeconomic losses during price shocks reflect both losses in aggregate
economic output and "allocative" losses. The former are reductions in the level of output that
the U.S. economy can produce by fully utilizing its available resources, while the latter stem
from temporary dislocation and underutilization of available resources due to the shock, such
as labor unemployment and idle plant capacity. The aggregate output effect, a reduction in
"potential" economic output, will persist as long as the price for oil remains elevated. Thus
its magnitude depends on the extent and duration of any disruption in the world supply of oil,
since these factors determine the extent of the resulting increase in prices for petroleum
products, as well as whether and how rapidly these prices return to their pre-disruption levels.
In addition to the aggregate contraction, there are "allocative" or "adjustment" costs
associated with dislocations in energy markets. Because supply disruptions and resulting
price increases occur suddenly, empirical evidence shows they also impose additional costs on
businesses and households for adjusting their use of petroleum and other productive factors
more rapidly than if the same price increase had occurred gradually. Dislocation effects
include the unemployment of workers and other resources during the time period required for
their inter-industry or interregional reallocation, as well as pauses in capital investment due to
uncertainty. These adjustments temporarily reduce the level of economic output that can be
achieved even below the "potential" output level that would ultimately be reached once the
economy's adaptation to higher petroleum prices was complete. The additional costs imposed
on businesses and households for making these adjustments reflect their limited ability to
4-33
-------
Economic and Other Assumptions Used in the Agencies' Analysis
adjust prices, output levels, and their use of energy, labor and other inputs quickly and
smoothly in response to rapid changes in prices for petroleum products.
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, only the change in these expected costs from potential disruptions
that results from the policy 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 in substituting other energy sources or inputs for petroleum use.
In summary, the steps needed to calculate the disruption or security premium are: (1)
determine the likelihood of an oil supply disruption in the future; (2) assess the likely impacts
of a potential oil supply disruption on the world oil price; (3) assess the impact of the oil price
shock on the U.S. economy (in terms of import costs and macroeconomic losses); and (4)
determine how these costs are likely to change with the level of U.S. oil imports. The
reduction in the expected value of costs and other macroeconomic losses that results from
lower oil imports represents the macroeconomic and adjustment cost portion of the oil import
premium.
4.2.8.6 Cost of existing U.S. energy security policies
The last often-identified component of the full economic costs of U.S. oil imports is
the costs to the U.S. taxpayers of existing U.S. energy security policies. The two primary
components of this cost are (1) the expenses associated with maintaining a U.S. military
presence - in part to help secure a stable oil supply - in potentially unstable regions of the
world; and (2) costs for maintaining the U.S. Strategic Petroleum Reserve (SPR). The SPR is
the largest stockpile of government-owned emergency crude oil in the world. Established in
the aftermath of the 1973-74 oil embargo, the SPR provides the U.S. a response option should
price increases triggered by 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.
The agencies recognize that potential national and energy security risks exist due to
the possibility of tension over oil supplies. Much of the world's oil and gas supplies are
located in countries facing social, economic, and demographic challenges, thus making them
even more vulnerable to potential local instability. For example, in 2010 just over 40 percent
of world oil supply came from OPEC nations, and the AEO 2012 projections through 2030
4-34
-------
Economic and Other Assumptions Used in the Agencies' Analysis
show no decline in this share.mm Approximately 28 percent of global supply is from Persian
Gulf countries alone. As another measure of concentration, of the 137 countries/principalities
that export either crude oil or refined petroleum product, the top 12 have recently accounted
for over 55 percent of exports.38 Eight of these countries are members of OPEC, and Russia is
ninth.11" In a market where even a 1-2 percent supply loss raises prices noticeably, and where
a 10 percent supply loss could lead to a significant price shock, this regional concentration is
of concern. Historically, the countries of the Middle East have been the source of eight of the
ten major world oil disruptions39 with the ninth originating in Venezuela, an OPEC member.
Because of U.S. dependence on oil, the military could be called on to protect energy
resources through such measures as securing shipping lanes from foreign oil fields to avoid
economic losses in the U.S. As Lieutenant General (Ret.) Richard Zilmer, commander of
U.S. coalition forces in Anbar province in Iraq in 2006-2007, testified at the Philadelphia
public hearing in support of the proposed standards: "better gas mileage is simply a matter of
national security."40 Lt. Gen. (Ret.) Zilmer contributed to a report of the Center for Naval
Analyses (CNA) that discussed the implications of oil import reductions and energy
security.41 The report focused on changes in the American transportation sector, in terms of
fuel efficiency, alternative fuels, and transportation habits that would be needed in order for
the U.S. economy to have enough resilience to sustain a drastic disruption in oil supply.
Among its findings and recommendations, the report states that "[t]he federal government fuel
economy standards have proven to be effective at increasing efficiency and reducing the use
of oil... These standards should be supported and strengthened as a means of making our
nation more secure." The report states that "[t]he benefits of efficiency are so obvious and
sizeable that it is amazing to consider how or why our country has failed to insist on (or at
least incentivize) it up to now." Finally, the report states "[w]hile our study focuses on
alternative fuels, we repeatedly found the best and most strategically promising alternative to
be efficiency."
Part of the goals of a U.S. military presence in the Persian Gulf is to avoid the impacts
oil price shocks from a supply cut-off on the U.S. economy. Although CNA did not conduct
an economy-wide analysis of an oil supply shock, it did consider the impact of such a shock
on one industrial sector that is heavily dependent on petroleum: the trucking transportation
industry. CNA then considered a 100% disruption in the flow of oil, lasting 30 days in the
Strait of Hormuz. They estimated that such a disruption would have caused losses of $3.3
billion or 2.9 percent of the U.S. trucking industry's output in 2009. According to CNA, this
disruption would have caused 37,500 truckers to lose their jobs. This analysis concludes with
"[i]f the U.S. - and this industry in particular - could reduce its use of petroleum by 30
percent, the effect of such supply disruptions would be nearly zero." Although CNA's report
focused on the trucking sector, the agencies believe that these findings are relevant to this rule
since both the heavy-duty and light-duty vehicles in the U.S. are highly dependent upon
petroleum.
m "DOE/EIA AEO2012, Table 21. International Liquids Supply and Disposition Summary".
1 The other three are Norway, Canada, and the EU, an exporter of product.
4-35
-------
Economic and Other Assumptions Used in the Agencies' Analysis
It is CNA's view that there are several other strategically important reasons for
maintaining a significant military presence in the Middle East beyond protecting oil routes.
Therefore, CNA does not necessarily believe that reduced oil consumption would
automatically lead to the return of troops stationed in the region.
Moreover, the military itself is heavily dependent on oil. To maintain such military
effectiveness and flexibility, the Department of Defense identified in the Quadrennial Defense
Review that it is "increasing its use of renewable energy supplies and reducing energy
demand to improve operational effectiveness, reduce greenhouse gas emissions in support of
U.S. climate change initiatives, and protect the Department from energy price fluctuations."42
The Department of the Navy has also stated that the Navy and Marine Corps rely far too much
on petroleum, which "degrades the strategic position of our country and the tactical
performance of our forces. The global supply of oil is finite, it is becoming increasingly
difficult to find and exploit, and over time cost continues to rise."43 In remarks given to the
White House Energy Security Summit on April 26, 2011, Deputy Security of Defense
William J. Lynn, III noted the direct impact of energy security on military readiness and
flexibility. According to Deputy Security Lynn, "Today, energy technology remains a critical
element of our military superiority. Addressing energy needs must be a fundamental part of
• 1 • 1 • n44
our military planning.
The agencies' analysis of energy security benefits from reducing U.S. oil imports did
not include an estimate of potential reductions in costs for maintaining a U.S. military
presence to help secure stable oil supply from potentially vulnerable regions of the world
because attributing military spending to particular missions or activities is difficult. Many
commenters in both written comments and at the agencies' public hearings expressed their
belief that these standards will have significant benefits for U.S. energy and national security.
A number of commenters, including consumer advocacy and environmental organizations,
organizations representing labor, and state and local governments, as well as energy security
advocates and numerous private individuals, felt that the agencies should quantify, to the
extent possible, a military component of the energy security benefits associated with this
rulemaking. These commenters felt that although they understand that the agencies would
have difficulties in determining a point estimate of the energy security benefits from reduced
military costs as a result of the rule, that even ranges would be useful. The American
Petroleum Institute commented that military expenditures will not likely change with a
reduction in U.S. oil imports, and therefore should not be included in the assessment of this
rulemaking.
However, the agencies have examined methodologies for estimating the military
component of the energy security benefits of our rules and have faced two major challenges:
"attribution" and "incremental" analysis. The attribution analysis challenge is to determine
which military programs and expenditures can properly be attributed to oil supply protection,
rather than to some other objective. The incremental analysis challenge is to estimate how
much the supply protection costs might vary if U.S. oil use is reduced or eliminated.
However, the agencies have reviewed a number of newer studies that attempt to overcome
these challenges.45
4-36
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Most commonly, analysts in recent studies 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. Some recent studies
seek to update, and in some cases significantly improve, the rigor of analysis. At the low end
of the range, the Council on Foreign Relations takes the view that substantial foreign policy-
related military missions will remain over the next 20 years, even without the oil security
mission. Alternatively, Delucchi and Murphy46 sought to deduct from the cost of Persian
Gulf military programs the costs associated with defending other U.S. interests (that is,
interests other than providing more stable domestic oil supply and price to the U.S. economy).
Excluding an estimate of cost for missions in the Persian Gulf unrelated to oil, and excluding
costs for providing military protection for other countries' oil import security, Delucchi and
Murphy estimated military costs for all U.S. domestic oil interests of between $18 and $59
billion in 2004.
In another recent study, RAND47 considered force reductions and cost savings that
could be achieved if oil security were no longer a consideration. 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, RAND concluded that $75-$91 billion, or 12-15 per cent of
the U.S. defense budget in 2009 could be reduced if U.S. dependence on imported oil were
eliminated entirely. However, the study also concludes that the reduction in military costs
from a partial reduction in the U.S. dependence on imported oil would be minimal. In another
study, Stern48 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
aircraft carrier deployment. For the 1976-2007 time frame, Stern estimated an average
military cost of $212 billion per year and $500 billion for 2007 alone that could be potentially
reduced with lower oil imports.
Although these recent studies provide significant, useful insights into the military
components of U.S. energy security, they do not provide enough substantive analysis to
develop a robust methodology for quantifying the military components of energy security for
this rulemaking. Even for studies that provide insight into the attribution of specific missions
to the objective of securing international oil production and distribution, they provide little
guidance on the degree to which incremental reductions in the U.S. dependence on imported
oil would reduce or eliminate those missions or programs. Thus, while the agencies plan to
continue to review newer studies and literature to better estimate the military components of
U.S. energy security benefits, for this rulemaking the agencies continue to exclude military
cost components in our quantified energy security benefits. To summarize, the agencies have
been unable to calculate the monetary benefit that the United States will receive from the
improvements in national security expected to result from our standards.
An additional potential component of the full economic costs of oil imports is the
costs of building and maintaining the SPR. The SPR is clearly related to U.S. oil use and
4-37
-------
Economic and Other Assumptions Used in the Agencies' Analysis
imports. Indeed, a stated purpose of the Energy Policy Conservation Act is "to provide for
the creation of a Strategic Petroleum Reserve capable of reducing the impact of severe energy
supply interruptions," a provision enacted following the 1973-74 Arab oil embargo.00
However, these costs have not varied historically in response to changes in U.S. oil import
levels. Thus although the influence of the SPR on oil price increases resulting from a
disruption of U.S oil imports is reflected in the ORNL estimate of the macroeconomic and
adjustment cost component of the oil import premium, potential changes in the cost of
maintaining the SPR associated with variation in U.S petroleum imports are excluded.
4.2.8.7 Total Energy Security Benefits
Much of 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
social cost of carbon (SCC) value, the question arises: how should the energy security
premium be measured from a global perspective? Monopsony benefits largely represent a
reduction in payments by consumers of petroleum products in the United States to foreign oil
producers that result from a decrease in the world oil price as the U.S. decreases its petroleum
consumption.
Although a reduction in these payments clearly represents a benefit to the U.S. when
considered from a domestic perspective, it represents an exactly offsetting loss to petroleum-
producing countries. Given the purely redistributive nature of this monopsony effect when
viewed from a global perspective, it is excluded in the energy security benefits calculations
for this program. The Union of Concerned Scientists recommended that the monopsony
benefits of the rules be included in the agencies' overall estimates of the energy security
benefits of their respective rules, since it is a benefit to the U.S. The agencies continue to
view energy security from a global perspective, and therefore exclude monopsony benefits to
the U.S. since these benefits are offset by losses to foreign oil producers.
In contrast, the other portion of the energy security premium, the U.S. macroeconomic
disruption and adjustment cost that arises from U.S. petroleum imports, does not have
offsetting impacts outside of the U.S., and is thus included in the energy security benefits
estimated for this program. Thus, the agencies have included only the macroeconomic
disruption portion of the energy security benefits to estimate the monetary value of the total
energy security benefits of this program. The agencies have calculated energy security in very
specific terms, as the reduction of both financial and strategic risks caused by potential sudden
disruptions in the supply of imported petroleum to the U.S. Reducing the amount of oil
imported reduces those risks, and thus increases the nation's energy security.
The energy security analysis conducted for this rulemaking estimates that the world
price of oil will fall modestly in response to lower U.S. demand for refined fuel. One
' See 42 U.S.C section 6201 (2) and Center for Auto Safety v. NHTSA. 739 F. 2d 1322, 1324 (D.C. Cir. 1986).
4-38
-------
Economic and Other Assumptions Used in the Agencies' Analysis
potential result of this decline in the world price of oil would be an increase in the
consumption of petroleum products, particularly outside the U.S. In addition, other fuels
could be displaced from the increasing use of oil worldwide. 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. An increase in
the consumption of petroleum products particularly outside the U.S., could lead to a modest
increase in emissions of greenhouse gases, criteria air pollutants, and airborne toxics from
their refining and use. However, lower usage of, for example, displaced coal would result in a
decrease in greenhouse gas emissions. Therefore, any assessment of the impacts on GHG
emissions from a potential increase in world oil demand would need to take into account the
impacts on all portions of the global energy sector. The agencies' analyses have not
attempted to estimate these effects.
4.2.9 Air pollutant emissions
Car and light truck use, fuel refining, and fuel distribution and retailing also generate
emissions of certain criteria air pollutants, including carbon monoxide (CO), hydrocarbon
compounds (usually referred to as "volatile organic compounds," or VOC), nitrogen oxides
(NOX), fine particulate matter (PM^.s), and sulfur dioxide (802). Due to regulatory structure,
emissions of most of these pollutants are associated with the number of vehicle miles driven,
rather than with the quantity of fuel consumed. Sulfur dioxide is an exception, as described
below.
While reductions in fuel refining and distribution that result from lower fuel
consumption will reduce emissions of criteria pollutants, additional vehicle use associated
with the rebound effect and additional electricity generation to meet the demand of plug-in
electric vehicles will increase emissions of these pollutants. Thus the net effect of stricter fuel
efficiency and GHG standards on total emissions of each criteria pollutant depends on the
relative magnitudes of reduced emissions during fuel refining and distribution, increases in
emissions from additional vehicle use, and changes in electricity generation (increases due to
EV/PHEVs and decreases due to reduced gasoline production at refineries). Because the
relationship between the emission rates in each sector (emissions per gallon refined of fuel,
mile driven, or kwh generated) is different for each criteria pollutant, the net effect of
increases in fuel efficiency and GHG standards on total emissions of each pollutant differs.
4.2.9.1 Emissions of criteria air pollutants
EPA and NHTSA calculated the net changes in domestic emissions of each criteria
pollutant by summing the increases in its emissions projected to result from increased vehicle
use, changes in electricity production, and the reductions in emissions anticipated to result
from lower domestic fuel refining and distribution. As indicated previously, the effect of
adopting improved fuel efficiency and GHG standards on total emissions of each criteria
pollutant depends on the relative magnitudes of the resulting reduction in emissions from fuel
refining and distribution, and the increase in emissions from additional vehicle use.
4-39
-------
Economic and Other Assumptions Used in the Agencies' Analysis
4.2.9.2 Vehicles
For the analysis of criteria emissions in this rule, EPA and NHTSA estimate 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 862, the agencies 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.
As in the MY 2012-2016 rulemaking and the proposal, the relevant emission rates
were estimated by U.S. EPA using the most recent version of the Motor Vehicle Emission
Simulator (MOVES2010a).49 The downstream emission rates are unchanged from the
proposal, and no comments were received on the use of the MOVES model or its
configuration. 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. The MOVES modeling conducted for this rule is assuming RFS2 volumes of
renewable fuel volumes in both the "reference case" and the control case.pp The emission
analysis assumed a 10% ethanol fuel supply.qqAs 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 tirewear and crankcase exhaust operations. EPA analysts selected the year
2050 in order to generate emission factors that were representative of lifetime average
emission rates for vehicles meeting the agency's Tier 2 emission standard.^ Separate
estimates were developed for each vehicle type and model year, as well as for each state and
month, in order to reflect the effects of regional and temporal variation in temperature and
other relevant variables on emissions.
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
pp The agencies assume 100 percent E10 in both the reference and control cases, which is a simplifying
assumption that is appropriate to the level of detail necessary for our rulemaking analyses.
qq More discussion on fuel supply and this rule is in Preamble Section III.F
" Because all light-duty emission rates in MOVES20 lOa are assumed to be invariant after MY 2010, a calendar-
year 2050 run produced a full set of emission rates that reflect anticipated deterioration in the effectiveness of
vehicles' emission control systems with increasing age and accumulated mileage for post-MY 2010 vehicles.
4-40
-------
Economic and Other Assumptions Used in the Agencies' Analysis
emission factors for each pollutant. The resulting emission rates represent average values
across the nation, and incorporate typical 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.ss
Emission rates for the criteria pollutant SC>2 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 862. These calculations assumed that
national average gasoline and diesel sulfur levels would remain at current levels, because
there are no current regulations which will change those levels, and we have no expectation
that the market will cause such changes on its own.tt Therefore, unlike many other criteria
pollutants, sulfur dioxide emissions from vehicle use decline in proportion to the decrease in
fuel consumption.
4.2.9.3 Fuel Production and Transport
Emissions of criteria air pollutants also occur during each phase of fuel production and
distribution, including crude oil extraction and transportation, fuel refining, and fuel storage
and transportation. The reduction in emissions during each of these phases depends on the
extent to which fuel savings result in lower imports of refined fuel, or in reduced domestic
fuel refining. To a lesser extent, they also depend on whether reductions in domestic gasoline
refining are reflected in reduced imports of crude oil or in reduced domestic extraction of
petroleum.
Both EPA's and NHTSA's analyses assume that reductions in imports of refined fuel
would reduce criteria pollutant emissions during fuel storage and distribution only.
Reductions in domestic fuel refining using imported crude oil as a feedstock are assumed to
reduce emissions during fuel refining, storage, and distribution, because each of these
activities would be reduced. Finally, reduced domestic fuel refining using domestically-
produced crude oil is assumed to reduce emissions during all phases of fuel production and
distribution.™ This includes reducing emissions from electric generating units that power the
refineries.
ss The national mix of fuel types includes county-level market shares of conventional and reformulated gasoline,
as well as county-level variation in sulfur content, ethanol fractions, and other fuel properties.
Inspection/maintenance programs at the county level account for detailed program design elements such as test
type, inspection frequency, and program coverage by vehicle type and age.
tt These are 30 and 15 parts per million (ppm, measured on a mass basis) for gasoline and diesel respectively,
which produces emission rates of 0.17 grams of SO2 per gallon of gasoline and 0.10 grams per gallon of diesel.
uu In effect, this assumes that the distances crude oil travels to U.S. refineries are approximately the same
regardless of whether it travels from domestic oilfields or import terminals, and that the distances that gasoline
travels from refineries to retail stations are approximately the same as those from import terminals to gasoline
stations.
4-41
-------
Economic and Other Assumptions Used in the Agencies' Analysis
The agencies estimated the reductions in criteria pollutant emissions from producing
and distributing fuel that would occur under alternative fuel efficiency and GHG standards
using emission rates obtained from Argonne National Laboratories' Greenhouse Gases and
Regulated Emissions in Transportation (GREET) model.5°'w The GREET model provides
separate estimates of air pollutant emissions that occur in four phases of fuel production and
distribution: crude oil extraction, crude oil transportation and storage, fuel refining, and fuel
distribution and storage.1™ EPA modified this version of the GREET model to change certain
assumptions about emissions during crude petroleum extraction and transportation, as well as
to update its emission rates to reflect adopted and pending EPA emission standards.51 EPA
also incorporated emission factors for the air toxics estimated in this analysis: benzene, 1,3-
butadiene, acetaldehyde, acrolein, and formaldehyde.52
The agency converted these emission rates from the mass per fuel energy content basis
on which GREET reports them to mass per gallon of fuel supplied using the estimates of fuel
energy content reported by GREET. The resulting emission rates were applied to both EPA's
and NHTSA's estimates of fuel consumption under alternative fuel efficiency standards to
develop estimates of total emissions of each criteria pollutant during fuel production and
distribution. The assumptions about the effects of changes in fuel consumption on domestic
and imported sources of fuel supply discussed above were then employed to calculate the
effects of reductions in fuel use from alternative fuel efficiency and GHG standards on
changes in domestic emissions of each criteria pollutant throughout the fuel supply and
distribution process. Additional discussion of the emission factors related to fuel production
and transport is provided in EPA's RIA.
Electricity Generation
For the NPRM, EPA and NHTSA utilized emission factors from EPA's Integrated
Planning Model (IPM) to assess the increased electricity used for EVs and PHEVs. As
discussed in our respective RIAs, EPA and NHTSA have independently developed updated
emission factors for use in estimating these emissions. Comments on estimation of these
emissions are also discussed in sections III and IV of the preamble, and in the agencies' RIAs.
4.2.9A Estimated values of reducing PM-related emissions in the model year analysis
The agencies' analysis of PM2.s-related benefits over the lifetime of specific model
years uses a "benefit-per-ton" method to estimate selected PM2.5-related health benefits.
vv GREET has been updated since the last major update of the EPA impact spreadsheet, most recently with
GREET 1 2012, released on June 28, 2012. Due to the lead time required for modeling, and the resultant timing
constraints, these updates have not been incorporated in this analysis. The agencies will monitor relevant
developments for future rulemakings.
ww Emissions that occur during vehicle refueling at retail gasoline stations (primarily evaporative emissions of
volatile organic compounds, or VOCs) are already accounted for in the "tailpipe" emission factors used to
estimate the emissions generated by increased light truck use. GREET estimates emissions in each phase of
gasoline production and distribution in mass per unit of gasoline energy content; these factors are then converted
to mass per gallon of gasoline using the average energy content of gasoline.
4-42
-------
Economic and Other Assumptions Used in the Agencies' Analysis
These PM2.s-related benefit-per-ton estimates provide the total monetized human health
benefits (the sum of premature mortality and premature morbidity) of reducing one ton of
directly emitted PM2.s, or one ton of a pollutant that contributes to secondarily-formed PM2.s
(such as NOx and SOx) from a specified source. Ideally, the human health benefits would be
estimated based on changes in ambient PM2.5 concentrations and population exposure, as
determined by complete air quality and exposure modeling. However, conducting such
detailed modeling for the model year analysis was not possible within the timeframe for the
final rule. Note that EPA and NHTSA conducted full-scale photochemical air quality
modeling for the calendar year analysis in 2030. Please refer to Chapter 6.2 of the RIA for a
description of EPA's air quality modeling results and to Chapter 6.3 for a description of the
quantified and monetized PM- and ozone-related health impacts of the FRM.
Due to analytical limitations, the estimated benefit-per-ton values do not include
comparable benefits related to reductions in other ambient concentrations of criteria pollutants
(such as ozone, NO2 or 802) or toxic air pollutants, nor do they monetize all of the potential
health and welfare effects associated with PM2.5 or the other criteria pollutants. As a result,
monetizing PM-related health impacts alone underestimates the benefits associated with
reductions of the suite of non-GHG pollutants that would be reduced by the final standards.
The dollar-per-ton estimates used to monetize reductions in emissions that contribute
to ambient concentrations of PM2.5 are provided in Table 4-11.
Table 4-11 PM2 5-related Benefits-per-ton Values (2010$)a
Year
All Sources'1
SO2
Upstream (Non-EGU)
Sources'1
NOX
Direct PM2 5
Mobile Sources
NOX
Direct PM2 5
Dollar-per-ton Derived from American Cancer Society Analysis (Pope et al, 2002) Using a 3
Percent Discount Rate0
2015
2020
2030
2040
$30,000
$33,000
$38,000
$45,000
$4,900
$5,400
$6,400
$7,600
$230,000
$250,000
$290,000
$340,000
$5,100
$5,600
$6,700
$8,000
$280,000
$310,000
$370,000
$440,000
Dollar-per-ton Derived from American Cancer Society Analysis (Pope et al., 2002) Estimated
Using a 7 Percent Discount Rate0
2015
2020
2030
2040
$27,000
$30,000
$35,000
$41,000
$4,500
$4,900
$5,800
$6,900
$210,000
$230,000
$270,000
$310,000
$4,600
$5,100
$6,100
$7,300
$250,000
$280,000
$330,000
$400,000
Dollar-per-ton Derived from Six Cities Analysis (Laden et al., 2006) Estimated Using a 3
Percent Discount Rate0
2015
2020
2030
2040
$73,000
$80,000
$94,000
$110,000
$12,000
$13,000
$16,000
$19,000
$560,000
$620,000
$720,000
$840,000
$12,000
$14,000
$16,000
$20,000
$680,000
$750,000
$900,000
$1,100,000
Dollar-per-ton Derived from Six Cities Analysis (Laden et al., 2006) Estimated Using a 7
Percent Discount Rate0
2015
2020
2030
2040
$66,000
$72,000
$84,000
$99,000
$11,000
$12,000
$14,000
$17,000
$510,000
$560,000
$650,000
$760,000
$11,000
$12,000
$15,000
$18,000
$620,000
$680,000
$810,000
$960,000
4-43
-------
Economic and Other Assumptions Used in the Agencies' Analysis
a Total dollar-per-ton estimates include monetized PM25-related premature mortality and morbidity endpoints.
Range of estimates are a function of the estimate of PM2 5-related premature mortality derived from either the
ACS study (Pope et al, 2002) or the Six-Cities study (Laden et al., 2006).
b Dollar-per-ton values were estimated for the years 2015, 2020, and 2030. For 2040, EPA extrapolated
exponentially based on the growth between 2020 and 2030.
0 The dollar-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.
d Note that the dollar-per-ton value for SO2 is based on the value for Stationary (Non-EGU) sources; no SO2
value was estimated for mobile sources.
As Table 4-11 indicates, EPA projects that the per-ton values for reducing emissions
of criteria pollutants from both vehicle use and stationary sources such as fuel refineries and
storage facilities will increase over time.™ These projected increases reflect rising income
levels, which are assumed to increase affected individuals' willingness to pay for reduced
exposure to health threats from air pollution. 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 in older age groups with the highest mortality risk.53>yy
For certain PM2.s-related pollutants (such as direct PM2.5 and NOx), EPA estimates
different per-ton values for reducing mobile source emissions than for reductions in
emissions of the same pollutant from stationary sources such as fuel refineries and storage
facilities. These reflect differences in the typical geographic distributions of emissions of
each pollutant by different sources, their contributions to ambient levels of PM2.5, and
resulting changes in population exposure. EPA and NHTSA apply these separate values to its
estimates of changes in emissions from vehicle use and from fuel production and distribution
to determine the net change in total economic damages from emissions of those pollutants.
The benefit per-ton technique has been used in previous analyses, including the 2012-
2016 Light-Duty Greenhouse Gas Rule,54 the Ozone National Ambient Air Quality Standards
(NAAQS) RIA,55 the Portland Cement National Emissions Standards for Hazardous Air
Pollutants (NESHAP) RIA,56 and the final NO2 NAAQS.57 Table 4-12 shows the quantified
and monetized PM2.s-related co-benefits that are captured in these benefit-per-ton estimates,
and also lists other effects that remain un-quantified and are thus excluded from the estimates.
Table 4-12 Human Health and Welfare Effects of PM2 s
Pollutant / Quantified and Monetized Un-quantified Effects
Effect in Primary Estimates Changes in:
PM2.5 Adult premature mortality Subchronic bronchitis cases
Bronchitis: chronic and acute Low birth weight
Hospital admissions: respiratory and Pulmonary function
xx As we discuss in the emissions chapter of EPA's RIA (Chapter 4), the rule would yield emission reductions
from upstream refining and fuel distribution due to decreased petroleum consumption.
yy 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)
4-44
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Pollutant / Quantified and Monetized Un-quantified Effects
Effect in Primary Estimates Changes in:
cardiovascular Chronic respiratory diseases other than
Emergency room visits for asthma chronic bronchitis
Nonfatal heart attacks (myocardial Non-asthma respiratory emergency room
infarction) visits
Lower and upper respiratory illness Visibility
Minor restricted-activity days Household soiling
Work loss days
Asthma exacerbations (asthmatic
population)
Infant mortality
Consistent with the NC>2 NAAQS,ZZ the benefits estimates utilize concentration-
response functions as reported in the epidemiology literature. Readers interested in reviewing
the complete methodology for creating the benefit-per-ton estimates used in this analysis can
consult the Technical Support Document (TSD)58 accompanying the final ozone NAAQS
RIA. Readers can also refer to Fann et al. (2009)59 for a detailed description of the benefit-
per-ton methodology.aaa
As described above, national per-ton estimates were developed for selected
pollutant/source category combinations. The per-ton values calculated therefore apply only to
tons reduced from those specific pollutant/source combinations (e.g., NC>2 emitted from
mobile sources; direct PM emitted from stationary sources). Our estimate of total PM2.5
benefits is therefore based on the total direct PIVb.s and PlV^.s-related precursor emissions
(NOx, SOx, and VOCs) controlled from each source and multiplied by the respective per-ton
values of reducing emissions from that source.
Reductions in PM-related mortality provide the majority of the monetized value in
each benefit-per-ton estimate. Typically, the premature mortality-related effect coefficients
that underlie the benefits-per-ton estimates are drawn from epidemiology studies that examine
two large population cohorts: the American Cancer Society cohort (Pope et al., 2002)60 and
the Harvard Six Cities cohort (Laden et al., 2006).61 The concentration-response (C-R)
function developed from the extended analysis of American Cancer Society (ACS) cohort, as
zz Although we summarize the main issues in this chapter, we encourage interested readers to see benefits
chapter of the NO2 NAAQS for a more detailed description of recent changes to the PM benefits presentation
and preference for the no-threshold model.
aaa The values included in this report are different from those presented in the article cited above. Benefits
methods change to reflect new information and evaluation of the science. Since publication of the June 2009
article, EPA has made two significant changes to its benefits methods: (1) We no longer assume that a threshold
exists in PM-related models of health impacts, which is consistent with the findings reported in published
research; and (2) We have revised the Value of a Statistical Life to equal $6.3 million (year 2000$), up from an
estimate of $5.5 million (year 2000$) used in the June 2009 report. Please refer to the following website for
updates to the dollar-per-ton estimates: http://www.epa.gov/air/benmap^pt.html
4-45
-------
Economic and Other Assumptions Used in the Agencies' Analysis
reported in Pope et al. (2002), has previously been used by EPA to generate its primary
benefits estimate. The extended analysis of the Harvard Six Cities cohort, as reported by
Laden et al (2006), was published after the completion of the Staff Paper for the 2006 PM2.5
NAAQS and has been used as an alternative estimate in the PM^.sNAAQS RIA and PM2.5 co-
benefits estimates in analyses completed since the PM2.5 NAAQS.
These studies provide logical choices for co-equal anchor points when presenting PM-
related benefits because, while both studies are well designed and peer-reviewed, there are
strengths and weaknesses inherent in each. Although EPA's primary method of
characterizing PM-related premature mortality is to use both studies to generate a co-equal
range of benefits estimates, EPA has chosen to present only the benefit-per-ton value derived
from the ACS study in its summary tables of total Model Year costs and benefits (See
Preamble Section III.H.10 and RIA Chapter 7). This decision was made to provide the reader
with summary tables that are easier to understand and interpret and does not convey any
preference for one study over the other. We note that this is also the more conservative of the
two estimates - PM-related benefits would be approximately 245 percent (or nearly two-and-
a-half times) larger had we used the per-ton benefit values based on the Harvard Six Cities
study instead.
As is the nature of benefits analyses, assumptions and methods evolve over time to
reflect the most current interpretation of the scientific and economic literature. For a period
of time (2004-2008), EPA's Office of Air and Radiation (OAR) valued mortality risk
reductions using a value of statistical life (VSL) estimate derived from a limited analysis of
some of the available studies. OAR arrived at a VSL using a range of $1 million to $10
million (2000$) consistent with two meta-analyses of the wage-risk literature.
The $1 million value represented the lower end of the interquartile range from the
Mrozek and Taylor (2002)62 meta-analysis of 33 studies. The $10 million value represented
the upper end of the interquartile range from the Viscusi and Aldy (2003)63 meta-analysis of
43 studies. The mean estimate of $5.5 million (2000$) was also consistent with the mean
VSL of $5.4 million estimated in the Kochi et al. (2006)64 meta-analysis. However, the
Agency neither changed its official guidance on the use of VSL in rulemakings nor subjected
the interim estimate to a scientific peer-review process through the Science Advisory Board
(SAB) or other peer-review group.
Until updated guidance is available, EPA determined that a single, peer-reviewed
estimate applied consistently best reflects the Science Advisory Board Environmental
Economics Advisory Committee (SAB-EEAC) advice it has received. Therefore, EPA has
decided to apply the VSL that was vetted and endorsed by the SAB in the Guidelines for
Preparing Economic Analyses (U.S. EPA, 2000)65 while they continue efforts to update their
guidance on this issue. This approach calculates a mean value across VSL estimates
bbb In the update of the Economic Guidelines (U.S. EPA, 2011), EPA retained the VSL endorsed by the SAB
with the understanding that further updates to the mortality risk valuation guidance would be forthcoming in the
4-46
-------
Economic and Other Assumptions Used in the Agencies' Analysis
derived from 26 labor market and contingent valuation studies published between 1974 and
1991. The mean VSL across these studies is $6.3 million (2000$). The dollar-per-ton
estimates used in this analysis are based on this revised VSL.000
The benefit-per-ton estimates are subject to a number of assumptions and
uncertainties.
• They do not reflect local variability in population density, meteorology, exposure,
baseline health incidence rates, or other local factors that might lead to an
overestimate or underestimate of the actual benefits of controlling fine particulates in
specific locations. Please refer to Chapter 6 of EPA's RIA for the description of the
agency's quantification and monetization of PM- and ozone-related health impacts for
the final standards.
• 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 engines and other industrial
sources. At the present time, however, no clear scientific grounds exist for supporting
differential effects estimates by particle type.
• This analysis assumes that the health impact function for fine particles is linear within
the range of ambient concentrations under consideration. Thus, the estimates include
health benefits from reducing fine particles in areas with varied initial concentrations
of PM2.5, including both regions that are in attainment with fine particle standard and
those that do not meet the standard, down to the lowest modeled concentrations.
• There are several health benefits categories that EPA and NHTSA were unable to
quantify due to limitations associated with using benefits-per-ton estimates, several of
which could be substantial. Because NOx and VOC emissions are also precursors to
ozone, changes in NOx and VOC would also impact ozone formation and the health
effects associated with ozone exposure. Benefits-per-ton estimates for ozone 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. Please refer to
Chapter 6.3 of EPA's RIA for a description of the unqualified co-pollutant benefits
associated with this rulemaking.
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 the localized impacts
associated with the rulemaking may vary significantly. Additionally, the atmospheric
near future. The update of the Economic Guidelines is available on the Internet at
http://yosemite.epa.gov/ee/epa/eed.nsf/pages/Guidelines.html/$file/Guidelines.pdf.
000 This value differs from the Department of Transportation's most recent estimate of the value of preventing
transportation-related fatalities, which is $6.1 million when expressed in today's (2011) dollars.
4-47
-------
Economic and Other Assumptions Used in the Agencies' Analysis
chemistry related to ambient concentrations of PM2.5, 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. For this final rule, EPA and
NHTSA conducted national-scale air quality modeling analyses for 2030 to analyze the
impacts of the standards on PM2.5, ozone, and selected air toxics.
4.2.10 Reductions in emissions of greenhouse gases
Emissions of carbon dioxide and other greenhouse gases (GHGs) occur throughout the
process of producing and distributing transportation fuels, as well as from fuel combustion
itself. By increasing fuel efficiency and thus reducing the volume of fuel consumed by
passenger cars and light trucks, the standards will reduce GHG emissions generated by fuel
use, as well as throughout the fuel supply cycle. Lowering these emissions is likely to slow
the projected pace and reduce the ultimate extent of future changes in the global climate, thus
reducing future economic damages that changes in the global climate are otherwise expected
to cause. Further, by reducing the probability that climate changes with potentially
catastrophic economic or environmental impacts will occur, lowering GHG emissions may
also result in economic benefits that exceed the resulting reduction in the expected future
economic costs caused by gradual changes in the earth's climatic systems. Quantifying and
monetizing benefits from reducing GHG emissions is thus an important step in estimating the
total economic benefits likely to result from establishing improved fuel efficiency and GHG
standards.
4.2.10.1 Estimating reductions in GHG emissions
NHTSA estimates emissions of carbon dioxide (€62) from passenger car and light
truck use by multiplying the number of gallons of each type of fuel (gasoline and diesel) they
are projected to consume with each alternative CAFE standard in effect by the quantity or
mass of CC>2 emissions released per gallon of fuel consumed. EPA directly calculates CO2
emissions from the projected CC>2 emissions of each vehicle under the CC>2 standards. This
calculation assumes that the entire carbon content of each fuel is ultimately converted to CO2
emissions during the combustion process. The weighted average CC>2 content of certification
gasoline is estimated to be 8,887 grams per gallon, while that of diesel fuel is estimated to be
approximately 10,200 grams per gallon. For details, please see EPA's and NHTSA's
respective RIAs. These same methods were used in the proposal and no comments were
received.
Although carbon dioxide emissions account for nearly 95 percent of total GHG
emissions that result from fuel combustion during vehicle use, emissions of other GHGs are
potentially significant as well because of their higher "potency" as GHGs than that of CO2
itself. EPA and NHTSA estimated the increases in emissions of methane (CH/j) and nitrous
oxide (N2O) from additional vehicle use by multiplying the increase in total miles driven by
cars and light trucks of each model year and age by emission rates per vehicle-mile for these
GHGs. These emission rates, which differ between cars and light trucks as well as between
gasoline and diesel vehicles, were estimated by EPA using MOVES 2010a.
4-48
-------
Economic and Other Assumptions Used in the Agencies' Analysis
The MOVES model assumes that the per-mile rates at which cars and light trucks emit
these non-CO2 GHGs are determined by the efficiency of fuel combustion during engine
operation and chemical reactions that occur during catalytic after-treatment of engine exhaust,
and are thus independent of vehicles' fuel consumption rates. Thus MOVES emission factors
for these GHGs, which are expressed per mile of vehicle travel, are assumed to be unaffected
by changes in fuel economy.
Much like criteria pollutants, emissions of GHGs occur during each phase of fuel
production and distribution, including crude oil extraction and transportation, fuel refining,
and fuel storage and transportation. Emissions of GHGs also occur in generating electricity,
which the agencies' analysis anticipates will account for an increased but likely still small
share of energy consumption in the model years subject to the standards. The agencies'
analyses assume that reductions in fuel consumption would reduce global GHG emissions
during all four phases of fuel production and distribution. Unlike criteria pollutants, the
agencies report both domestic and international reductions in GHG emissions. EPA derived
GHG emission rates corresponding to producing and distributing fuel from Argonne National
Laboratories' Greenhouse Gases and Regulated Emissions in Transportation (GREET)
model.66666
For the NPRM, EPA and NHTSA utilized emission factors from EPA's Integrated
Planning Model (TPM) to assess the increased electricity used for EVs and PHEVs. As
discussed in our respective RIAs, EPA and NHTSA have independently developed updated
emission factors for use in estimating these emissions. Comments on estimation of these
emissions are also discussed in sections III and IV of the preamble, and in the agencies' RIAs.
Increases in emissions of non-CO2 GHGs are converted to equivalent increases in CO2
emissions using estimates of the Global Warming Potential (GWP) of methane (CH/j), nitrous
oxide (N2O), and hydrofluorocarbons (HFC-134a). These GWPs are one way of accounting
ddd The four stages are crude oil extraction, crude oil transportation and storage, fuel refining, and fuel
distribution and storage
eee This version of the model was modified, and is discussed in section 4.2.9.1
fff As in the MY 2012-2016 LD rules and in the MY 2014-2018 MD and HD rules, the global warming potentials
(GWP) used in this rulemaking are consistent with the 100-year time frame values in the 2007 Intergovernmental
Panel on Climate Change (IPCC) Fourth Assessment Report (AR4). At this time, the 100-year GWP values
from the 1995 IPCC Second Assessment Report (SAR) are used in the official U.S. GHG inventory submission
to the United Nations Framework Convention on Climate Change (UNFCCC) per the reporting requirements
under that international convention. The UNFCCC recently agreed on revisions to the national GHG inventory
reporting requirements, and will begin using the 100-year GWP values from AR4 for inventory submissions in
the future (United Nations Framework Convention on Climate Change, "Decisions adopted by the Conference of
the Parties: 15/CP. 17 'Revision of the UNFCCC reporting guidelines on annual inventories for Parties included
in Annex I to the Convention'," FCCC/CP/201 l/9/Add.2, Durban, South Africa, December 2011). According to
the AR4, N2O has a 100-year GWP of 298, CH4 has a 100-year GWP of 25, and HFC-134a has a 100-year GWP
of 1430.
4-49
-------
Economic and Other Assumptions Used in the Agencies' Analysis
for the higher radiative forcing capacity and differing lifetimes of methane and nitrous oxide
when they are released into the earth's atmosphere, measured relative to that of CO2 itself.
Because these gases differ in atmospheric lifetimes, their relative damages are not constant
over time. Impacts other than temperature change also vary across gases in ways that are not
captured by GWP. For instance, CC>2 emissions, unlike methane and other greenhouse gases,
contribute to ocean acidification. Methane contributes to health and ecosystem effects arising
from increases in tropospheric ozone, while damages from methane emissions are not offset
by the positive effect of CO2 fertilization. Noting these caveats, the CO2 equivalents of
increases in emissions of these gases are then added to the increases in emissions of CC>2 itself
to summarize the effect of the total increase in CO2-equivalent GHG emissions from vehicle
use.
4.2.10.2 Economic benefits from reducing GHG emissions
NHTSA and EPA have taken the economic benefits of reducing CC>2 emissions (or
avoiding damages from increased emissions) into account in developing the final GHG and
CAFE standards and in assessing the economic benefits of the final standards. Specifically,
NHTSA and EPA have assigned dollar values to reductions in carbon dioxide (€62)
emissions using estimates of the global "social cost of carbon" (SCC). The SCC is an
estimate of the monetized damages associated with an incremental increase in carbon
emissions in a given year. It is intended to include (but is not limited to) changes in net
agricultural productivity, human health, property damages from increased flood risk, and the
value of ecosystem services due to climate change. The SCC is expressed in constant dollars
per additional metric ton of CC>2 emissions occurring during a specific year, and is higher for
more distant future years because the damages caused by an additional ton of emissions
increase with larger concentrations of CC>2 in the earth's atmosphere.
The estimates used in this analysis were developed through an interagency process
that included EPA, DOT/NHTSA, and other executive branch entities, and concluded in
February 2010. The interagency group focused on global SCC values because emissions of
CO2 involve a global externality: greenhouse gases contribute to damages around the world
wherever they are emitted. Consequently, to address the global nature of the climate change
problem, the SCC must incorporate the full (global) damages caused by GHG emissions.
Furthermore, climate change occurs over very long time horizons and represents a problem
that the United States cannot solve independently. The February 2010 SCC Technical
Support Document (SCC TSD) provides a complete discussion of the SCC estimates and the
methods used to develop them.67
We first used these SCC estimates in the benefits analysis for the final joint EPA/DOT
Rulemaking to establish 2012-2016 Light-Duty Vehicle Greenhouse Gas Emission Standards
and Corporate Average Fuel Economy Standards; see the rule's preamble for discussion about
application of the SCC (75 FR 25324; May 7, 2010). We have continued to use these
4-50
-------
Economic and Other Assumptions Used in the Agencies' Analysis
estimates in other rulemaking analyses, including the Greenhouse Gas Emission Standards
and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles (76 FR
57106; 9/15/11). Finally, see also preamble Section III.H.6, Section IV.C.3.1, EPA RIA
Chapter 7.1, and NHTSA RIA VIII.C for discussion about the application of SCC estimates to
this final rule.
The SCC estimates corresponding to assumed values of the discount rate are shown
below in Table 4-13.
Table 4-13 Social Cost of of CO2,2017 - 2050a (in 2010 dollars)
Year
2017
2020
2025
2030
2035
2040
2045
2050
Discount Rate and Statistic
5% Average
$6
$7
$9
$10
$12
$13
$15
$16
3% Average
$26
$27
$31
$34
$37
$41
$44
$47
2.5% Average
$41
$43
$48
$52
$56
$61
$64
$68
3%
95th percentile
$79
$84
$94
$104
$114
$124
$133
$142
a The SCC values apply to emissions occurring during each year
shown (in 2010 dollars), and represent the present value of future damages as
of the year shown.
As Table 4-13 shows, the SCC estimates selected by the interagency group for use in
regulatory analyses range from roughly $6 to about $79 (in 2010 dollars) for emissions
occurring in the year 2017. The first three estimates are based on the average SCC estimated
using different models and reflect discount rates of 5, 3, and 2.5 percent, respectively. The
fourth value is included to represent the possibility of higher-than-expected impacts from
accumulation of GHGs in the earth's atmosphere, and the consequently larger economic
damages. For this purpose, the interagency group elected to use the SCC value for the 95th
percentile at a 3 percent discount rate.
The value that the interagency group centered its attention on is the average SCC
estimate at a 3 percent discount rate, or about $26 per metric ton in 2017. To capture the
uncertainties involved in regulatory impact analysis, however, the group emphasized the
importance of considering the full range of estimated SCC values. As the table shows, the
SCC estimates rise over time because future emissions are expected to produce larger
incremental damages as physical and economic systems become more stressed in response to
greater climatic change; for example, the average SCC at 3 percent increases from about $26
per ton of CO2 in 2017 to approximately $34 per ton of CO2 by 2030.
Reductions in CO2 emissions that are projected to result from lower fuel consumption,
refining, and distribution during each future year are multiplied by the appropriate SCC
estimates for that year, to determine the range of total economic benefits from reduced
4-51
-------
Economic and Other Assumptions Used in the Agencies' Analysis
emissions during that year. For internal consistency, these annual benefits are discounted
back to net present value terms using a discount rate that is consistent with that used to
develop each SCC estimate.
This final rulemaking also responds to comments regarding the valuation of non-CC>2
GHGs and analyzes changes in non-CC>2 GHGs. The 2010 interagency group, however, did
not directly estimate the social cost of non-CC>2 GHGs. One way to approximate the value of
marginal non-CO2 GHG emission reductions in the absence of direct model estimates is to
convert the reductions to CO2-equivalents which may then be valued using the SCC.
Conversion to CO2-e is typically done using the global warming potential (GWP) for the non-
CO2 gas.ggg We refer to this as the "GWP approach."
One potential problem with using temporally aggregated statistics such as GWP is that
the additional radiative forcing from the GHG perturbation is not constant over time and any
differences in temporal dynamics between gases will be lost. This is a potentially
confounding issue given that the social cost of GHGs is based on a discounted stream of
damages that are non-linear in temperature. For example, methane has an expected adjusted
atmospheric lifetime of about 12 years and associated GWP of 25 (TPCC Fourth Assessment
Report (AR4) 100-year GWP estimate). Gases with a relatively shorter lifetime, such as
methane, have impacts that occur primarily in the near term and thus are not discounted as
heavily as those caused by longer-lived gases such as CO2, while the GWP treats additional
forcing the same independent of when it occurs in time. Furthermore, the baseline
temperature change is lower in the near term and therefore the additional warming from
relatively short lived gases will have a lower marginal impact relative to longer lived gases
that have an impact further out in the future when baseline warming is higher. In addition,
impacts other than temperature change also vary across gases in ways that are not captured by
GWP. For instance, CO2 emissions, unlike methane will result in CO2 passive fertilization to
plants.
In short, the GWP-weighted approach will produce social cost estimates that are less
accurate than the directly modeled estimates. A limited number of studies in the published
literature explore these differences. A recent working paper (Marten and Newbold, 2011),
found that the GWP-weighted benefit estimates for CH4 and N2O are likely to be lower than
those that would be derived using a directly modeled social cost of the non-CO2 GHGs for a
variety of reasons.68 hhh The GWP reflects only the integrated radiative forcing of a gas over
100 years. In contrast, the directly modeled social cost differs from the GWP because the
differences in timing of the warming between gases are explicitly modeled, the non-linear
888 The GWP is an aggregate measure that approximates the additional energy trapped in the atmosphere over a
given timeframe from a perturbation of a non-CO2 gas relative to CO2.
^ As discussed in Marten and Newbold, the discount rate influences the relative social cost of a gas, i.e., the
ratio of the social cost of the gas and the social cost of CO2. Methane is a short-lived gas and therefore at higher
discount rates, the relative social cost is higher than at low discount rates. Depending on the discount rate, the
relative social cost of methane ranged from 22 to 41 in 2015, compared to 25 for the AR4 GWP. The relative
social cost of N2O was calculated to be at least 372, much higher than the AR4 value of 298.
4-52
-------
Economic and Other Assumptions Used in the Agencies' Analysis
effects of temperature change on economic damages are included, and rather than treating all
impacts over a hundred years equally, the modeled social cost applies a discount rate but
calculates impacts through the year 2300.
The agencies recognize the importance of considering the economic impacts from
changes to non-CC>2 GHGs under this final rule. Therefore, in the absence of direct model
estimates from the interagency analysis, EPA and NHTSA have used the GWP approach to
estimate the dollar value of this rule's non-CC>2 GHG benefits in a sensitivity analysis; these
estimates are presented for illustrative purposes and therefore not included in the total benefits
estimate for the rulemaking. NHTSA and EPA converted CH4 and N2O emissions to CC>2
equivalents using the GWP of each gas, then multiplied these CO2-equivalent emission by the
interagency social cost of CC>2 estimates. EPA also converted HFC-134a emissions to CC>2
equivalents and applied the social cost of carbon.111 Please see NHTSA's preamble IV.G.4 and
RIA Chapter X and EPA's preamble Section III.H.6 and RIA Chapter 7 for more details about
the agencies' respective sensitivity analyses and results.
4.2.11 Benefits due to reduced refueling time
Direct estimates of the value of extended vehicle range are not available in the
literature, so the agencies instead calculate the reduction in the required annual number of
refueling cycles due to improved fuel economy, and assess the economic value of the
resulting benefits. Chief among these benefits is the time that owners save by spending less
time both in search of fueling stations and in the act of pumping and paying for fuel.
NHTSA conducted an analysis to estimate the benefits associated with reduced
refueling, which both agencies are using in their respective analyses of overall programmatic
costs and benefits, but which the agencies are applying slightly differently. See chapter VIII
of NHTSA's RIA and chapter 7 of EPA's RIA for more details.
111 As in the MY 2012-2016 LD rules and MY 2014-2018 MD and HD rules, the global warming potentials
(GWP) used in this rulemaking are consistent with the 100-year time frame values in the 2007 Intergovernmental
Panel on Climate Change (IPCC) Fourth Assessment Report (AR4). At this time, the 100-year GWP values
from the 1995 IPCC Second Assessment Report (SAR) are used in the official U.S. GHG inventory submission
to the United Nations Framework Convention on Climate Change (UNFCCC) per the reporting requirements
under that international convention. The UNFCCC recently agreed on revisions to the national GHG inventory
reporting requirements, and will begin using the 100-year GWP values from AR4 for inventory submissions in
the future (United Nations Framework Convention on Climate Change, "Decisions adopted by the Conference of
the Parties: 15/CP.17 'Revision of the UNFCCC reporting guidelines on annual inventories for Parties included
in Annex I to the Convention'," FCCC/CP/201 l/9/Add.2, Durban, South Africa, December 2011). According to
the AR4, N2O has a 100-year GWP of 298, CH4 has a 100-year GWP of 25, and HFC-134a has a 100-year GWP
of 1430.
4-53
-------
Economic and Other Assumptions Used in the Agencies' Analysis
4.2.12 Discounting future benefits and costs
Discounting future fuel savings and other benefits is intended to account for the
reduction in their value to society when they are deferred until some future date, rather than
received immediately. The discount rate expresses the percent decline in the value of these
benefits - as viewed from the current perspective - for each year they are deferred into the
future. In evaluating the benefits from alternative increases in fuel economy and GHG
standards for MY 2017-2025 passenger cars and light trucks, EPA and NHTSA consider
discount rates of both 3 and 7 percent per year.
Three percent may be the appropriate rate for discounting future benefits from
increased fuel economy and GHG standards because most or all of vehicle manufacturers'
costs for complying with improved fuel economy and GHG standards are likely to be
reflected in higher sales prices for their new vehicle models. By increasing sales prices for
new cars and light trucks, GHG and CAFE regulations will thus primarily affect vehicle
purchases and other private consumption decisions.
Both economic theory and OMB guidance on discounting indicate that the future
benefits and costs of regulations that mainly affect private consumption should be discounted
at the consumption rate of time preference.69 OMB guidance indicates that savers appear to
discount future consumption at an average real (that is, adjusted to remove the effect of
inflation) rate of about 3 percent when they face little risk about its likely level, which makes
it a reasonable estimate of the consumption rate of time preference.70 Because there is some
uncertainty about the extent to which vehicle manufacturers will be able to recover their costs
for complying with improved fuel economy and GHG standards by increasing vehicle sales
prices, however, the use of a higher percent discount rate may also be appropriate. OMB
guidance indicates that the real economy-wide opportunity cost of capital is the appropriate
discount rate to apply to future benefits and costs when the primary effect of a regulation is
".. .to displace or alter the use of capital in the private sector," and estimates that this rate
currently averages about 7 percent.71 Thus the agencies have employed both 3 and 7 percent
rates to discount projected future benefits and costs resulting from improved fuel economy
and GHG standards for MY 2017-2025 passenger cars and light trucks.
One important exception to these values are the rates used to discount benefits from
reducing CO2 emissions from the years in which reduced emissions occur, which span the
lifetimes of model year 2017-2025 cars and light trucks, to their present values. In order to
ensure consistency in the derivation and use of the SCC estimates of the unit values of
reducing CO2 emissions, the total benefits from reducing those emissions during each future
year are discounted using the same rates that were used to derive the alternative values of
reducing each ton of CO2 emissions (2.5, 3.0, and 5.0 percent).
4.2.13 Additional Costs of Vehicle Ownership
Sales Taxes:
Consumers 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 cost of the
4-54
-------
Economic and Other Assumptions Used in the Agencies' Analysis
program, but 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. The
agencies took the most recent auto sales taxes by state1" and weighted them by population by
state to determine a national weighted-average sales tax of 5.46 percent. The agencies sought
to weight sales taxes by new vehicle sales by state; however, such data were unavailable. It is
recognized that for this purpose, new vehicle sales by state is a superior weighting mechanism
to Census population; in effort to approximate new vehicle sales by state, 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.
Financing Costs:
The agencies considered that 70 percent of new vehicle purchasers take out loans to
finance their purchases. As these costs are transfer payments, they are not included in the
societal cost of the program, but they are included as one of the increased costs to the
consumer for these standards. Using proprietary forecasts available from Global Insight,
estimates of 48-month bank and auto finance company loan rates for years 2017 through
2025 were developed, which - when deflated by Global Insight's corresponding forecasts of
the CPI - range from 3.73% to 5.38%, averaging 5.16 percent over the nine years.mmm In the
construction of this estimate, it was assumed that there will be an equal distribution of bank
and auto finance company loans - an assumption necessitated by the lack of data on the
distribution of the volume of loans between the differing types of creditors. The agencies
opted to adjust future loan rates using the CPI rather than the GDP deflator, as this analysis is
JJJ See http://www.factorywarrantylist.com/car-tax-bv-state.html (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.
^ Bird, Colin. "Should I Pay Cash, Lease or Finance My New Car?",
http://www.cars.com/go/advice/Story .jsp?section=fin&story=should-i-pay-cash&subject=loan-quick-
start&referer=advice&aff=sacbee , July 12, 2011, citing CNW Marketing Research. Accessed 9/27/11.
111 No projections were available for rates of loan terms of 60 months. The agencies compared the historical
difference of 48-month and 60-month loan rates and determined the 48-month rate to be a suitable proxy for the
60-month rate.
mmm Global Insight data are available on a fee basis at http://www.ihs.com/products/global-insight/country-
analysis/us-economic-forecasts.aspx. Analysis of future auto loan rates is based on Global Insight data available
as of March, 2012.
4-55
-------
Economic and Other Assumptions Used in the Agencies' Analysis
intended to facilitate further analysis from the perspective of the consumer, for which the CPI
is the preferred deflation factor.
Insurance Costs:
The agencies considered the rule's 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. A recent study by Quality Planning11"11 provides the average
value of collision plus comprehensive insurance for new vehicles, in 2010$, is $521 ($396 of
which is collision and $125 of which is comprehensive). 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.°°° Using sales volumes
from the Bureau, we determined an average passenger car and an average light truck price
was $27,953 in $2010 dollars.™5
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% 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 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.qqq 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%, 1.82%, 1.75%, 1.64%, and 1.50%, respectively, or 8.57% 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
111111 "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).
000 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
http://www.bea.gov/national/nipaweb/nipa underlying/Table7.2.5s (last accessed May 4, 2012)
ppp 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.
qqqlnsurance 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.
4-56
-------
Economic and Other Assumptions Used in the Agencies' Analysis
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 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
rule.
As discussed earlier, the scope of this analysis is to estimate the increased cost to the
consumer for these standards, not the increase in societal costs or benefits.
4-57
-------
Economic and Other Assumptions Used in the Agencies' Analysis
References:
, Final Rulemaking. Motor Vehicle Emissions Federal Test Procedure Revisions
10/1996. 61 FR at 54852
9 _ ___
EPA, Fuel Economy Labeling of Motor Vehicles: Revisions To Improve Calculation of Fuel
Economy Estimates; Final Rule, 40 CFR Parts 86 and 600, 71 Fed. Reg, 77872 (Dec. 27,
2006). Available at http://www.epa.gov/fedrgstr/EPA-AIR/2006/December/Day-
277a9749.pdf.
3 EPA, Final Technical Support Document: Fuel Economy Labeling of Motor Vehicle
Revisions to Improve Calculation of Fuel Economy Estimates, Office of Transportation and
Air Quality EPA420-R-06-017 December 2006, Chapter II,
http://www.epa.gov/fueleconomy/420r06017.pdf.
4 See 71 FR at 77887, and U.S. Environmental Protection Agency, Final Technical Support
Document, Fuel Economy Labeling of Motor Vehicle Revisions to Improve Calculation of
Fuel Economy Estimates, EPA420-R-06-017, December 2006 for general background on the
analysis. See also EPA's Response to Comments (EPA-420-R-1 1-005) to the 201 1 labeling
rule, page 189.
136 EPA and NHTSA, § 600.210-12 Calculation of fuel economy and CO2 emission values
for labeling. Revisions and Additions to Motor Vehicle Fuel Economy Label. Federal
Register/ Vol. 76, No. 129 / Wednesday,76 Fed. Reg. 39478 (July 6, 2011). § 600.115-11 .
6 EPA. E15 Waiver Decision, EPA420-F- 10-054. October 2010.
7 EPA, Final Technical Support Document: Fuel Economy Labeling of Motor Vehicle
Revisions to Improve Calculation of Fuel Economy Estimates, at 70. Office of Transportation
and Air Quality EPA420-R-06-017 December 2006, Chapter II,
http://www.epa.gov/fueleconomy/420r06017.pdf.
8 75 FRM at 25372
9 OMB Circular A-4, September 17, 2003.
http://www.whitehouse.gov/omb/assets/omb/circulars/a004/a-4.pdf
10 Lu, S., NHTSA, Regulatory Analysis and Evaluation Division, "Vehicle Survivability and
Travel Mileage Schedules," DOT HS 809 952, 8-11 (January 2006). Available at
http://www-nrd.nhtsa.dot.gov/pdf/nrd-30/NCSA/Rpts/2006/809952.pdf (last accessed Sept.
9,2011).
11 FHWA, Highway Statistics, Summary to 1995, Table vm201 at
http://www.fhwa.dot.gov/ohim/summary95/vm201a.xlw , and annual editions 1996-2005,
Table VM-1 at http://www.fhwa. dot.gov/policy/ohpi/hss/hsspubs. htm (last accessed Feb. 15,
2010).
4-58
-------
Economic and Other Assumptions Used in the Agencies' Analysis
12 Energy Information Adminstration, AEO 2011 Final Relase, Light Duty Vehicle Miles
Traveled by Technology Type.
http://www.eia.gov/oiaf/aeo/tablebrowser/#release=AEO2011&subject=0-
AEO2011 &table=51 - AEO2011 ®ion=0-0&cases=ref2011 -d020911 a
13 Sorrell, S. and J. Dimitropoulos, 2007. "UKERC Review of Evidence for the Rebound
Effect, Technical Report 2: Econometric Studies", UKERC/WP/TPA/2007/010, UK Energy
Research Centre, London, October and Greening, L.A., D.L. Greene and C. Difiglio, 2000.
"Energy Efficiency and Consumption - The Rebound Effect - A Survey", Energy Policy, vol.
28, pp. 389-401. (Docket EPA-HQ-OAR-2010-0799)
14 Pickrell, D. and P. Schimek, 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. (Document ID EPA-HQ-OAR-2010-0799-0027)
15 Puller, Steven and Lorna Greening. 1999. "Household Adjustment to Gasoline Price
Change: An Analysis Using Nine Years of U.S. Survey Data," Energy Economics 21(1):37-
52. (Docket EPA-HQ-OAR-2010-0799)
16 Small, K. and K. Van Dender, 2007a. "Fuel Efficiency and Motor Vehicle Travel: The
Declining Rebound Effect." The Energy Journal, vol. 28, no. 1, pp. 25-51. (Docket EPA-HQ-
OAR-2010-0799)
17 Small, K. and K. Van Dender, 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. (Docket EPA-HQ-OAR-2010-0799)
1 &
Report by Kenneth A. Small of University of California at Irvine to EPA, "The Rebound
Effect from Fuel Efficiency Standards: Measurement and Projection to 2030,", June 12, 2009.
(See docket: EPA-HQ-OAR-2010-0799)
19 Hymel, Kent M., Kenneth A. Small, and Kurt Van Dender, "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.
(Docket EPA-HQ-OAR-2010-0799)
20Greene, David, 2012. "Rebound 2007: Analysis of U.S. light-duty vehicle travel statistics,"
Energy Policy, vol. 41, pp. 14-28. (Docket EPA-HQ-OAR-2010-0799)
21 Gillingham, Kenneth. "The Consumer Response to Gasoline Price Changes: Empirical
Evidence and Policy Implications." Ph.D. diss., Stanford University, 2011. (Document ID
EPA-HQ-OAR-2010-0799-0760). The California Air Resources Board submitted this work
for independent academic review in September 2011 and the three reviewers found the results
and conclusions to be valid and reasonable. These reviews are available in Docket EPA-HQ-
OAR-2010-0799 under the title, California Air Resources Board Independent Academic
4-59
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Review of Gillingham, Kenneth. "The Consumer Response to Gasoline Price Changes:
Empirical Evidence and Policy Implications." Ph.D. diss., Stanford University, 2011.
99
Dargay, J.M.,Gately,D.,1997. "The demand for transportation fuels: imperfect price-
reversibility?," Transportation Research PartB 31(1). (Docket EPA-HQ-OAR-2010-0799)
23 Sentenac-Chemin, E.,2012. "Is the price effect on fuel consumption symmetric? Some
evidence from an empirical study," Energy Policy, vol. 41, pp. 59-65. (Docket EPA-HQ-
OAR-2010-0799)
94
Dermot Gately, 1993. "The Imperfect Price-Reversibility of World Oil Demand," The
Energy Journal, International Association for Energy Economics, vol. 14(4), pp. 163-182.
(Docket EPA-HQ-OAR-2010-0799)
9S
Bento, Antonio M., Lawrence H. Goulder, Mark R. Jacobsen, and Roger H. von Haefen,
2009, "Distributional and Efficiency Impacts of Increased US Gasoline Taxes," American
Economic Review, v. 99, pp. 1-37 and Su, Qing, 2012, "A Quantile Regresssion Analysis of
the Rebound Effect: Evidence form the 2009 National Household Transportation Survey in
the United States," Energy Policy, v. 45, pp. 368-377. (Docket EPA-HQ-OAR-2010-0799)
26 FHWA. 1997 Federal Highway Cost Allocation Study;
http://www.fhwa.dot.gov/policy/hcas/fmal/index.htm (last accessed Sept. 9, 2011).
27 Federal Highway Administration, 1997 Federal Highway Cost Allocation Study.,
http://www.fhwa.dot.gov/policy/hcas/fmal/index.htm, Tables V-22, V-23, and V-24 (last
accessed Sept. 9,2011).
9R
http://www.eia.gov/total energy/data/monthly/pdf/secl_l 3.pdf
29http://www.eia.gov/totalenergy/data/monthly/pdf/secl_13.pdf
30 http://www.eia.gov/totalenergy/data/monthly/pdf/sec3_3.pdf
31 http://www.eia.gov/totalenergy/data/monthly/pdf/sec3_3 .pdf
32 U.S. Department of Energy, Annual Energy Review, 2008, Report No. DOE/EIA-
0384/2008 Tables 5.1 and 5.1 c, June 26, 2009
33 http://www.census.gov/foreign-trade/statistics/historical/gands.txt.
34 Leiby, Paul N. "Estimating the Energy Security Benefits of Reduced U.S. Oil Imports,"
Oak Ridge National Laboratory, ORNL/TM-2007/028, Final Report, 2008.
35 ORNL-TM-2007-028_Oil_Import_Premium_revisited_2008Marl4 Rev8.pdf, "Estimating
the U.S. Oil Security Premium for the Proposed 2017-2025 Light -Duty Vehicle GHG/Fuel
Economy Rule", Paul N. Leiby, Oak Ridge National Laboratory. March 2008.
36
Ibid. References are in the back of the document.
4-60
-------
Economic and Other Assumptions Used in the Agencies' Analysis
37 Estimating the U.S. Oil Security Premium for the Proposed 2017-2025 Light -Duty Vehicle
GHG/Fuel Economy Rule", Paul N. Leiby, Oak Ridge National Laboratory (ORNL), July 15
2012
38 Based on data from the CIA, combining various recent years,
https://www.cia.gov/library/publications/the-world-factbook/rankorder/2176rank.html.
39 IE A 2011 "IE A Response System for Oil Supply Emergencies".
40 Transcript of Philadelphia Public Hearing, p. 175 (January 19, 2012).
41 Center for Naval Analyses, "Ensuring America's Freedom of Movement: A National
Security Imperative to Reduce U.S. Oil Dependence" October 2011.
42 U.S. Department of Defense. 2010. Quadrennial Defense Review Report. Secretary of
Defense: Washington, D.C. 128 pages.
43 The Department of the Navy's Energy Goals
(http://www.navy.mil/features/Navy_EnergySecurity.pdf) (Last accessed May 31, 2011).
44 U.S. Department of Defense, Speech: Remarks at the White House Energy Security
Summit. Tuesday, April 26, 2011.
(http://www.defense.gov/speeches/speech.aspx?speechid=1556) (Last accessed May 31,
2011).
45
Leiby, Paul. "Military Costs of Energy Security". 2012.
46 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 (June 2008): 2253-2264.
doi: 10.1016/j.enpol.2008.03.006.
http://linkinghub.elsevier.com/retrieve/pii/S0301421508001262.
47 Crane, Keith, et al. "Does Imported Oil Threaten U.S. National Security?" RAND, 2009.
48 Stern, RJ. "United States cost of military force projection in the Persian Gulf, 1976-2007",
Energy Policy, 2010.
49 EPA MOVES model ver 2010a (http://www.epa.gov/otaq/models/moves/index.htm).
50 Argonne National Laboratories, The Greenhouse Gas and Regulated Emissions from
Transportation (GREET) Model., Version 1.8, June 2007, available at http://greet.es.anl.gov/
(last accessed Sept. 9, 2011).
4-61
-------
Economic and Other Assumptions Used in the Agencies' Analysis
51 Calculation of Upstream Emissions for the Greenhouse Gas (GHG) Vehicle Rule,
Memorandum from Craig A. Harvey to Docket EPA-HQ-OAR-2009-0472, September 14,
2009
52 Calculation of Upstream Emissions for the Greenhouse Gas (GHG) Vehicle Rule,
Memorandum from Craig A. Harvey to Docket EPA-HQ-OAR-2009-0472, September 14,
2009
53 The issue is discussed in more detail in the PM NAAQS RIA from 2006. See U.S.
Environmental Protection Agency. 2006. Final Regulatory Impact Analysis (RIA) for the
Proposed National Ambient Air Quality Standards for Particulate Matter. Prepared by: Office
of Air and Radiation. October 2006. Available at http://www.epa.gov/ttn/ecas/ria.html.
54 U.S. Environmental Protection Agency (U.S. EPA), 2010. Regulatory Impact Analysis,
Final Rulemaking to Establish Light-Duty Vehicle Greenhouse Gas Emission Standards and
Corporate Average Fuel Economy Standards. Office of Transportation and Air Quality.
April. Available at http://www.epa.gov/otaq/climate/regulations/420rl0009.pdf. EPA-420-R-
10-009
55 U.S. Environmental Protection Agency (U.S. EPA). 2008. Regulatory Impact Analysis,
2008 National Ambient Air Quality Standards for Ground-level Ozone, Chapter 6. Office of
Air Quality Planning and Standards, Research Triangle Park, NC. March. Available at
. EPA-HQ-OAR-2009-
0472-0238
56 U.S. Environmental Protection Agency (U.S. EPA). 2010. Regulatory Impact Analysis:
National Emission Standards for Hazardous Air Pollutants from the Portland Cement
Manufacturing Industry. Office of Air Quality Planning and Standards, Research Triangle
Park, NC. Augues. Available on the Internet at <
http://www.epa.gov/ttn/ecas/regdata/RIAs/portlandcementfmalria.pdf >. EP A-HQ-O AR-
2009-0472-0241
57 U.S. Environmental Protection Agency (U.S. EPA). 2010. Final NO2 NAAQS Regulatory
Impact Analysis (RIA). Office of Air Quality Planning and Standards, Research Triangle
Park, NC. April. Available on the Internet at
http://www.epa.gov/ttn/ecas/regdata/RIAs/FinalNO2RIAfulldocument.pdf. Accessed March
15, 2010. EP A-HQ-O AR-2009-0472-023 7
58 U.S. Environmental Protection Agency (U.S. EPA). 2008. Technical Support Document:
Calculating Benefit Per-Ton estimates, Ozone NAAQS Docket #EPA-HQ-OAR-2007-0225-
0284. Office of Air Quality Planning and Standards, Research Triangle Park, NC. March.
Available on the Internet at . EPA-HQ-OAR-2009-0472-0228
59 Fann, N. et al. (2009). The influence of location, source, and emission type in estimates of
the human health benefits of reducing a ton of air pollution. Air Qual Atmos Health.
Published online: 09 June, 2009. EPA-HQ-OAR-2009-0472-0229
60 Pope, C.A., III, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston.
2002. "Lung Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine Particulate
4-62
-------
Economic and Other Assumptions Used in the Agencies' Analysis
Air Pollution." Journal of the American Medical Association 287:1132-1141. EPA-HQ-
OAR-2009-0472-0263
61 Laden, F., J. Schwartz, F.E. Speizer, and D.W. Dockery. 2006. "Reduction in Fine
Particulate Air Pollution and Mortality." American Journal of Respiratory and Critical Care
Medicine 173:667-672. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press. EPA-HQ-OAR-2009-0472-
1661
62 Mrozek, J.R., and L.O. Taylor. 2002. "What Determines the Value of Life? A Meta-
Analysis." Journal of Policy Analysis and Management 21(2):253-270. EPA-HQ-OAR-2009-
0472-1677
63 Viscusi, V.K., and J.E. Aldy. 2003. "The Value of a Statistical Life: A Critical Review of
Market Estimates throughout the World." Journal of Risk and Uncertainty 27(l):5-76. EPA-
HQ-OAR-2009-0472-0245
64 Kochi, I, B. Hubbell, and R. Kramer. 2006. An Empirical Bayes Approach to Combining
Estimates of the Value of Statistical Life for Environmental Policy Analysis. Environmental
and Resource Economics. 34: 385-406. EPA-HQ-OAR-2009-0472-0235
65 U.S. Environmental Protection Agency (U.S. EPA). 2000. Guidelines for Preparing
Economic Analyses. EPA 240-R-00-003. National Center for Environmental Economics,
Office of Policy Economics and Innovation. Washington, DC. September. Available on the
Internet at
. EPA-
HQ-OAR-2009-0472-0226
66 Argonne National Laboratories, The Greenhouse Gas and Regulated Emissions from
Transportation (GREET) Model, Version 1.8, June 2007, available at
http://www.transportation.anl.gov/software/GREET/index.html (last accessed April 20,
2008).
67 Docket ID EPA-HQ-OAR-2009-0472-114577, Technical Support Document: Social Cost
of Carbon for Regulatory Impact Analysis Under Executive Order 12866, Inter agency
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 (February 2010). Also available at http://epa.gov/otaq/climate/regulations.htm
68 Marten, A. and S. Newbold. 2011. "Estimating the Social Cost of Non-CO2 GHG
Emissions: Methane and Nitrous Oxide." NCEE Working Paper Series #11-01.
http://yosemite.epa.gov/ee/epa/eed.nsf/WPNumber/201 l-01?opendocument. Accessed May
24, 2012.
69 Id.
4-63
-------
Economic and Other Assumptions Used in the Agencies' Analysis
70 Office of Management and Budget, Circular A-4, "Regulatory Analysis," September 17,
2003, 33. Available at http://www.whitehouse.gov/omb/circulars/a004/a-4.pdf (last accessed
July 24, 2009).
71 Id.
4-64
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Chapter 5: Air Conditioning, Off-Cycle Credits, and Other
Flexibilities
5.1 Air conditioning technologies and credits
5.1.1 Overview
Air conditioning (A/C) is virtually standard equipment in new cars and trucks, as over
95% of the new cars and light trucks in the United States are equipped with mobile air
conditioning (or MAC) systems. Given the large penetration of A/C in today's light duty
vehicle fleet, its impact on the amount of energy consumed is significant. In the MYs 2012-
2016 Light-Duty Greenhouse Gas Rule, EPA structured the rule to allow vehicle
manufacturers' to generate credits for improved air conditioner systems in complying with the
CC>2 fleetwide average standards and accounted for these AC improvements in determining
the stringency of the GHG standards. EPA will continue with and expand upon these
provisions, and manufacturers can generate credits for improved performance of both direct
(A/C leakage) and indirect (tailpipe emissions attributable to A/C use) A/C emissions. In
addition, EPA is acting pursuant to its authority under EPCA to allow manufacturers to
generate fuel consumption improvement values for purposes of CAFE compliance based on
the use of A/C efficiency-improving technologies. In the 2012-2016 rule, EPA and NHTSA
did not allow manufacturers to include reductions in fuel consumption resulting from A/C
efficiency improvements) in the CAFE calculations. As was the case in the MYs 2012-2016
rule, manufacturers do not to count reductions in A/C leakage toward their CAFE calculations
since these improvements do not affect fuel economy. In the sections below, the agencies will
first describe the structure of the EPA A/C program, followed by a description of the A/C
program under CAFE.
Through model years 2012-2016, EPA expects that all manufacturers will generate
A/C credits offered (for reduced leakage and improved efficiency) to help come into
compliance with the standards. EPA estimated that there would be significant penetration of
A/C technologies to gain credits, and this was reflected in the stringency of the standards.3
Consistent with the 2008-based fleet definitions, the base level of A/C technologies in 2008
forms the A/C "baseline", and the A/C technologies projected to penetrate to the fleet in 2016
is referred to as the A/C "reference". For this MYs 2017-2025 rule, EPA will maintain the
credit program the amount of credit being determined in relation to the MY 2008 baseline .
The credits should continue to the present rule since without them, a manufacturer utilizing
credits in MY 2016 could suddenly find in MY 2017 that the stringency of the standards are
artificially increased due to discontinued A/C credits. In this chapter, A/C credits are
a NHTSA will also be referencing these efficiency improving A/C technologies in its rule, referring to them as
"fuel consumption improvements."
b Put another way, the MY 2016 GHG standards would remain even if there were no new MY 2017-2025
standards and A/C credits would also continue. Thus, if the AC credits were removed or significantly changed
from these (perpetuated) post-2016 standards, the stringency of those standards would effectively be increased.
5-1
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
calculated from the 2008 baseline vehicle fleet (though there are some changes to the credit
program), while costs are calculated from the 2008 model year based reference fleet. Any
additional A/C credits projected for MYs 2017-2025 are reflected in the stringency of the
standards as described in Section III.C.l of the preamble.
A/C is different from the other technologies described in Chapter 3 of the joint TSD in
several ways. First, most of the technologies described earlier directly affect the efficiency of
the engine, transmission, and vehicle systems. As such, these systems are almost always
active while the vehicle is moving down the road or while being tested on a dynamometer for
the fuel economy and emissions test drive cycles. A/C, on the other hand, is a parasitic load
on the engine that only burdens the engine when the vehicle occupants demand it. Since it is
not tested as a part of the fuel economy and GHG emissions standards compliance test drive
cycles (the A/C system is off while the vehicle is operated on the two test cycles — the FTP
and FIFET — used for compliance purposes) it is referred to as an "off-cycle" effect. There
are many other off-cycle loads that can be switched on by the occupant that affect the engine;
these include lights, wipers, stereo systems, electrical defroster/defogger, heated seats, power
windows, etc. However, these electrical loads individually amount to a very small effect on
the engine (although together they can be significant). The A/C system (by itself) adds a
significant load on the engine (especially on sunny, hot, and/or humid days), resulting in
increased fuel consumption, or "indirect" CC>2 emissions.
There are two mechanisms by which A/C systems contribute to the emissions of
greenhouse gases. The first is through direct leakage of the refrigerant into the air. The
hydrofluorocarbon (FIFC) refrigerant compound currently used in all recent model year
vehicles is R-134a (also known as 1,1,1,2-Tetrafluoroethane, or HFC-134a). Based on the
higher global warming potential of HFCs, a small leakage of the refrigerant has a greater
global warming impact than a similar amount of emissions of some other mobile source
GHGs. R-134a has a global warming potential (GWP) of 1,430. This means that 1 gram of
R-134a has the equivalent global warming potential of 1,430 grams of CO2 (which has a
GWP of 1). In order for the A/C system to take advantage of the refrigerant's
thermodynamic properties and to exchange heat properly, the system must be kept at high
pressures even when not in operation. Typical static pressures can range from 50-80 psi
depending on the temperature, and during operation, these pressures can get to several
hundred psi. At these pressures leakage can occur through a variety of mechanisms. The
refrigerant can leak slowly through seals, gaskets, and even small failures in the containment
of the refrigerant. Through normal use, the rate of leakage may also increase due to wear on
the system components. Leakage may also increase more quickly through rapid component
deterioration such as during vehicle accidents, maintenance or end-of-life vehicle scrappage
(especially when refrigerant capture and recycling programs are less efficient). Small
amounts of leakage can also occur continuously even in extremely "leak-tight" systems by
permeating through hose membranes and seals. This last mechanism is not dissimilar to fuel
permeation through porous fuel lines (and seals). Manufacturers may be able to reduce these
leakage emissions through the implementation of technologies/designs such as leak-tight,
non-porous, durable components. The global warming impact of leakage emissions also can
be addressed by using alternative refrigerants, such as HFO-1234yf, R-744 (CO2), HFC-152a
(R-152a), or other refrigerants under development with lower global warming potentials than
R -134a. Refrigerant emissions can also occur during maintenance and at the end of the
5-2
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
vehicle's life (as well as emissions during the initial charging of the system with refrigerant),
and these emissions are already addressed by the CAA Title VI stratospheric ozone program,
as described below.0
The second mechanism by which vehicle A/C systems contribute to GHG emissions is
through the consumption of additional fuel required to provide power to the A/C system and
from carrying around the weight of the A/C system hardware year-round. These indirect
emissions result from the additional fuel which is required to provide power to the A/C
system (and the additional fuel is converted into CO2 by the engine during combustion).
These increased emissions due to A/C operation can be reduced by increasing the overall
efficiency of the vehicle's A/C system, as described below. The final rules do not provide
credits for the weight of the A/C system, since the incremental increase in CC>2 emissions and
fuel consumption due to carrying the A/C system is directly measured during the normal (2-
cycle) federal test procedure, and is thus already accounted for in the CC>2 tailpipe standard.
EPA's analysis from the MYs 2012-2016 rule indicates that A/C-related indirect
emissions represent about 3.9% of the total greenhouse gas emissions from cars and light
trucks. In this document, EPA will separate the discussion of these two categories of A/C-
related emissions because of the fundamental differences in the emission mechanisms and the
methods of emission control. Refrigerant leakage control is akin in many respects to past
EPA fuel evaporation control programs (in that containment of a fluid is the key feature),
while efficiency improvements are more similar to the vehicle-based control of CC>2 using the
technologies described in chapter 3 of the joint TSD in that emission reductions would be
achieved through specific hardware and controls. Finally, the accounting for credits for
control of direct and indirect A/C improvement credits is independent, which allows for a
separate discussion of these two categories.
5.1.2 Air Conditioner Leakage
5.1.2.1 Impacts of Refrigerant Leakage on Greenhouse Gas Emissions
There have been several studies in the literature which have attempted to quantify the
emissions (and impact) of air conditioner HFC emissions from light duty vehicles. In this
section, several of these studies are discussed. These inventories and impacts form the basis
for the air conditioner environmental credits, and in this final rule, we are using the same
emissions inventory and analysis method for refrigerant leakage as we did in the 2012-2016
rule as described in section 5.1.2.3.3.
0 Even if A/C systems utilize a "low-GWP" refrigerant, such as HFO-1234yf (GWP = 4), emissions are still a
concern. First, as refrigerant leaks from the system, once the refrigerant level drops to 40 to 50 percent of its
normal capacity, the operating efficiency of the system will degrade, resulting in an increase in fuel consumption
due to A/C use, and an increase in indirect emissions. Second, if systems do leak refrigerant at an excessive rate,
there is a higher probability that someone will unlawfully recharge the system with a cheaper, and higher-GWP
refrigerant, resulting in increased direct emissions.
5-3
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Based on measurements from 300 European vehicles (collected in 2002 and 2003),
Schwarz and Harnisch estimate that the average HFC direct leakage rate from modern A/C
systems was 53 g/yr.1 This corresponds to a leakage rate of 6.9% per year. This was
estimated by extracting the refrigerant from recruited vehicles and comparing the amount
extracted to the amount originally filled (as per the vehicle specifications). The fleet and size
of vehicles differs from Europe and the United States, therefore it is conceivable that vehicles
in the United States could have a different leakage rate. The authors measured the average
charge of refrigerant at initial fill to be about 747 grams (it is somewhat higher in the U.S. at
770g), and that the smaller cars (684 gram charge) emitted less than the higher charge
vehicles (883 gram charge). Moreover, due to the climate differences, the A/C usage patterns
also vary between the two continents, which may influence leakage rates.
Vincent et al., from the California Air Resources Board estimated the in-use
9
refrigerant leakage rate to be 80 g/yr. This value is based on the consumption of refrigerant
in commercial fleets, and surveys of vehicle owners and technicians. The study assumed an
average A/C charge size of 950 grams and a recharge rate of 1 in 16 years (lifetime). The
recharges occurred when the system was 52% empty and the fraction recovered at end-of-life
was 8.5%.
5.1.2.1.1 Emission Inventory
The EPA publishes an inventory of greenhouse gases and sinks on an annual basis.
The refrigerant emissions numbers that are used in the present analysis are from the Vintaging
model, which is used to generate the emissions included in this EPA inventory source. The
HFC refrigerant emissions from light duty vehicle A/C systems was estimated to be 61.8 Tg
CC>2 equivalent in 2005 by the Vintaging model.3'd In 2005, refrigerant leakage accounted for
about 5.1% of total greenhouse gas emissions from light duty sources. From a vehicle
standpoint, the Vintaging model assumes that 42% of the refrigerant emissions are due to
direct leakage (or "regular" emissions), 49% for service and maintenance (or "irregular"
emissions), and 9% occurs at disposal or end-of-life as shown in the following table. These
are based on assumptions of the average amount of chemical leaked by a vehicle every year,
how much is lost during service of a vehicle (from professional service center and do-it-
yourself practices), and the amount lost at disposal. These numbers vary somewhat over time
based on the characteristics (e.g. average charge size and leakage rate) of each "vintage" of
A/C system, assumptions of how new A/C systems enter the market, and the number of
vehicles disposed of in any given year.
Table 5-1 Light Duty Vehicle HFC-134a Emissions in 2005 from Vintaging Model - HFC Emissions
Multiplied by 1430 GWP to Convert to CO2 Equivalent
Emission Process
HFC emissions (metric
tons)
Fraction of total
d EPA reported the MVAC emissions at 56.6 Tg CO2 EQ, using a GWP of 1300. This number has been adjusted
using a GWP of 1430.
5-4
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Leakage
Maintenance/servicing
Disposal/end-of-life
Total
18,151
21,176
3,890
43,217
0.42
0.49
0.09
1.0
5.1.2.2 Alternative Refrigerants
Leakage emissions can also be reduced with the use of refrigerants other than R-134a,
which has a global warming potential (GWP) of 1430. To address future GHG regulations in
the Europe Union and the State of California, air conditioning systems which use alternative
refrigerants are under development, and have been demonstrated in prototypes by vehicle
manufacturers and A/C component suppliers. The European Union has enacted regulations
which require the use of refrigerants with a GWP less than 150. Phase-in of these EU
regulations began with new vehicle platforms in MY 2011, and will be completely phased-in
for all vehicles by MY 2017. Some of the alternative refrigerants under development by
manufacturers and A/C component suppliers include HFO-1234yf, CC>2, HFC-152a, and low-
GWP blends of existing refrigerants. The air conditioning component and refrigerant
manufacturers, as well as automotive manufacturers, are actively studying the performance,
efficiency, safety, and cost of these alternative refrigerants.
HFO-1234yf, with a GWP of 4, is a leading candidate as an alternative to R-134a
refrigerant. For example, General Motors has selected HFO1234yf for use in certain model
year 2013 vehicles.4 While the performance and efficiency of A/C systems using HFO-
1234yf can be equivalent to those using HFC-134a, the higher cost of implementing this
refrigerant - estimated at $67 (2010$, direct manufacturing cost) per vehicle in model year
2016 (see section 5.1.4) - is causing the industry to consider other solutions which are lower-
cost.
A so-called "natural refrigerant" under consideration is CC>2, which has a GWP of 1.
While this refrigerant is environmentally neutral from a GWP perspective (i.e. relative to a
CC>2 baseline), and is currently used in some commercial refrigeration units, its use in
automotive applications is challenging due to the higher operating pressure of CO2 systems,
where the peak pressure can be as high as 2000 PSI, compared to the peak pressure in HFC-
134a systems of around 450 PSI. Several European auto manufacturers have successfully
developed CC>2 A/C systems, but none have been produced for use in new vehicles at this
time. An A/C system which uses CC>2 is estimated to cost from about $140 to $210 more than
an equivalent HFC-134a system; however, the cost of the refrigerant itself is expected to be
considerably less than HFO-1234yf.5
F£FC-152a (1,1-difluoroethane) is a flammable refrigerant with a GWP of 120 and an
ASFIRAE flammability designation of Class 2. Given the flammability of this refrigerant, we
expect that manufacturers would either need to design their A/C systems with a secondary
loop or with directed relief valves to mitigate safety concerns within the cabin area, and to
comply with the use conditions at 40 CFR Part 82 Subpart G Appendix B. With a secondary
loop design, the evaporator is not located inside the passenger cabin area, but inside a chiller
in an underhood location, where a secondary fluid (such as an ethylene glycol-water mixture)
is circulated to transfer heat from the cabin to the chiller. This approach requires additional
system components (chiller, pump, reservoir, and plumbing for secondary loop), which adds
5-5
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
an estimated 12 Ibs. of mass to the vehicle.6 Secondary loop systems have added value in that
they have the ability to store cooling within the loop, which in turn allows for "free" cooling
to occur during deceleration events, and then delivered to the cabin during engine idle off
conditions (for example). With the directed relief valve design, the refrigerant within the A/C
system is vented and ducted to the atmosphere by opening high and low-side relief valves
when a leak is detected.7 The advantage of the directed relief valve approach (relative to a
secondary loop) is that fewer components are needed, potentially minimizing the mass and
cost of the system.
Other alternative refrigerants which may be used in the future may include low-GWP
blends. Recent studies have shown that the low-GWP refrigerant blends ACS and AC6 from
the chemical manufacturer Mexichem, have performance and efficiency characteristics which
are similar to HFC-134a under high-load (maximum cooling) conditions, and slightly reduced
performance and efficiency under low-load conditions. These mildly-flammable (similar to
HFO-1234yf) refrigerant blends, being comprised of several different refrigerant components,
have zeotropic properties. This means that the fraction of each component in the gas and
liquid phases is not constant, and varies with temperature and pressure within the system.8
Zeotropic behavior may result in mal-distribution of the refrigerant within the evaporator and
condenser, which negatively affects system efficiency, especially at low loads.9 However, it
is believed that optimization of evaporator and condenser design can improve the low-load
efficiency. These blends may be similar enough in performance and in their physical
characteristics to HFC-134a and HFO-1234yf that they may be used in current production
A/C systems designs with relatively minor modifications
We expect that stakeholders in the automotive A/C industry will continue to study and
develop low-GWP refrigerant solutions in order to minimize the direct and indirect impact of
A/C-related emissions. With the statutory requirements for low-GWP refrigerants in the
European Union, which began in model year 2011 for new vehicles designs, we expect that
one or more of these low-GWP solutions will be available for at least 20% of the U.S. vehicle
fleet by model year 2017, and that an additional 20% of the fleet can adopt the alternative
refrigerant in each subsequent model year. EPA expects that manufacturers would be
changing over to alternative refrigerants at the time of complete vehicle redesign, which
occurs about every 5 years, though in confidential meetings, some manufacturers/suppliers
have informed EPA that it may be possible to modify the hardware for some alternative
refrigerant systems between redesign periods.
5.1.2.3 A/C Leakage Credit
The level to which each technology can reduce leakage can be estimated using the
February 2012 version of SAE Surface Vehicle Standard J2727 - HFC-134a Mobile Air
Conditioning System Refrigerant Emission Chart. While this standard was developed for
leakage of F£FC-134a refrigerant, it is also applicable to the alternative refrigerant FIFO-
1234yf, and may be applicable to other low-GWP refrigerants as well. To convert J2727
chart emission (leak) rates from HFC-134a to HFO-1234yf leakage rates, the result is
multiplied by 1.03. This conversion factor for F£FO-1234yf is derived by multiplying the ratio
of the molecular weights of the two refrigerants (114 kg/kmol for F£FO-1234yf and 102
5-6
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
kg/kmol for HFC-134a) by the inverse ratio of the dynamic viscosities of the two refrigerants
(ll.lx 10-6Pa-sforHFC-134aand 12.0 x 10-6 Pa-s for HFO-1234yf).
The J2727 standard was developed by SAE and the cooperative industry and
government EVIAC (Improved Mobile Air Conditioning) program using industry experience,
laboratory testing of components and systems, and field data to establish a method for
calculating leakage. With refrigerant leakage rates as low as 10 g/yr, it would be exceedingly
difficult to measure such low levels in a test chamber (or shed). Since the J2727 method has
been correlated to "mini-shed", or SAE J2763, results (where A/C components are tested for
leakage in a small chamber, simulating real-world driving cycles), the EPA considers this
method to be an appropriate surrogate for vehicle testing of leakage.10 It is also referenced by
the California Air Resources Board in their Environmental Performance Label regulation and
the State of Minnesota in their GHG reporting regulation.11'12
5.1.2.3.1 Why Is EPA Continuing to Rely on a Design-Based Approach to Quantify
Leakage?
EPA is not reopening, reconsidering, or otherwise reevaluating its approach to
quantifying A/C leakage in the MYs 2012-2016 final rule. However, as in the MYs 2012-
2016 rule, EPA will continue to use a design-based method for quantifying refrigerant leakage
from A/C systems. In the time since the MYs 2012-2016 rule was finalized, the Agency was
not informed of any new approaches or methods for measuring actual refrigerant leak rates.
While EPA generally prefers performance testing for emissions, a feasible method for
measuring refrigerant emissions accurately from a vehicle is not available, and we are
finalizing for MYs 2017-2025 a continuation of the SAE 12727-based approach adopted in the
MYs 2012-2016 rule. EPA believes that the SAE J2727 method, as discussed below, is an
appropriate method for quantifying the expected yearly refrigerant leakage rate from A/C
systems.
5.1.2.3.2 How Will Leakage Credits Be Calculated?
For model years 2017 through 2025, the A/C credit available to manufacturers will be
calculated based on how much a particular vehicle's annual leakage value is reduced
compared to an average MY 2008 vehicle with baseline levels of A/C leakage technology,
and will be calculated using a method drawn directly from the updated SAE J2727 approach
(for details on these updates, see 5.1.2.3.2.2). By scoring the minimum leakage rate possible
on the J2727 components enumerated in the rule (expressed as a measure of annual leakage),
a manufacturer can generate the maximum A/C credit (on a gram per mile basis). To avoid
backsliding on leakage rates when using low-GWP refrigerants, where manufacturers could
choose less costly sealing technologies and/or materials, EPA is finalizing the proposed
disincentive credit for "high leak" on alternative refrigerant systems. The maximum value for
this high leak disincentive credit (or HiLeakDisincentive) is 1.8 g/mi for cars and 2.1 g/mi for
trucks, with lower amounts possible for leakage rates between the minimum leakage score
(MinScore) and the average impact (Avglmpact). The terms used for calculating the A/C
Leakage Credit as well as the HiLeakDisincentive are discussed later in this section.
5-7
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
The A/C credit available to manufacturers will be calculated based on the reduction to
a vehicle's yearly leakage rate, using the following equation for HFC-134a refrigerant:
Equation 5-1 Credit Equation for HFC-134a Refrigerant
A/CLeakage Credit = (MaxCredit) *[1 - (§86.166-12Score/AvglmpacF) *
(GWPRefrigerant /1430)]
and the following equation for low-GWP, alternative refrigerants:
Equation 5-2 Credit Equation for Alternative Refrigerants
A/CLeakage Credit = (MaxCredit) *[1 - (§86.166-12 Score / AvglmpacF) *
(GWPRefrigerant/1430)] - HiLeakDisincentive
where the HiLeakDisincentive is determined in accordance with one of the following three
conditions, depending on the refrigerant capacity (RefrigCapacity), or charge level, of the A/C
system:
For A/C systems with a refrigerant capacity <= 733g:
HiLeakDis = 0.0, t/ Score < 11.0 g/yr
/Score — 11\
HiLeakDis = 1.8 * ( — \,if 11.0 < Score < 14.3,
\ O i O /
HiLeakDis = 1.8, if Score > 14.3
For A/C systems with a refrigerant capacity > 733g:
HiLeakDis = 0.0, if Score < RefrigCapacity* 0.015
HiLeakDis = 1.8 * (Score - (RefrigCapacity * 0.015)/3.3),i/ RefrigCapacity * 0.015
< Score < RefrigCapacity * 0.015 + 3.3
HiLeakDis = 1.8, if Score > RefrigCapacity * 0.015 + 3.3
There are four significant terms to the credit equation. Each is briefly summarized below, and
is then explained more thoroughly in the following sections. Please note that the values of
' Section 86.166-12 sets out the individual component leakage values based on the SAE value.
5-S
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
many of these terms change depending on whether HFC-134a or an alternative refrigerant is
used. The values are shown in Table 5-2, and are documented in the following sections.
• "MaxCredit" is a term for the maximum amount of credit entered into the equation
before constraints are applied to terms. The maximum credits that could be generated
by a manufacturer is limited by the choice of refrigerant and by assumptions regarding
maximum achievable leakage reductions. Some of these values may have changed
since the 2012-2016 rule.
• "Score" is the leakage score of the A/C system as measured according to the §86.166-
12 calculation in units of g/yr, where the minimum score which is deemed feasible is
fixed.
• "Avglmpact" is a term which represents the annual average impact of A/C leakage.
• "MinScore" is the lowest leak score that EPA projects is possible, when starting from
a baseline, or Avglmpact, system. The MinScore represents a 50% reduction in
leakage from the baseline levels based on the feasibility analysis detailed below.
• "GWPRefrigerant" is the global warming potential for direct radiative forcing of the
refrigerant as defined by EPA (or IPCC).
• "HiLeakDisincentive" is a term for the disincentive credit deducted for low-GWP
alternative refrigerant systems which have a leakage rate greater than the minimum
leakage score of 11.0 g/year for cars and trucks. The maximum Disincentive is 1.8
g/mile for cars and 2.1 g/mile for trucks. The 11.0 g/year threshold for generating a
HiLeakDisincentive is based on the analysis we used for the MY 2014-2018 GHG
Emissions Standards for Heavy-Duty Engines and Vehicles, where a maximum
refrigerant leak rate standard of 11.0 g/year was set for vehicles with a refrigerant
capacity of 733 g or lower, and 1.5 percent of the refrigerant capacity (in grams) for
systems with a refrigerant capacity greater than 733 g. .
Table 5-2 Components of the A/C Credit Calculation
MaxCredit equation input (grams/mile CO 2 EQ)
A/C credit maximum (grams/mile CC>2 EQ)"
§86.166-12 MinScore (grams HFC/year)
Avg Impact (grams HFC/year)
HFC-134a
Cars
12.6
6.3
8.3
16.6
Trucks
15.6
7.8
10.4
20.7
Lowest-GWP
Refrigerant
(GWP=1)
Cars
13.8
13.8
8.3
16.6
Trucks
17.2
17.2
10.4
20.7
a With electric compressor, value increases to 9.5 and 11.7 for cars and trucks, respectively.
5.1.2.3.2.1 Max Credit Term
In order to determine the maximum possible credit on a gram per mile basis, it was
necessary to determine the projected real world HFC emissions per mile. This calculation is
done exactly the same as it was done for the MYs 2012-2016 final rule. Because HFC is a
5-9
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
leakage type of emission, it is largely disconnected from vehicle miles traveled (VMT).
Consequently, EPA calculated the total HFC inventory (in 2016), and then calculated the
VMT for that year separately. The quotient of these two terms is the HFC contribution per
mile.
Consistent with the methodology presented in the MYs 2012-2016 rule, the HFC
emission inventories were estimated from a number of existing data sources. The per-vehicle
per-year HFC emission of the current vehicle fleet was determined using averaged 2005 and
2006 registration data from the Transportation Energy Databook (TEDB) and 2005 and 2006
mobile HFC leakage estimates from the EPA Emissions and Sinks report described above.3'13
The per-vehicle per-year emission rates were then adjusted to account for the new definitions
of car and truck classes by increasing the car contribution proportionally by the percentage of
former trucks that are reclassified as cars.g This inventory calculation assumes that the
leakage rates and charge sizes of future fleets (absent any standards) are equivalent to the fleet
present in the 2005/2006 reference years. Preliminary EPA analysis indicates that this may
increasingly overstate the future HFC inventory, as charge sizes are decreasing, though more
is discussed on this topic below.
The per-vehicle per-year average emission rate was then scaled by the projected
vehicle fleet in each future year (using the fleet predicted in the emissions analysis) to
estimate the HFC emission inventory if no further controls were enacted on the fleet. After
dividing the 2016 inventory by total predicted VMT in 2016, an average per mile HFC
emission rate ("base rate") was obtained.
The base rate is an average in-use number, which includes both old vehicles with
significant leakage, as well as newer vehicles with very little leakage. The new vehicle
leakage rate is discussed in section 5.1.2.4, while deterioration is discussed in section 5.1.2.6.
• Max Credit with Conventional Refrigerant (HFC-134a)
Two adjustments were made to the base rate in order to calculate the Maximum HFC
credit with conventional refrigerant. First, EPA has determined that 50% leakage
prevention is the maximum potentially feasible prevention rate in the timeframe of this
rule (see section 5.1.2.4). Some leaks will occur and are expected, regardless of
prevention efforts. The accuracy of the J2727 approach (as expressed in §86.112), as
a design based test, decreases as the amount of expected leakage diminishes. 50% of
the base rate is therefore set as the maximum potential leakage credit for
improvements to HFC leakage using conventional refrigerant.
f In short, leakage emissions occur even while the car is parked, so the connection to a gram/mile credit is not
straightforward. However, HFC emissions must be converted to a gram/mile basis in order to create a relevant
credit.
8 Many of these "older" references still use the old definition of car and truck. The new definitions do not apply
until model year 2011.
5-10
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Second, EPA expects that improvements to conventional refrigerant systems will
affect both leakage and service emissions, but will not affect end of life emissions.
EPA expects that reductions in the leakage rate from A/C systems will result in fewer
visits for maintenance and recharges. This will have the side benefit of reducing the
emissions leftover from can heels (leftover in the recharge cans) and the other releases
that occur during maintenance. However, as disposal/end of life emissions will be
unaffected by the leakage improvements (and also are subject to control under the
rules implementing Title VI of the CAA), the base rate was decreased by a further 9%
(Table 5-1).
• Max Credit with Alternative Refrigerant
Emission reductions greater than 50% are possible with alternative refrigerants. As an
example, if a refrigerant with a GWP of 0 were used, it would be possible to eliminate
all refrigerant GHG emissions. In addition, for alternative refrigerants, the EPA
believes that vehicles with reduced GWP refrigerants should get credit for end of life
emission reductions. Thus, the maximum credit with alternative refrigerant is about
9% higher than twice the maximum leakage reduction.
As discussed above, EPA recognizes that substituting a refrigerant with a significantly
lower GWP will be a very effective way to reduce the impact of all forms of refrigerant
emissions, including maintenance, accidents, and vehicle scrappage.
The A/C Leakage Credits that will be available will be a function of the GWP of the
alternative refrigerant as well as of leakage, with the largest credits being available for
refrigerants with GWPs at or approaching a value of 1, while also maintaining a low leakage
rate. For a hypothetical alternative refrigerant with a GWP of 1 (e.g., CO2 as a refrigerant),
effectively eliminating leakage as a GHG concern, our credit calculation method could result
in maximum credits equal to total average emissions, or credits of 13.8 and 17.2 g/mi CO2eq
for cars and trucks, respectively, as incorporated into the A/C Leakage Credit formula above
as the "MaxCredit" term.
As we did for the MYs 2012-2017 rule, EPA made a final adjustment to each credit to
account for the difference between real-world HFC emissions and test-cycle CO2 emissions.
It has been shown that the tests currently used for CAFE certification represents an
approximately 20% gap from real world fuel consumption and the resulting CO2 emissions.14
Because the credits from direct A/C improvements are taken from a real world source, and are
being traded for an increase in fuel consumption due to increased CO2 emissions, the credit
was multiplied by 0.8 to maintain environmental neutrality (Table 5-3).
Table 5-3 HFC Credit Calculation for Cars and Trucks Based on a GWP of 1430
HFC
Inventory
(MMT
CO2 EQ)
VMT
(Billions
of Miles)
Total HFC
EmissionsPer
Mile
(CO2EQ
Gram/mile)
HFC
Leakage and
Service
EmissionsPer
Mile
(C02EQ
Maximum
Credit w/
alternative
refrigerant
(Adjusted
for On-
Maximum
Credit w/o
alternative
refrigerant
(50% of
Adjusted
5-11
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Car
Truck
Total
27.4
30.4
57.8
1,580
1,392
2,972
17.2
21.5
18.6
Gram/mile)
15.5
19.6
16.9
road gap &
including
end of life)
13.8
17.2
14.9
HFC&
excluding
end of life)
6.3
7.8
6.8
5.1.2.3.2.2 Section 86.166-12, implementing the J2727 Score Term
The J2727 score is the SAE J2727 yearly leakage estimate of the A/C system as
calculated according to the J2727 procedure. In the time since the MYs 2012-2016 Light-
Duty GHG Rule, there have been several refinements to the J2727 procedure which EPA has
incorporated into the EPA regulations. First, a provision was made for system joints where
100 percent of the joints are leak test with helium and a mass-spectrometer leak detector. If
the joints pass this leak test, they can be considered to have a leakage factor equivalent to that
of a seal washer, which is next to the lowest factor possible for system joints. Second, a
requirement was added to use SAE J2064 hose permeation test results in place of the discrete
values for various hose material and construction types that were provided in previous
versions of the J2727 test method. By using the test chamber results for refrigerant
permeation through hoses, a more representative leakage estimate for the overall system is
achieved. The minimum J2727 score for cars and trucks is a fixed value, and the section
below describes the derivation of the minimum leakage scores that can be achieved using the
J2727 procedure.
To calculate a J2727 score and credit for the alternative refrigerant HFO-1234yf, all
values relevant to the credit calculations, as well as the J2727 score, shall be adjusted to
account for the higher molecular weight of this refrigerant. In contrast to the studies
discussed in section 5.1.2.6 which examines the HFC emission rate of the in-use fleet (which
includes vehicles at all stages of life), the SAE J2727 estimates leakage from new vehicles. In
the development of J2727, two relevant studies were assessed to quantify new vehicle
emission rates. In the first study, measurements from relatively new (properly functioning
and manufactured) Japanese-market vehicles were collected. This study was based on 78 in-
use vehicles (56 single evap, 22 dual evap) from 7 Japanese auto makers driven in Tokyo and
Nagoya from April, 2004 to December, 2005. The study also measured a higher emissions
level of 16 g/yr for 26 vehicles in a hotter climate (Okinawa). This study indicated the
leakage rate to be close to 8.6 g/yr for single evaporator systems and 13.3 g/yr for dual
evaporator systems.15 A weighted (test) average gives 9.9 g/yr. In the second study,
emissions were measured on European-market vehicles up to seven years age driven from
November, 2002 to January, 2003.16 The European vehicle emission rates were slightly
higher than the Japanese fleet, but overall, they were consistent. The average emission rate
from this analysis is 17.0 g/yr with a standard deviation of 4.4 g/yr. European vehicles,
because they have smaller charge sizes, likely understate the leakage rate relative to the
United States. To these emission rates, the J2727 authors added a factor to account for
occasional defective parts and/or improper assembly and to calibrate the result of the SAE
J2727 calculation with the leakage measured in the vehicle and component leakage studies.
5-12
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
We adjust this rate up slightly by a factor proportional to the average European
refrigerant charge to the average United States charge (i.e. 770/747 from the Vintaging model
and Schwarz studies respectively). The newer vehicle emission rate is thus 18 g/yr for the
average newer vehicle emissions (average for car and truck).
To derive the minimum score, the 18 gram per year rate was used as a ratio to convert
the gram per mile emission impact into a new vehicle gram per year for the test. The car or
truck direct a/c emission factor (gram per mile) was divided by the average emission factor
(gram per mile) and then multiplied by the new vehicle average leakage rate (gram per year)
Equation 5-3 - J2727 Minimum Score
J2727 Minimum Score = Car or truck average pre control emissions (gram per
mile)/ Fleet average pre-control emissions (grams per mile) x New vehicle annual
leakage rate (grams per year) x Minimum Fraction
By applying this equation, the minimum J2727 score is fixed at 8.3 g/yr for cars and
10.4 g/yr for trucks. This corresponds to a total fleet average of 18 grams per year, with a
maximum reduction fraction of 50%.
The GWP Refrigerant term in Equation 1 allows for the accounting of refrigerants
with lower GWP (so that this term can be as low as zero in the equation), which is why the
same minimum score is kept regardless of refrigerant used. It is technically feasible for the
J2727 Minimum score to be less than the values presented in the table. But this will usually
require the use of an electric compressor (see below for technology description).
5.1.2.3.2.3 Avglmpact Term
Avglmpact is the average annual impact of A/C leakage, which is 16.6 and 20.7 g/yr
for cars and trucks respectively. This was derived using Equation 2, but by setting the
minimum fraction to one.
5.1.2.3.2.4 GWPRefrigerant Term
This term is relates to the global warming potential (GWP) of the refrigerant as
documented by EPA. A full discussion of GWP and its derivation is too lengthy for this
space, but can be found in many EPA documents.40 This term is used to correct for
refrigerants with global warming potentials that differ from HFC-134a.
5.1.2.3.2.5 HiLeakDisincentive
As proposed, EPA is adding (compared to the MYs 2012-2016 rule's formula) a
disincentive to the leakage credit formula for systems which use a low GWP refrigerant, but
"backslide" on low leakage levels. As stated above, low leakage levels provide an
environmental benefit by maintaining the charge of the system. This has two advantageous
effects. First, it preserves the efficiency of the system. Reduced refrigerant charge levels can
reduce overall efficiency, especially if the compressor starts "short-cycling". Also, since
lubrication is combined with most of the current and likely future refrigerants, the shortage of
5-13
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
lubrication can wear out the compressor and cause it to seize and malfunction. CARB testing
has shown that preserving the refrigerant charge level in a conventional A/C system results in
improved system efficiency.17 Second, by reducing the leak rate of the low GWP system, the
probability that the new system will run out of charge will be minimized. When a system
runs out of charge, vehicle owners can either drive without A/C, or have a professional
recharge the system, or recharge the system themselves. The latter are called "do-it-yourself-
ers" (DIYers). It is possible that DIYers (and some repair shops) may refill a low-GWP
system (e.g., HFO-1234yf) with a high-GWP refrigerant (e.g., HFC-134a), in order to save on
costs. Due to the demand from the legacy fleet, refill containers of HFC-134a would be
available to the market for many years to come (so it would be available to DIYers). Since
the thermodynamic properties of HFC-134a and HFO-1234yf are similar, HFO-1234yf
systems may function with HFC-134a, although with some reduced effectiveness, and in
some systems may lead to long term damage.1 Unfortunately, the extent to which this will
occur is difficult to predict. EPA regulations prohibit topping-off a system with a refrigerant
other than the original (for which the system was designed). EPA will use this disincentive
credit to maintain low refrigerant leakage emission levels and to reduce the potential for
leakage of high GWP refrigerants from systems that have been improperly recharged. Thus,
EPA believes that there are real, but unquantifiable, benefits for a leakage disincentive credit,
and we are finalizing a (Max)HiLeakDisincentive of 1.8 g/mi for cars and 2.1 g/mi for trucks.
The EPA believes that these numbers strike a balance in that it is a large enough incentive to
maintain low leakage levels, but it is not too large as to diminish incentive to switch to an
alternative refrigerant.
The leakage rate at which a disincentive, or negative credit, would be generated
(MinScore) was increased from the levels proposed in the NPRM. Most commenters
requested that the EPA remove this negative credit for the final rule. Other commenters
believed that the 8.3 g/yr and 10.4 g/yr thresholds for cars and trucks that we proposed were
overly stringent, and that a more realistic value should be specified, based what the current
fleet of vehicles is achieving. In response to these comments, we examined the approach we
used in the MYs 2014-2018 rule for GHG Emissions Standards for Heavy-Duty Engines and
Vehicles, where a maximum refrigerant leak rate standard of 11.0 g/year was set for vehicles
with a refrigerant capacity of 733 g or less, and 1.5 percent of the system refrigerant capacity
for vehicles with a refrigerant capacity greater than 733 g. We believe that the approach used
in the heavy-duty GHG rule is appropriate for setting the threshold for which a
HiLeakDisincentive (a "negative" credit) is generated in the MYs 2017-2025 light-duty GHG
rule, as the air conditioning systems in both categories are similar, in terms of the components
and materials used, as well as general system design and layout. Furthermore, analysis of the
2012 model year 12727-based leakage rate data in the State of Minnesota reporting database
affirms that this approach will require leakage reductions in a large portion of the vehicle
h Refilling/ topping-off systems designed for use for one refrigerant (e.g., HFO-1234yf) with another refrigerant
(e.g., HFC-134a) is a violation of CAA Section 612 regulations. See 40 CFR part 82, subpart G, appendix D,
sections.
1 Based on discussions with vehicle manufacturers.
5-14
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
fleet, while allowing early adopters of low-leakage technologies to avoid the disincentive. In
addition, this scaled approach, where larger A/C systems with higher refrigerant capacities
can have a higher leakage rate before triggering the HiLeakDisincentive, provides
manufacturers with the flexibility they need to install appropriately-sized A/C systems in
larger vehicles, while using common low-leak technologies and components across their
vehicle models. If a single threshold were applied to cars and trucks, extraordinary leakage
mitigation measures would likely be necessary on larger systems in order to avoid the
disincentive. Figure 5-1 illustrates how this approach to the HiLeakDisincentive fits within
the 2012 model leakage scores for light-duty vehicles.
O 2012 MN Data
-2017-2025
HiLeakDisincentiv
e Min. Threshold
' 2017-2025
HiLeakDisincentiv
e Max. Threshold
800 1000 1200 1400
Refrigerant Charge Size (g)
Figure 5-1 2012 State of Minnesota Leakage Reporting Data (All Vehicles) with HiLeakDis Thresholds
5.1.2.3.3 Why are the leakage credits different from the 2010 Technical Assessment
Report?
The 2010 Technical Assessment report employed a different methodology for
calculating the HFC credit, which resulted in significantly fewer credits available for A/C
leakage compared to the MYs 2012-2016 final rule (approximately 40% less). The TAR
analysis decreased the average charge size and leakage assumed in its analysis of future model
years as compared to the MYs 2012-2016 final rule. In the present rule, we maintain the MYs
2012-2016 credit value. EPA chose this approach for both technical and policy reasons.
5-15
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Like any inventory, the refrigerant inventory produced by the Vintaging model has
uncertainties associated with it. This is especially true given that we do not know how many
"high emitters" exist in the U.S. fleet. A high emitter is a vehicle that rapidly leaks HFC, but is
also continually recharged. A typical light duty vehicle may require recharge approximately
every seven years (see section 5.1.2.6). However, the owner of a high emitter may continually
charge their systems each summer, thereby increasing the overall average emissions of the fleet.
In the 2009-2010 study of the leakage rates from 70 in-use heavy duty vehicles, the California Air
Resources Board found a relatively high prevalence of high emitting vehicles. Of the 70, 5 had
leakage rates that were greater than one-half a charge per year, while seven additional vehicles
had annualized leakage rates greater than one-quarter charge per year.j These values could
potentially be used to recalculate the HFC inventories from the TAR and recalculate the leakage
credit.
EPA considered the lower inventory discussed in the TAR as well as the CARB study
when determining the leakage credit for this rule. While there is ultimately a mass balance
between HFC produced and HFC leaked, this balance is not closed on an annual basis, and is
difficult to directly verify. Given the counterbalancing factors, EPA made the policy decision to
maintain continuity with the MYs 2012-2016 FRM analysis, and will incorporate this level of the
credit in the standard setting process. A reduction in A/C credits (in 2017 compared to 2016 for
example) would artificially increase the stringency of the standard for those manufacturers who
generated leakage (and alternative refrigerant) credits in 2016 as a means of compliance. With
little lead time, these manufacturers would need to add other technologies to their fleet in order to
close the gap their compliance target created by a reduction in the maximum potential A/C credit.
Alternately, the stringency of the 2017 standards would have to be relaxed, and in some cases may
even be less stringent than 2016 standards if this adjustment is made.
ICCT expressed the concern that maintaining continuity with the earlier rule would
encourage manufacturers to seek to generate leakage credits more aggressively than they
otherwise would. As we stated above, we acknowledge that some manufacturers might choose a
slightly different technology approach to compliance if fewer A/C leakage credits were available.
However, we believe that the disruption to the transition from the MYs 2012-2016 rule to the
MYs 2017-2025 rule that would result is not acceptable. Given the need for stability for the
standards (and stringencies), EPA is "freezing" the credit assessment based on what we presented
in the MYs 2012-2016 rule, and also presented again above.
5.1.2.4 Technologies That Reduce Refrigerant Leakage and their Effectiveness
In this section, the analysis used in the MYs 2012-2016 rule is again applied to the
baseline technology levels and the effectiveness for leakage-reducing technologies. For the
MYs 2012-2016 rule, EPA conducted an analysis to determine the historic leakage emission
rate for motor vehicle A/C systems, and it was estimated in section 5.1.2.3.2.2 that the A/C
systems in new vehicles would leak refrigerant at an average rate of 18 g/yr - a value which
EPA believes represented the types of A/C components and technologies in use prior to MY
J While the Vintaging model assumes an average annualized leakage rate of 18% + 43% at end of life, it assumes
that the MAC unit only lasts 12 years. Actual MACs, particularly those that are recharged, may last longer.
5-16
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
2007. EPA believes, through utilization of the leakage-reducing technologies described
below, that it will be possible for manufacturers to reduce refrigerant leakage 50%, relative to
the 18 g/yr baseline level.18 EPA also believes that all of these leakage-reducing technologies
are currently available, and that manufacturers will use these technologies to generate credits
under provisions of the 2012-2016 rule, as well as under the provisions of this rule.
In describing the technologies below, only the relative effectiveness figures are
presented, as the individual piece costs are not known. The EPA only has costs of complete
systems based on the literature, and the individual technologies are described below.
5.1.2.4.1 Baseline Technologies
The baseline technologies assumed for A/C systems which have an average annual
leak rate of 18 g/yr are common to many mass-produced vehicles in the United States. In
these mass-produced vehicles, the need to maintain A/C system integrity (and the need to
avoid the customer inconvenience of having their A/C system serviced due to loss of
refrigerant) is often balanced against the cost of the individual A/C components. For
manufacturers seeking improved system reliability, components and technologies which
reduce leakage (and possibly increased cost) are selected, whereas other manufacturers may
choose to emphasize lower system cost over reliability, and choose components or
technologies prone to increased leakage. In EPA's baseline scenario, the following
assumptions were made concerning the definition of a baseline A/C system:
- all flexible hose material is rubber, without leakage-reducing barriers or veneers, of
approximately 650 mm in length for both the high and low pressure lines
- all system fittings and connections are sealed with a single o-rings
- the compressor shaft seal is a single-lip design
- one access port each on the high and low pressure lines
- two of the following components: pressure switch, pressure relief valves, or pressure
transducer
- one thermostatic expansion valve (TXV)
The design assumptions of EPA baseline scenario are also similar to the sample
worksheet included in SAE's surface vehicle standard J2727 - (R) HFC-134a Mobile Air
Conditioning System Refrigerant Emission Chart.10 In the J2727 emission chart, it is the
baseline technologies which are assigned the highest leakage rates, and the inclusion of
improved components and technologies in an A/C system will reduce this annual leakage rate,
as a function of their effectiveness relative to the baseline. EPA considers these 'baseline'
technologies to be representative of recent model year vehicles, which, on average, can
experience a refrigerant loss of 18 g/yr. However, depending on the design of a particular
vehicle's A/C system (e.g. materials, length of flexible hoses, number of fittings and adaptor
plates, etc.), it is possible to achieve a leakage score much higher (i.e. worse) than 18 g/yr.
According to manufacturer data submitted to the State of Minnesota, 19% of 2009 model year
vehicles have a J2727 refrigerant score greater than 18 g/yr, with the highest-scoring vehicle
reporting a leakage rate of 30.1 g/yr.19 For the 2010 model year, the average J2727 leakage
score reporting database was 14.0 g/yr for cars, and 14.8 g/yr for trucks, but this is simply the
average result of all vehicles in the database, and does not reflect sales weighting of the
5-17
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
leakage scores nor does it eliminate identical models (vehicles with different brands or
nameplates, but identical with respect to the A/C system design and components) when
calculating the average score.
Here again, the 18g/yr baseline is maintained at the MYs 2012-2016 rule levels for
both technical and policy reasons. As mentioned earlier, there is great uncertainty in the
leakage emissions from vehicles. The J2727 scoring system, which is calibrated to in-use
emissions from properly functioning vehicles, does not include high emitters. EPA considers
J2727 to be a surrogate for in-use emissions, and not necessarily an accurate representation of
real-world emissions. Thus, to maintain continuity with MY 2016 standards (and credits),
EPA is "freezing" the baseline assumption of leakage rate from the fleet.
5.1.2.4.2 Flexible Hoses
The flexible hoses on an automotive A/C system are needed to isolate the system from
engine vibration and to allow for the engine to roll within its mounts as the vehicle accelerates
and decelerates. Since the compressor is typically mounted to the engine, the lines going to-
and-from the compressor (i.e. the suction and pressure lines) must be flexible, or unwanted
vibration would be transferred to the body of the vehicle (or other components), and excessive
strain on the lines would result. It has been industry practice for many years to manufacture
these hoses from rubber, which is relatively inexpensive and durable. However, rubber hoses
are not impermeable, and refrigerant gases will eventually migrate into the atmosphere. To
reduce permeation, two alternative hose material can be specified. The first material, is
known as a standard 'veneer' (or 'barrier') hose, where a polyamide (polymer) layer - which
has lower permeability than rubber - is encased by a rubber hose. The barrier hose is similar
to a veneer hose, except that an additional layer of rubber is added inside the polyamide layer,
creating three-layer hose (rubber-polyamide-rubber). The second material is known as 'ultra-
low permeation', and can be used in a veneer or barrier hose design. This ultra-low
permeation hose is the most effective at reducing permeation, followed by the standard veneer
or barrier hose. Permeation is most prevalent during high pressure conditions, thus it is even
more important that these low permeable hoses are employed on the high pressure side, more
so than on the low pressure side.
According to J2727, standard barrier veneer hoses have 25% the permeation rate of
rubber hose, and ultra low permeable barrier veneer hoses have 10% the permeation rate (as
compared to a standard baseline rubber hose of the same length and diameter). However, in
the February 2012 version of J2727, manufacturers are required to use actual SAE J2064 hose
permeation data, instead of the discrete values provided for various hose material and
construction types, as was specified in previous versions of the J2727 method.
5.1.2.4.3 System Fittings and Connections
Within an automotive A/C system and the various components it contains (e.g.
expansion valves, hoses, rigid lines, compressors, accumulators, heat exchangers, etc.), it is
necessary that there be an interface, or connection, between these components. These
interfaces may exist for design, manufacturing, assembly, or serviceability reasons, but all
A/C systems have them to some degree, and each interface is a potential path for refrigerant
5-18
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
leakage to the atmosphere. In SAE J2727 emission chart, these interfaces are described as
fittings and connections, and each type of fitting or connection type is assigned an emission
value based on its leakage potential; with a single O-ring (the baseline technology) having the
highest leak potential; and a metal gasket having the lowest. In between these two extremes,
a variety of sealing technologies, such as multiple o-rings, seal washers, and seal washers with
o-rings, are available to manufacturers for the purpose of reducing leakage. It is expected that
manufacturers will choose from among these sealing technology options to create an A/C
system which offers the best cost-vs-leakage rate trade-off for their products.
The relative effectiveness of the fitting and connector technology is presented in Table
5-4. For example, the relative leakage factor of 125 for the baseline single O-ring is 125
times more "leaky" than the best technology - the metal gasket.
Table 5-4 Effectiveness of Fitting and Connector Technology
Fitting or Connector
Single O-ring
Single Captured O-ring
Multiple O-ring
Seal Washer
Seal Washer with O-ring
Metal Gasket
100% Helium Leakiest
Relative
Leakage
125
75
50
10
5
1
10
5.1.2.4.4 Compressor Shaft Seal
A major source of refrigerant leakage in automotive A/C systems is the compressor
shaft seal. This seal is needed to prevent pressurized refrigerant gasses from escaping the
compressor housing. As the load on the A/C system increases, the pressure and the leakage
past the seal also increase. In addition, with a belt-driven A/C compressor, a side load is
placed on the compressor shaft by the belt, which can cause the shaft to deflect slightly. The
compressor shaft seal must have adequate flexibility to compensate for this deflection, or
movement, of the compressor shaft to ensure that the high-pressure refrigerant does not leak
past the seal lip and into the atmosphere. When a compressor is static (not running), not only
are the system pressures lower, the only side load on the compressor shaft is that from tension
on the belt, and leakage past the compressor shaft is at a minimum. However, when the
compressor is running, the system pressure is higher and the side load on the compressor shaft
is higher (i.e. the side load is proportional to the power required to turn the compressor shaft)
- both of which can increase refrigerant leakage past the compressor shaft seal. It is estimated
that the rate of refrigerant leakage when a compressor is running can be 20 times that of a
static condition.20 Due to the higher leakage rate under running conditions, SAE J2727
assigns a higher level of impact to the compressor shaft seal. In the example shown in the
August 2008 version of the J2727 document, the compressor is responsible for 58% of the
system refrigerant leakage, and of that 58%, over half of that leakage is due to the shaft seal
alone (the remainder comes from compressor housing and adaptor plate seals). To address
refrigerant leakage past the compressor shaft, manufacturers can use multiple-lip seals in
place of the single-lip seals.
5-19
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
5.1.2.5 Technical Feasibility of Leakage-Reducing Technologies
EPA believes that the leakage-reducing technologies discussed in the previous
sections are available to manufacturers today, are relatively low in cost, and that their
feasibility and effectiveness have been demonstrated by the SAE EVIAC teams. EPA also
believes - as has been demonstrated in the J2727 calculations submitted by manufacturers to
the State of Minnesota - that reductions in leakage from 18 g/yr to 9 g/yr are possible (e.g. the
2009 Saturn Vue has a reported leakage score of 8.5 g/yr). In addition to generating credit for
reduced refrigerant leakage, we expect many manufacturers to choose to introduce alternative
refrigerant systems, such as HFO-1234yf, as discussed in Section III.C.l of the preamble to
this rule.
5.1.2.6 Deterioration in Leakage Controls
In the MYs 2012-2016 rule as well as in the proposal for this rule, EPA presented a
"model" of the deterioration of leakage systems. This analysis would have been necessary if
EPA wanted to quantify the maintenance benefits of leakage control. EPA received no
comments on this model for the proposal. For this final rule, EPA is not claiming
maintenance cost savings due to refrigerant leakage regulations. This is due to uncertainty as
to how these low-leak technologies will actually perform and deteriorate in use. While we
expect that low-leak systems will have a lower probability of a requiring a recharge
maintenance event, or at least a longer period before such an event is required, it is difficult to
quantify such factors, and such quantification is necessary for their inclusion in a cost-benefit
analysis. Moreover, EPA is estimating that the predominant technology in the MYs 2017-
2025 timeframe will be alternative refrigerants , thus minimizing the need for this
deterioration model.
Despite the fact that we are not using the deterioration model (as proposed), EPA
believes that it is important to address the issue as it relates to the hi-leak disincentive. Since
the deterioration model was presented, EPA has reconsidered some of the assumptions that
went into the model. Given that the deterioration mechanisms are not fully understood and
quantified, it is difficult to project the precise rate at which leak-reducing A/C technologies
will deteriorate compared to conventional technologies. But we do know that even if a
similar rate of deterioration is assumed, the A/C system with a lower initial leakage rate will
have a lower frequency of required recharge events over its lifetime, as it will take the low-
leak vehicle longer to reach a level of 50 percent charge remaining in the system. In addition,
we believe that many of the leak-reducing technologies, such as use of seal washers in place
of O-rings, and 100 percent leak testing of assembled components with helium, are inherently
beneficial to system durability. In the case of seal washers, a rigid, metal connection is
created at system joints instead of a less-rigid o-ring seal, which can be susceptible to damage
upon installation. In the case of helium leak testing, any defective joints, connections, or
components are detected prior to installation in the vehicle, which reduces the probability of
k Though we are encouraging manufacturers to keep the leakage scores low with the hi leak disincentive.
5-20
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
vehicles with higher-than-expected leakage leaving the assembly plant. EPA believes that
establishing an incentive to achieve low-leak systems through a HiLeakDisincentive (see
5.1.2.3.2.5) will result in lower deterioration rates and extend the interval for which a system
recharge is required. For these reasons, EPA believes that the deterioration model presented
in the proposal was potentially overly "conservative". EPA will continue to monitor data in
the future on the issue of leakage deterioration and the effects of the hi-leak disincentive.
5.1.2.7 Other Benefits of Improving A/C Leakage Performance
The EPA is assuming that a reduction in leakage emissions from new vehicles will
also improve the leakage over the lifetime of the vehicle. There is ample evidence to show
that A/C systems that leak more also have other problems that occur (especially with the
compressor) due to the lack of oil circulating in the system. Thus, it is expected that an A/C
system which utilizes leak-reducing components and technologies should, on average, last
longer than one which does not.
A European study conducted in 2001 (by Schwarz) found that the condenser is the
component most likely to fail and result in a total leak.21 The study also found that
compressor component was most likely the culprit when other malfunctions were present
(other than total loss). A more recent (and larger) study found that condensers required
replacement at half the rate of a compressor (10% vs 19% of the entire part replacement rate),
and that evaporators and accumulators failed more often.16 The same study also found that
many of the repairs occurred when the vehicles were aged 5-10 years. Both these studies
indicate that the condenser and compressor are among the major causes of failure in an A/C
system. Leakage reductions in the system are expected to greatly reduce the incidence of
compressor repair, since one of the main root causes of compressor failure is a shortage of
lubricating oil, which originates from a shortage of refrigerant flowing through the system
(and it is a refrigerant-oil mixture which carries lubricating oil to the compressor).22
Monitoring of refrigerant volume throughout the life of the A/C system may provide
an opportunity to circumvent some previously described failures specifically related to
refrigerant loss. Similar to approaches used today by the engine on-board diagnostic systems
(OBD) to monitor engine emissions, a monitoring system that informed the vehicle operator
of a low refrigerant level could potentially result in significant reductions in A/C refrigerant
emissions due to component failure(s) by creating an opportunity for early repair actions.
While most A/C systems contain sensors capable of detecting the low refrigerant pressures
which result from significant refrigerant loss, these systems are generally not designed to
inform the vehicle operator of the refrigerant loss, and that further operation of the unrepaired
system can result in additional component damage (e.g. compressor failure). Electronic
monitoring of the refrigerant may be achieved by using a combination of existing A/C system
sensors and new software designed to detect refrigerant loss before it progresses to a level
where component failure is likely to occur.
EPA requested comment in the 2012-2016 NPRM on allowing additional leakage
credits for systems that monitor the leak levels, especially where manufacturers are willing to
warrant such systems. Presently, the EPA is not aware if any such technology exists to
accurately monitor refrigerant levels, as the technical challenges are high. As a result, there
5-21
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
were no manufacturers who expressed interest in this credit, and the EPA did not finalize such
credits in the 2012-2016 program. EPA again sought comment during the MY 2017-2025
proposal on allowing these credits again, in the hopes of encouraging innovative technologies
to monitor leakage levels. ICCT supported EPA request for monitoring technology in their
comments, however, no manufacturer (or supplier) provided comment on this issue and EPA
remains unaware of any such technology in existence (much less in implementation).
5.1.3 COi Emissions and Fuel Consumption due to Air Conditioners
As stated above, for model years 2012 to 2016, EPA provided credits for the use of
A/C technologies that improve efficiency and achieve reductions in indirect CC>2 emissions
related to A/C use. These credits were not previously applicable to the CAFE program fuel
economy calculations. For this rule, the agencies are finalizing provisions that the A/C
indirect credits are applicable to both the greenhouse gas and fuel economy calculations.
5.1.3.1 Impact of Air Conditioning Use on Fuel Consumption and COi Emissions
Three studies have been performed in recent years which estimate the impact of A/C
use on the fuel consumption of motor vehicles in the United States. In the first study, the
National Renewable Energy Laboratory (NREL) and the Office of Atmospheric Programs
(OAP) within EPA have performed a series of A/C related fuel use studies.23'24 The energy
needed to operate the A/C compressor under a range of load and ambient conditions was
based on testing performed by Delphi, an A/C system supplier. They used a vehicle
simulation model, ADVISOR, to convert these loads to fuel use over the EPA's FTP test
cycle. They developed a personal "thermal comfort"-based model to predict the percentage of
drivers which will turn on their A/C systems under various ambient conditions. Overall,
NREL estimated A/C use to represent 5.5% of car and light truck fuel consumption in the
U.S.
In the second study, the California Air Resources Board (CARB) estimated the impact
of A/C use on fuel consumption as part of their GHG emission rulemaking.25 The primary
technical analysis utilized by ARB is summarized in a report published by NESCCAF for
CARB. The bulk of the technical work was performed by two contractors: AVL Powertrain
Engineering and Meszler Engineering Services. This work is founded on that performed by
NREL-OAP. Meszler used the same Delphi testing to estimate the load of the A/C
compressor at typical ambient conditions. The impact of this load on onroad fuel
consumption was estimated using a vehicle simulation model developed by AVL - the
CRUISE model - which is more sophisticated than ADVISOR. These estimates were made
for both the EPA FTP and HFET test cycles. (This is the combination of test cycle results
used to determine compliance with NHTSA's current CAFE standards.) NREL's thermal
comfort model was used to predict A/C system use in various states and seasons.
The NESCCAF results were taken from Table 3-1 of their report and are summarized
in Table 5-5.26
Table 5-5 CO2 Emissions Over 55/45 FTP/HFET Tests and From A/C Use (g/mi) Based on the NESCCAF
study
5-22
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
55/45 FTP/HFET
Indirect A/C
Fuel Use (g/mi CO2)
Total
Indirect A/C
Fuel Use
Small Car
278
16.8
294.8
5.7%
Large Car
329
19.1
348.1
5.5%
Minivan
376
23.5
399.5
5.9%
Small Truck
426
23.5
449.5
5.2%
Large Truck
493
23.5
516.5
4.6%
NESCCAF estimated that nationwide, the average impact of A/C use on vehicle fuel
consumption ranged from 4.6% for a large truck or SUV, to 5.9% for a minivan. The impact
on vehicle CC>2 and fuel consumption resulting from A/C use was determined using a
55%/45% weighting of CC>2 emissions from EPA FTP and HFET tests (hereafter referred to
simply as FTP/HFET). Simulation modeling to assess A/C related fuel consumption and CC>2
emissions was first conducted without the load from the A/C system followed by modeling
which included the load from the A/C system. For the purposes of this analysis of A/C system
fuel use, the percentage of CO2 emissions and fuel use are equivalent, since the type of fuel
being used is always gasoline.1
In the MYs 2012-2016 rule, there was a third analysis presented along with a thorough
comparison of these studies. While not repeated here, it was estimated that the impact of A/C
on onroad fuel consumption was 3.9% based on a combination of the results from these
studies. This resulted in an average impact of 14.3g/mi (independent of car or truck type) and
hence a maximum of 5.7 g/mi credit, identical for car and truck (based on a 40%
improvement feasibility). For this rule, EPA has conducted a new analysis, which supports
the results achieved in the MYs 2012-2016 final rule, though there is now a distinction made
between cars and trucks as it relates to A/C efficiency impacts (and credits).
5.1.3.2 Updated Analysis of Efficiency Impacts
As just mentioned, in the Light-Duty GHG final rule for model years 2012 through
2016, EPA estimated that the average CO2 emission increase due to A/C use would be 14.3
g/mi taking into account both manual and automatic climate control systems with market
penetrations of 62% and 38%, respectively. For this study of the A/C compressor load impact
on vehicle fuel economy, EPA relied on comparisons of measured fuel economy over two
warmed up bags (or phases) of the FTP test (without A/C operating) and the SC03 test (A/C
emissions test). EPA had based its estimates on testing of over 600 production vehicles.
These test results were combined with the Phoenix study, where the A/C compressor on-time
was estimated to be 23.9% for manual climate control systems and 35% for automatic climate
control systems. For more technical details, one can refer to the Regulatory Impact Analysis
for the model year 2012 to 2016 final rule.
Because NESCCAF estimated A/C fuel use nationwide, while ARB focused on that in California, the
NESCCAF and EPA methodologies and results are compared below.
5-23
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
For the proposed and final rule, EPA developed a more robust and systematic method
of estimating vehicle CO2 emissions related to A/C usage. This method is based on a
sophisticated, newly-developed EPA vehicle simulation tool. The next few paragraphs
provide an overview of the vehicle simulation tool and describe how this approach improves
on the earlier analysis. More detailed descriptions about the vehicle simulation tool and its
use for the A/C indirect impact analysis are in Chapter 2 of the EPA Regulatory Impact
Analysis.
Over the past year, EPA has developed full vehicle simulation capabilities in order to
support regulations and vehicle compliance by quantifying the effectiveness of different
technologies with scientific rigor over a wide range of engine and vehicle operating
conditions. This in-house simulation tool has been developed for modeling a wide variety of
light, medium, and heavy duty vehicle applications over various driving cycles. In order to
ensure transparency of the models and free public access, EPA has developed the tool in
MATLAB/Simulink environment with the completely open source code. To support these
simulation capabilities in part, EPA is upgrading its testing infrastructure (such as engine test
cells, vehicle dynamometers, Portable Emissions Measurement Systems, and a battery
laboratory) at the National Vehicle and Fuel Emissions Laboratory in Ann Arbor, Michigan.
This testing infrastructure provides necessary data to calibrate and validate vehicle
simulations, such as engine fuel maps, engine torque maps, vehicle aerodynamic parameters,
battery, electrical component parameters, etc.
EPA's first application of the vehicle simulation tool was for purposes of heavy-duty
vehicle compliance and certification. For the model years 2014 to 2018 final rule for medium
and heavy duty trucks, EPA created the "Greenhouse gas Emissions Model" (GEM), which is
used both to assess Class 2b-8 vocational vehicle and Class 7/8 combination tractor GHG
emissions and fuel efficiency and to demonstrate compliance with the vocational vehicle and
combination tractor standards. See 40 CFR sections 1037.520 and 1037.810. This GEM
documentation is also currently in publication.27
For light-duty vehicles, EPA has developed a conventional (non-hybrid) vehicle
simulation tool and used it to estimate indirect A/C CO2 emissions. These estimates are used,
in turn, to quantify the maximum amount of indirect A/C credit (i.e. the maximum credit
potential). As mentioned previously, the tool is based on MATLAB/Simulink and is a
forward-looking full vehicle model that uses the same physical principles as other
commercially available vehicle simulation tools (e.g. Autonomie, AVL-CRUISE, GT-Drive,
etc.) to derive the governing equations. These governing equations describe steady-state and
transient behaviors of each electrical, engine, transmission, driveline, and vehicle systems,
and they are integrated together to provide overall system behavior during transient conditions
as well as steady-state operations. Chapter 2 of EPA's Regulatory Impact Analysis provides
more details on this light-duty vehicle simulation tool used for estimating indirect A/C impact
on fuel consumption.
In the light-duty vehicle simulation tool, there are four key system elements that
describe the overall vehicle dynamics behavior and the corresponding fuel efficiency:
electrical, engine, transmission, and vehicle. The electrical system model consists of parasitic
electrical load and A/C blower fan, both of which were assumed to be constant. The engine
5-24
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
system model is comprised of engine torque and fueling maps. For estimating indirect A/C
impact on fuel consumption increase, two engine maps were used: baseline and EGR boost
engines. These engine maps were obtained by reverse-engineering the vehicle simulation
results provided by Ricardo Inc. For the transmission system, a Dual-Clutch Transmission
(DCT) model was created and used along with the gear ratios and shifting schedules used for
the earlier Ricardo simulation work. For the vehicle system, four vehicles were modeled:
small, mid, large size passenger vehicles, and a light-duty pick-up truck. The transient
behavior and thermodynamic properties of the A/C system was not explicitly simulated, in
favor of a simpler approach of capturing the compressor load based on national average
ambient conditions. We believe this simplification is justified since the goal is to capture the
behavior on the average of a fleet of vehicles (not the behavior of an individual make or
model).
In order to properly represent average load values to the engine caused by various A/C
compressors and vehicle types, EPA has adopted power consumption curves of A/C systems,
published by an A/C equipment supplier, Delphi. 8'29 Also, in an effort to characterize an
average A/C compressor load in the presence of widely varying environmental conditions in
the United States, EPA has adopted data from the National Renewable Energy Laboratory
(NREL) to estimate environmental conditions associated with typical vehicle A/C
usage.30'31'32 Based on the NREL data, EPA selected an A/C power consumption curve as a
function of engine speed that was acquired by Delphi at 27°C and 60% relative humidity as a
representative average condition. This power consumption curve data was taken from a fixed
displacement compressor with a displacement volume of 210 cc. Thus, the curve includes the
effect of compressor cycling as well as non-summer defrost/defog usage. In order to associate
each vehicle type with the appropriate A/C compressor displacement, EPA scaled the curve
based on the displacement volume ratio. For determining indirect A/C impact on fuel
consumption increase, EPA estimated A/C compressor sizes of 120 cc, 140 cc, 160 cc, and
190 cc for small, medium, large passenger vehicles, and light-duty pick-up truck, respectively.
By applying the displacement volume ratios to the 210 cc power consumption curve, EPA
created A/C load curves for four vehicle types, as shown in Figure 5-2.
A/C Load Demand
§ } SO -
-210 cc
-190 cc
-160 cc
-140CC
120 cc
500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500
Engine Speed (RPM)
Figure 5-2 Representative A/C Compressor Load Curves
5-25
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
With these A/C compressor load curves, EPA ran full vehicle simulations based on the
following matrix. In this matrix, the baseline engine represents a typical Spark-Ignition (SI),
Port-Fuel Injection (PFI), Naturally-Aspirated (NA) engine equipped with a Variable Value
Actuation (VVA) technology. In this technology, the valve timing (both intake and exhaust)
is continuously varied over a wide range of engine operating conditions in order to result in
optimal engine breathing efficiency. On the other hand, the EGR boost engine uses
turbocharging and cooled EGR to increase engine's Brake Mean Effective Pressure (BMEP)
level while managing combustion and exhaust temperatures. This engine usually has a peak
BMEP of 25 to 30 bar, which supports significant downsizing (e.g. about 50%) compared to
the baseline engines. Table 5-6 provides simulation results over SC03 driving cycle with an
EGR boost engine for various vehicle classes.
• Small, medium, large cars, and pick-up truck
• FTP, Highway, and SC03 cycles
• Baseline and EGR boost engines
• A/C off and A/C on
Table 5-6 Vehicle Simulation Results on CO2 Emissions over SC03 Cycle with EGR Boost Engine
SC03 Cycle
CO2 with A/C off
CO2 Increase with A/C on
Total CO2 with A/C
Indirect A/C Fuel Use
[g/mi]
[g/mi]
[g/mi]
[%]
Small Car
196.4
11.7
208.1
5.6
Medium Car
235.7
12.0
247.7
4.8
Large Car
293.7
13.8
307.5
4.5
Truck
472.4
17.2
489.6
3.5
EPA ran the SC03 cycle simulations instead of FTP/Highway combined cycle
simulations so that the simulation results would represent the actual A/C cycle test. EPA also
assumed the EGR boost engine during vehicle simulations because EGR boost engine better
represents engine technology more likely to be implemented in model years 2017 to 2025 and
because the A/C impact on CO2 increase in the EGR boost engine is similar to that in the
baseline engine as shown in Table 5-6 and Table 5-7. Details of this analysis which showed
impact of A/C usage on fuel consumption is independent of engine technology are provided in
the next section. Moreover, EPA assumed 38% of a market penetration for automatic climate
control systems as well as 23.9% and 35.0% of A/C on-time for manual and automatic climate
control systems, respectively. These are the same assumptions made in the MYs 2012-2016
rule.33 In order to come up with overall impact of A/C usage on CO2 emissions for passenger
cars, the simulation results for cars shown in Table 5-6 were sales-weighted for each model
year from 2017 to 2025. For the end result, the impact of A/C usage was estimated at 11.9
CO2 g/mile for cars and 17.2 CO2 g/mile for trucks. This corresponds to an impact of
approximately 14.0 CO2 g/mile for the (2012) fleet, which is comparable to the MYs 2012-
2016 final rule result, but still lower than the two studies by NREL and NESCCAF cited
above.
5-26
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
5.1.3.2.1 Effect of Engine Technology on Fuel Consumption by A/C System
In order to continue to maintain the credit levels from the MYs2012-2016 rule, EPA
had to first demonstrate that the fuel economy and CC>2 emissions due to A/C was relatively
insensitive to the engine technologies that may be expected to be prevalent in the future. If
for example, more efficient engines are able to run the A/C system more efficiently such that
the incremental increase in emissions due to A/C decreased compared to the base engines,
then credits for the same A/C technologies must decrease over time as engines become more
efficient. This would correspond to a decrease in credits proportional (or multiplicative) to
the increase in efficiency of the engine. Conversely, if the incremental increase in emissions
due to A/C remained relatively constant, then the credits available for A/C efficiency should
also remain stable. This would correspond to the credits (A/C impact) being additive to the
base emissions rate, thus being independent of engine efficiency). The EPA based the
hypothesis on the latter assumption.
In order to prove out this hypothesis, EPA carried out vehicle simulations for several
cases, including two engine technologies: baseline and EGR boost engines (a surrogate for a
future advanced efficient engine). Table 5-7 shows the vehicle simulation results of CO2
emissions over the SC03 driving cycle when baseline engines are used, as opposed to the
advanced EGR boost engines. By comparing the values of CO2 increase with A/C on in Table
5-6 and Table 5-7, it is evident that the impact of A/C usage on fuel consumption is not very
dependent on the engine technologies. In fact, the difference in the CO2 increase with A/C on
(2n row in table) between the emissions from the baseline and EGR boost engines is less than
10% for all vehicle classes.
Table 5-7 Vehicle Simulation Results on COi Emissions over SC03 Cycle with Baseline Engine
SC03 Cycle
CO2 with A/C off
CO2 Increase with A/C on
Total CO2 with A/C
Indirect A/C Fuel Use
[g/mi]
[g/mi]
[g/mi]
[%]
Small Car
259.3
11.3
270.6
4.2
Medium Car
348.0
11.1
359.1
3.1
Large Car
425.4
12.5
437.9
2.9
Truck
628.1
16.2
644.3
2.5
Figure 5-3 depicts zoomed-in BSFC maps for baseline and EGR boost engines. The
circles on these maps represent average operating conditions of the engines over the FTP
(city) drive cycle. The blue circle represents a simulated average operating condition without
A/C while the red circle represents an average operating condition with A/C. As can be seen
in the figure, the engines operate at higher load levels when the A/C is on. In this figure, the x
and y axes present engine speed in RPM and torque in Nm, respectively.
5-27
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
100
80
60
40
(a) Baseline Engine
(b) EGR Boost Engine
Figure 5-3 Average Engine Operating Conditions with A/C Off and A/C On over Fueling Maps for
Baseline and EGR Boost Engines
For the baseline engine case, the engine efficiency improves significantly (375 g/kW-h
to almost 330 g/kW-h) as it moves along the BSFC surface, whereas the improvement is
much less for the EGR boost engine as it moves from approximately 250 g/kW-h to 240
g/kW-h. However, the large improvement in AC efficiency for the baseline engine is offset
by the fact that the engine itself is less efficient than the EGR boost engine. Conversely, the
small AC efficiency improvement for the EGR boost engine is compensated by the fact that
the engine is much more efficient than the baseline engine. As a result, the CO2 increase seen
by both engines due to A/C usage becomes similar between the two different technologies.
This result allows us to approximate the A/C impact on vehicle fuel consumption as an
additive effect rather than as a multiplicative effect since it is independent of engine
technologies. For the same reason, it also means that A/C credits for a given technology can
remain constant over time, which will greatly simplify the progression of future credits.™
5.1.3.3 Technologies That Improve Efficiency of Air Conditioning and Their
Effectiveness
Most of the excess load on the engine comes from the compressor, which pumps the
refrigerant around the system loop. Significant additional load on the engine may also come
from electrical or hydraulic fan units used for heat exchange across the condenser and
m It also means that the last row in the above two tables are a bit misleading as A/C impact should not be
quantified as a fraction of the total emissions, but rather an additive increment. The numbers are left onto the
tables for comparison purposes to studies in the literature that use this convention.
5-28
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
radiator. The controls that EPA and NHTSA believe manufacturers would use to generate
credits for improved A/C efficiency and to improve fuel efficiency in the CAFE program
through the use of an adjustment in calculated fuel economy would focus primarily, but not
exclusively, on the compressor, electric motor controls, and system controls which reduce
load on the A/C system (e.g. reduced 'reheat' of the cooled air and increased of use
recirculated cabin air). EPA and NHTSA are finalizing a program that will result in improved
efficiency of the A/C system (without sacrificing passenger comfort) while improving the fuel
efficiency of, and reducing the CO2 emissions from, the vehicle.
The cooperative EVIAC program described above has demonstrated that average A/C
efficiency can be improved by 36.4% (compared to an average MY 2008 baseline A/C
system), when utilizing "best-of-best" technologies.34 EPA and NHTSA consider a baseline
A/C system to contain the following components and technologies; internally-controlled fixed
displacement compressor (in which the compressor clutch is controlled based on 'internal'
system parameters, such as head pressure, suction pressure, and/or evaporator outlet
temperature); blower and fan motor controls which create waste heat (energy) when running
at lower speeds; thermostatic expansion valves; standard efficiency evaporators and
condensers; and systems which circulate compressor oil throughout the A/C system. These
baseline systems are also extraordinarily wasteful in their energy consumption because they
add heat to the cooled air out of the evaporator in order to control the temperature inside the
passenger compartment. Moreover, many systems default to a fresh air setting, which brings
hot outside air into the cabin, rather than recirculating the already-cooled air within the cabin.
The EVIAC program indicates that improvements can be accomplished by a number of
methods related only to the A/C system components and their controls including: improved
component efficiency, improved refrigerant cycle controls, and reduced reheat of the cooled
air. The program EPA and NHTSA are finalizing will encourage the reduction of A/C CC>2
emissions from cars and trucks by up to 42% from current baseline levels through a CO2
credit and fuel economy improvement system. EPA and NHTSA believe that the component
efficiency improvements demonstrated in the EVIAC program, combined with improvements
in the control of the supporting mechanical and electrical devices (i.e. engine speeds and
electrical heat exchanger fans), can go beyond the EVIAC levels and achieve a total efficiency
improvement of 42% through incremental improvements beyond that shown in the study due
to the long lead time before MY 2017. The following sections describe the technologies the
agencies believe manufacturers can use to attain these efficiency improvements.
Based on the new vehicle simulation research conducted by the EPA described above,
the EPA believes that the impact of A/C on average CC>2 emissions amounts to 11.9 CC>2
g/mile for cars and 17.2 CC>2 g/mile for trucks (0.001339/0.001935 gallons of gasoline per
mile car/truck improvement) and that these results are relatively insensitive to the engine and
transmission efficiency improvements expected to be seen during the rule timeframe. A 42%
improvement on this emissions rate leads to the maximum credit opportunity of 5.0 g CO2/mi
for cars and 7.2 g CO2/mi for trucks (-0.000563 / -0.000810 gallons per mile car/truck
improvement). This compares to the 5.7 g/mi (identical for cars and trucks) finalized in the
2012-2016 final rule. When cars and trucks are combined, the new final rule maximum
credits are consistent (on a fleet level) with those finalized in the previous rule, though for
cars the credits are now somewhat reduced and for trucks increased. The agencies believe
5-29
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
that the modification of these credits for this rule is justified given the simulation work
conducted, which shows that A/C emissions tends to be larger for the larger vehicles (and
trucks tend to be larger than passenger cars).
The following sections discuss each of the A/C efficiency-improving technologies that
EPA recognizes in the efficiency credit menu. We estimated the effectiveness of each of
these technologies for the MYs 2012-2017 rule based on a variety of sources, including
testing under the EVIAC program and internal EPA testing. We did not receive comments
challenging these estimates and continue to base the efficiency credits (and CAFE
improvement values) on these estimates.
5.1.3.3.1 Reduced Reheat Using a Externally-Controlled, Variable-Displacement
Compressor
The term 'external control' of a variable-displacement compressor is defined as a
mechanism or control strategy where the displacement of the compressor adjusted
electronically, based on the temperature setpoint and/or cooling demand of the A/C system
control settings inside the passenger compartment. External controls differ from 'internal
controls' that internal controls adjust the displacement of the compressor based on conditions
within the A/C system, such has head pressure, suction pressure, or evaporator outlet
temperature. By controlling the displacement of the compressor by external means, the
compressor load can be matched to the cooling demand of the cabin. With internal controls,
the amount of cooling delivered by the system may be greater than desired, at which point the
cooled cabin air is then 'reheated' to achieve the desired cabin comfort. It is this reheating of
the air which results reduces the efficiency of the A/C system - compressor power is
consumed to cool air to a temperature less than what is desired.
Reducing reheat through external control of the compressor is a very effective strategy
for improving A/C system efficiency. The SAE IMAC team determined that an annual
efficiency improvement of 24.1% was possible using this technology alone.34 The agencies
estimate that additional improvements to this technology are possible (e.g. the increased use
of recirculated cabin air), and that when A/C control systems and components are fully
developed, calibrated, and optimized to particular vehicle's cooling needs, an efficiency
improvement of 42% can be achieved, compared to the baseline system.
5.1.3.3.2 Reduced Reheat Using a Externally-Controlled, Fixed-Displacement or
Pneumatic Variable-Displacement Compressor
When using a fixed-displacement or pneumatic variable-displacement compressor
(which controls the stroke, or displacement, of the compressor based on system suction
pressure), reduced reheat can be realized by disengaging the compressor clutch momentarily
to achieve the desired evaporator air temperature. This disengaging, or cycling, of the
compressor clutch must be externally-controlled in a manner similar to that described in
2.3.2.1. The agencies believe that a reduced reheat strategy for fixed-displacement and
pneumatic variable-displacement compressors can result in an efficiency improvement of
20%. This lower efficiency improvement estimate (compared to an externally-controlled
5-30
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
variable displacement compressor) is due to the thermal and kinetic energy losses resulting
from cycling a compressor clutch off-and-on repeatedly.
5.1.3.3.3 Defaulting to Recirculated Cabin Air
In ambient conditions where air temperature outside the vehicle is much higher than
the air inside the passenger compartment, most A/C systems draw air from outside the vehicle
and cool it to the desired comfort level inside the vehicle. This approach wastes energy
because the system is continuously cooling the hotter outside air instead of having the A/C
system draw its supply air from the cooler air inside the vehicle (also known as recirculated
air, or 'recirc'). By only cooling this inside air (i.e. air that has been previously cooled by the
A/C system), less energy is required, and A/C Idle Tests conducted by EPA indicate that an
efficiency improvement of 35-to-40% improvement is possible under idle conditions.
Ongoing testing on the new AC 17 test, described below, may in the future shed light on the
overall effectiveness of this technology during other driving conditions.
A mechanically-controlled door on the A/C system's air intake typically controls
whether outside air, inside air, or a mixture of both, is drawn into the system. Since the
typical 'default' position of this air intake door is outside air (except in cases where maximum
cooling capacity is required, in which case, many systems automatically switch this door to
the recirculated air position), EPA and NHTSA are specifying that, as cabin comfort and de-
fogging conditions allow, an efficiency credit be granted if a manufacturer defaults to
recirculated air whenever the outside ambient temperature is greater than 75°F. To maintain
the desired quality inside the cabin (in terms of freshness and humidity), EPA believes some
manufacturers will control the air supply in a 'closed-loop' manner, equipping their A/C
systems with humidity sensors or fog sensors (which detect condensation on the inside glass),
allowing them to adjust the blend of fresh-to-recirculated air and optimize the controls for
maximum efficiency. Vehicles with closed-loop control of the air supply (i.e. sensor
feedback is used to control the interior air quality) will qualify for a 1.7 g/mi CO2 credit and a
0.000124 gal/mi fuel consumption improvement. Vehicles with open-loop control (where
sensor feedback is not used to control interior air quality) will qualify for a 1.1 g/mi CO2
credit and a 0.000124 gal/mi fuel consumption improvement. We believe that the closed-loop
control system will be inherently more efficient than the open-loop control system because the
former can maximize the amount to recirculation to achieve a desired air quality and interior
humidity level, whereas the latter will use a fixed 'default' amount of recirculated air which
provides the desired air quality under worst case conditions (e.g. maximum number of
passengers in the vehicle).
Electric drive vehicles such as HEVs, PHEVs and EVs may require some fraction of
the A/C cooling capacity to control the battery temperature under hot conditions. PHEVs are
most likely to require A/C cooling because their batteries have higher current requirements for
all-electric driving than HEVs, and much less battery mass and energy storage than pure EVs.
Some electrified vehicles today, such as the Nissan Leaf, cool their batteries with outside air,
while others, such as the Toyota Prius and Ford Fusion Hybrid, use cooled cabin air, and the
Chevrolet Volt is an example of a vehicle which uses a refrigerant loop to cool the battery
coolant and thus to cool the battery. With the increased penetration of these electrified
vehicles, it is possible that there will be some loss of efficiency of the A/C system (especially
5-31
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
as it relates to cabin air recirculation). Vehicles which use cabin air to cool the battery must
discharge this heated air outside the vehicle, rather than recirculating it through the climate
control system. Currently, EPA does not account for this A/C efficiency loss in the credit
menu. EPA and NHTSA requested comments on the technical merits or applicability of
accounting for this loss of efficiency within the crediting and fuel economy improvement
scheme, however there were no comments submitted. The agencies are therefore not making
any adjustment in the credit menu for electric drive vehicles. As these types of vehicles
become more common, the agencies will continue to study the effectiveness of air
conditioning technologies.
5.1.3.3.4 Improved Blower and Fan Motor Controls
In controlling the speed of the direct current (DC) electric motors in an air
conditioning system, manufacturers often utilize resistive elements to reduce the voltage
supplied to the motor, which in turn reduces its speed. In reducing the voltage however, these
resistive elements produce heat, which is typically dissipated into the air ducts of the A/C
system. Not only does this waste heat consume electrical energy, it contributes to the heat
load on the A/C system. One method for controlling DC voltage is to use a pulsewidth
modulated (PWM) controller on the motor. A PWM controller can reduce the amount of
energy wasted, and based on Delphi estimates of power consumption for these devices, EPA
and NHTSA believe that when more efficient speed controls are applied to either the blower
or fan motors, an overall improvement in A/C system efficiency of 15% is possible.35
5.1.3.3.5 Internal Heat Exchanger
An internal heat exchanger (MX), which is alternatively described as a suction line heat
exchanger, transfers heat from the high pressure liquid entering the evaporator to the gas
exiting the evaporator, which reduces compressor power consumption and improves the
efficiency of the A/C system. In the MYs 2012-2016 rule, we considered that IHX
technology would be required with the changeover to an alternative refrigerant such as HFO-
1234yf, as the different expansion characteristics of that refrigerant (compared to R-134a)
would necessitate an IHX. The agencies believe that a 20% improvement in efficiency
relative to the baseline configuration can be realized if the system includes an IHX, and a 1.1
g/mi credit and a 0.000124 gal/mi fuel consumption improvement for an IHX.
5.1.3.3.6 Improved-Efficiency Evaporators and Condensers
The evaporators and condensers in an A/C system are designed to transfer heat to and
from the refrigerant - the evaporator absorbs heat from the cabin air and transfers it to the
refrigerant, and the condenser transfer heat from the refrigerant to the outside ambient air.
The efficiency, or effectiveness, of this heat transfer process directly effects the efficiency of
the overall system, as more work, or energy, is required if the process is inefficient. A
method for measuring the heat transfer effectiveness of these components is to determine the
Coefficient of Performance (COP) for the system using the industry-consensus method
described in the SAE surface vehicle standard J2765 - Procedure for Measuring System COP
of a Mobile Air Conditioning System on a Test Bench.36 The bench test based engineering
analysis that a manufacturer will submit at time of certification. We will consider the baseline
5-32
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
component to be the version which a manufacturer most recently had in production on the
same vehicle or a vehicle in a similar EPA vehicle classification. The design characteristics
of the baseline component (e.g. tube configuration/thickness/spacing and fin density) are to be
documented in an engineering analysis and compared to the improved components, along
with data demonstrating the COP improvement. This same engineering analysis can be
applied to evaporators and condensers on other vehicles and models (even if the overall size
of the heat exchanger is different), as long as the design characteristics of the baseline and
improved components are the same. If these components can demonstrate a 10%
improvement in COP versus the baseline components, EPA and NHTSA estimate that a 20%
improvement in overall system efficiency is possible.
5.1.3.3.7 Oil Separator
The oil present in a typical A/C system circulates throughout the system for the
purpose of lubricating the compressor. Because this oil is in contact with inner surfaces of
evaporator and condenser, and a coating of oil reduces the heat transfer effectiveness of these
devices, the overall system efficiency is reduced.37 It also adds inefficiency to the system to
be "pushing around and cooling" an extraneous fluid that results in a dilution of the
thermodynamic properties of the refrigerant. If the oil can be contained only to that part of
the system where it is needed - the compressor - the heat transfer effectiveness of the
evaporator and condenser will improve. The overall COP will also improve due to a
reduction in the flow of dilutent. The SAE EVIAC team estimated that overall system COP
could be improved by 8% if an oil separator was used. EPA and NHTSA believe that if oil is
prevented from prevented from circulating throughout the A/C system, an overall system
efficiency improvement of 10% can be realized. Whether the oil separator is a standalone
component or is integral to the compressor design, manufacturers can submit an engineering
analysis to demonstrate the effectiveness of the oil separation technology.
5.1.3.4 Technical Feasibility of A/C Efficiency-Improving Technologies
EPA and NHTSA believe that the efficiency-improving technologies discussed in the
previous sections are available to manufacturers today, are relatively low in cost, and their
feasibility and effectiveness has been demonstrated by the SAE EVIAC teams and various
industry sources. The agencies also believe that when these individual components and
technologies are fully designed, developed, and integrated into A/C system designs,
manufacturers will be able to achieve the estimated reductions in CO2 emissions and generate
appropriate A/C Efficiency Credits, which are discussed in the following section.
5.1.3.5 A/C Efficiency Test Procedures
As proposed, the agencies have two test procedures to determine eligibility for A/C
efficiency credits. The two test procedures are the Idle and the AC 17 test procedures. The
test procedures play different roles depending on the model year for which the test is
conducted. For model years 2014 to 2016, there are 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 CO2 threshold requirements, 2) running the A/C Idle
and demonstrating compliance with engine displacement adjusted CO2 threshold
5-33
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
requirements, and 3) running a newly-developed AC 17 test and reporting the test results. For
model years 2017-2019, the AC 17 test will be 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 2017-2021, 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.
Details of the AC 17 test requirements as well as the modified idle test thresholds are
described in detail in this section.
In the MYs 2012-2016 final rule, manufacturers were required, starting in MY 2014,
to demonstrate the efficiency of a vehicle's A/C system by running an A/C Idle Test as a
prerequisite to credit eligibility (the amount of credit determined separately by means of the
credit menu). If a vehicle met the emissions threshold of 14.9 g/min CC>2 or lower on this test,
a manufacturer was eligible to receive full credit 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. A
revised version of this technology menu is described below. 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 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) still apply for model years 2014-2016, so this final rule is not
making any fundamental changes to the previous rule. EPA is, however providing an optional
new threshold requirement for MYs 2014-2016 reflecting both the proposed rule and the
comments submitted on the idle test.
In order to establish the value of this eligibility threshold for the MYs 2012-2016 final
rule, the EPA conducted an extensive laboratory testing program to measure the amount of
additional CO2 a vehicle generated on the Idle Test due to A/C use. The results of this test
program are summarized in Table 5-8, and represent a wide cross-section of vehicle types in
the U.S. market. The average A/C CO2 result from this group of vehicles is the value against
which results from vehicle testing will be compared. The EPA conducted laboratory tests on
over 60 vehicles representing a wide range of vehicle types (e.g. compact cars, midsize cars,
large cars, sport utility vehicles, small station wagons, and standard pickup trucks).
Table 5-8 Summary of A/C Idle Test Study Conducted by EPA at the National Vehicle Fuel and Emissions
Laboratory
Vehicle Makes Tested 19
Vehicle Models Tested 29
5-34
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Model Years Represented (number of vehicles in 1999 (2), 2006 (21), 2007
each model year) (39)
EPA Size Classes Represented Minicompact, Compact,
Midsize, and Large Cars
Sport Utility Vehicles
Small Station Wagons
Standard Pickup Trucks
Total Number of A/C Idle Tests 62
Average A/C CO2 (g/min) 21.3
Standard Deviation of Test Results (+ g/min) 5_8
The majority of vehicles tested were from the 2006 and 2007 model years and their
A/C systems are representative of the 'baseline' technologies, in terms of efficiency (i.e. to
EPA's knowledge, these vehicles do not utilize any of the efficiency-improving technologies
described in the credit menu finalized for the MYs 2012-2016 rule). For the MYs 2012-2016
rule, EPA attempted to find a correlation between the A/C CO2 results and a vehicle's interior
volume, footprint, and engine displacement, but found it to be minimal, as there is significant
"scatter" in the test results. This scatter is generally not test-to-test variation, but scatter
amongst the various vehicle models and types. This original analysis covered a wide range of
vehicle size classes and vehicle types: from compact cars to light-duty trucks, some of which
did not have readily-available SAE and CAFE interior volume numbers (i.e. the interior
volume for small station wagons and pickup trucks had to be inferred from other published
sources). Due to the variability in the data, EPA chose a constant threshold value for the Idle
Test performance, which provided access to the credit menu.
Since the previous rule, manufacturers have had the opportunity to run the Idle Test on
a wide variety of vehicles and have 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.38 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.
When the EPA test data is sorted according to engine displacement, the relationship
between engine displacement and idle test result are somewhat apparent, though there is
significant variability as is evident in Figure 5-4. The threshold value from the MYs 2012-
2016 rule is included in the figure for comparison purposes.
5-35
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
14.9 g/min Idle Test Threshold
(forfull credit value)
3.0 4.0
Engine Displacement (L)
Figure 5-4 Relationship Between EPA A/C Idle Test Results and Engine Displacement.
One factor which may explain part of this observed phenomenon is that the brake-
specific fuel consumption (bsfc) of a smaller displacement engine is generally lower at idle
than that of a larger displacement engine. At the idle condition, without A/C load applied, a
smaller engine is generally more efficient (i.e. has a lower bsfc) than a larger engine, in terms
of how well it converts fuel heat energy into power. When additional load from the A/C
system is added to the small displacement engine, the bsfc does not improve as dramatically
as it does on a larger displacement engine, and if both the small- and large-displacement
engines require a similar amount of engine power to run the A/C system, the larger engine
will move from a "less-efficient" to "more-efficient" operating condition, whereas the smaller
engine remains relatively flat, in terms of bsfc. The result is that a larger displacement engine
uses less fuel to run the A/C system, relative to a smaller displacement engine, because its
baseline condition (A/C off) is "less-efficient", and the incremental amount of fuel used is
lower. The slope of the linear regression line for this data set is -1.58 g/min/L, with a zero
intercept of 26.9 g/min.n
In the MYs2012-2016 rule, the EPA chose a threshold of 30% improvement on the
Idle Test as the threshold for accessing the credit menu (the justification and feasibility
n The R2 for this fit is 0.09 reflecting the scatter and variability of the data. The slope is statistically significant at
the 2% confidence level (Significance F) indicating that the slope is statistically significant.
5-36
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
argument is presented again below). This corresponded to a 6.4 g/min reduction from the
average Idle Test result (20.5 g/min). Thus the 30% improvement is the average idle test
result (20.5 g/min) minus 30% (6.4 g/min) which equals 14.9 g/min (in the previous rule). In
this rule, EPA will maintain the 6.4g/min gap between the average emissions (equation of the
line) and the threshold. Equation 5-4 results in an idle test threshold which is scaled
according to engine displacement, in liters. The threshold equation is overlaid on the data in
Figure 5-5. Using this equation, the idle test threshold for a 1.2L engine for example (to
receive full credit) would be 18.6 g/min for a 6.0L engine the threshold would be 11.0 g/min.
Equation 5-4 - A/C Idle Test Threshold
In the MYs 2017-2025 NPRM, we acknowledged that the idle test may not fully
capture the effect of each and every technology, but believed that the test did reflect the
overall efficiency of the vehicle's A/C system under a commonly encountered operating
condition. See 76 FR at 74938. For this reason, we continue to allow the use of the original
Idle Test through model year 2016. In addition, we have now combined the Idle Test with a
displacement-adjusted "threshold," which some manufacturers wishing to use the Idle Test
may choose to apply. Overall, however, we now believe that the newly developed AC 17 test
is a more accurate method for determining A/C fuel use and CO2 emissions, and that the A/C
Idle Test requirement in both its forms can eventually be phased out (as described below).
We believe that part of the variation in the relationship between displacement and Idle
Test result that is evident in the figure above, was due to the type of components a
manufacturer chose to use in a particular vehicle. Components such as compressors are shared
across vehicle model types (e.g. a compressor may be 'over-sized' for one application, but the
use of a common part amongst multiple model types results in a cost savings to the
manufacturer), rather than being designed for one particular cabin size. Some of the variation
may also be due to the amount of cooling capacity a vehicle has at idle. For instance, if the
cooling capacity (or cooling performance) of a particular vehicle was less-than-optimal at idle
(due to factors such as limitations of the compressor design, pulley ratio, or packaging), this
vehicle could produce below-average A/C CO2 results, because the amount of energy required
by the compressor would be lower. Yet at higher engine and/or vehicle speeds, this same
vehicle may have cooling capacity typical of other vehicles. Therefore, a test which is limited
to one area of A/C operation is limited in its ability to determine overall A/C system
efficiency.
Some of this variation between vehicle models may also be due to the efficiency of the
fan(s) which draw air across the condenser - since an external fan is not placed in front of the
vehicle during the A/C Idle Test, it is the vehicle's radiator fan which is responsible for
rejecting heat from the condenser (and some models may do this more efficiently than others).
In this case, EPA believes that an SC03-type test - run in a full environmental chamber with a
"road-speed" fan on the front of the vehicle - would be a better measure of how a vehicle's
A/C system performs under transient conditions, and any limitations the system may have at
idle could be counter-balanced by improved performance and efficiency elsewhere in the
drive cycle.
5-37
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Since the MYs 2012-2016 final rule, EPA has received a number of idle test results
from several manufacturers. Testing by Ford, General Motors, and Chrysler has shown that
there are some significant limitations to the idle test procedure. As mentioned above, there
was significant test-to-test variability noted, and many vehicles - especially those with
smaller displacement engines - failed to meet the required test threshold (14.9 g/min) to
qualify for A/C credits - even when such vehicles are equipped with a significant number of
efficiency-improving technologies listed in the menu. These tested vehicles were from
upcoming model years and had a variety of air conditioner components and controls strategies
(from among the technologies described above and in the menu) implemented. The results
are shown in the Table 5-9 and are printed with permission from the manufacturers.
Table 5-9 A/C Idle Test Results from Various Manufacturers
Engine Displacement (liters)
1.4
2.0
2.0
2.4
2.4
3.5
3.6
3.6
5.7
A/C Idle Test Result
(gCO2/min)
19.4
22.4
20.0
28.0
18.3
12.0
24.0
16.0
26.0
The test-to-test variability observed by the manufacturers was significant, and is likely
due to high dilution of the exhaust sample (exhaust mass flow is low at idle), which results in
greater measurement error, as there is less CO2 present for sampling than there would be
under normal operating conditions. Furthermore, fluctuations in cell ambient conditions (e.g.
temperature and humidity), or in the way the driver is positioned in the seat, make accurate
test-to-test comparisons of the results difficult to achieve. In Figure 5-5, these new data
points from the manufacturers are overlaid onto the idle test data collected in support of the
MYs 2012-2016 final rule by the EPA. Most of the EPA vehicles tested over the past two
years did not contain a significant amount of efficient air conditioner components (off of the
menu list). The manufacturer data is largely consistent with the EPA data. The data support
the notion that it might be more appropriate to use an increasing function of emissions as a
function of engine displacement for a threshold, rather than the flat function we finalized in
the MYs 2012-2016 rule.
The test cells on which Idle Tests are conducted are typically the same cells which are
used for FTP testing for criteria pollutants, where the allowable ambient temperature is 68-to-
86 °F, and there is no humidity specification. Since there are normal, seasonal fluctuations in
humidity level for this type of test cell, controlling the ambient conditions to those specified
in the Idle Test procedure is difficult. EPA is modifying the allowable ambient air
temperature condition from to 75 ± 2 °F on average to 73-to-80°F on average, and the
ambient humidity within the test cell be modified from 50 ± 5 grains of water per pound of
dry air to 40-to-60 grains of water per pound of dry air.
5-38
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
E
--^.fieso/ts
FRM
*
• *
Threshold 14.9g/min Idle Test Threshold (for J $ 4"
fullcredit value in 2012 rule) X
3.0 4.0
Engine Displacement (L)
EPA Data X OEM Data
Linear (New Threshold)
Figure 5-5 EPA A/C Idle Test Results with Results from Various Manufacturers
Based on manufacturer data, with the revised threshold, it is still possible for a vehicle
test to have some A/C technologies but still fail to meet the threshold for the credit menu. For
the 2014 to 2016 model years, where a manufacturer chooses to run the A/C Idle Test, EPA is
allowing partial credits for vehicles that fail to meet the threshold but that show an
improvement over the baseline To qualify for the full credit, it will be necessary for each
vehicle certified to achieve an A/C CC>2 result less than or equal to the threshold function. As
previously described, the threshold function is 30% less than the average value observed in
the EPA testing. EPA chose the 30% improvement over the "average" value to drive the fleet
of vehicles toward A/C systems which approach or exceed the efficiency of best-in-class
vehicles. EPA test results on three vehicle size classes (large car, SUV, and pickup truck)
indicate that significant reductions in fuel consumption can be achieved by simply switching
A/C control from outside air (OSA) to recirculated cabin air. As shown in Table 5-11, the
percentage reduction in the CO2 and fuel consumption due to A/C use was greater than 30%
in all three cases.
Table 5-10 Effect of Outside Air and Recirculated Cabin Air on A/C Idle Test Results (EPA Testing)
Vehicle Type
Large Car
SUV
A/C CO2 R(
w/Outside Air
25.9
17.4
;sult (g/min)
w '/Retire Cabin Air
14.0
11.4
Change in A/C CO2
w/Recirc (%)
-45.9
-34.5
5-39
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
[Pickup Truck
14.1
9.0
-36.2 1
EPA believes the threshold approach will cause manufacturers to tailor the size of A/C
components and systems to the cooling needs of a particular vehicle model and focus on the
overall efficiency of their A/C systems. However, as explained above, the agency is
establishing as proposed an intermediate level of credit for vehicles that do not meet or
exceed the Idle Test threshold (either the original fixed threshold from the earlier rule or the
displacement-adjusted threshold finalized in this rule), so as not to set an all or nothing
threshold to qualify for credits. EPA will allow an intermediate amount of credit as long as
the Idle Test performance remains better than the best fit regression obtained from EPA
testing. A multiplier would be applied to the credits (based on the menu) such that if the
difference between the Idle Test result and the threshold value (hereafter referred to as the
"gap") at the vehicle's engine displacement is greater than 6.4g/min, then the multiplier would
be 1.0. If the gap is 0.0 g/min or less, then the multiplier would be 0.0, and the multiplier in
between would follow a linear function as shown in the following figure. Figure 5-6 shows
the change in credit received (y axis) as the idle test result in g/min (x axis) decreases. In the
Figure, as the Idle test result (in g/min) decreases, the difference between the Idle test
measurement and the threshold increases. The EPA is finalizing an option that allows
manufacturers the option of using these threshold adjustments as early as MY 2014.
1 2
In
•U
'T 0 8
|e.
*ff
~s n e.
,.0.6
_.£
'•5
f U.'l
U
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
quantify the incremental improvement of a given technology to generate an actual credit over
non-idle operation (without a menu). In order to generate a credit value a more complex test
procedure is required that can do an "A" to "B" comparison, where B is with the technology
and A is without. There were comments from a number of stakeholders both after
promulgation of the MY 2012-2016 rule and in response to the proposed 2017-2025 model
year standards reiterating some of these limitations. In preparation for the 2017-2025 NPRM,
EPA initiated a study that engaged automotive manufacturers, USCAR, component suppliers,
SAE, and CARB in developing a new test procedure for determining A/C system efficiency
and credits. 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. The
result of this effort is the new AC 17 test, 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. EPA believes that this new test procedure more accurately reflects the
impact that A/C use (and in particular, efficiency-improving components and control
strategies) has on tailpipe CO2 emissions. EPA proposed use of this test, to be phased in
starting in MY 2014 as an option, in MYs 2017-2019 as the exclusive means of determining
eligibility for A/C efficiency credits, and thereafter as both an eligibility test and as a partial
means of determining credit amount. See 76 FR 74938, 74940.
The new AC17 test has four elements: a pre-conditioning cycle, a 30-minute solar
soak period; Bag 1 is an SC03 drive cycle at 77 °F (to capture the "pull-down," or post-soak,
interior cool-down portion of A/C operation0); and Bag 2 is a highway fuel economy cycle (to
capture the "steady-state" portion of A/C operation). The test cycle is first run with the A/C
on (Bags 1 and 2) and then re-run with the A/C off (Bags 3 and 4). The A/C-related CO2
emissions are the difference between the A/C on and A/C off test results, with both bag results
being averaged (i.e. weighted equally), and the difference between A/C on and A/C off
averages producing the overall test result for the vehicle, which represents the incremental
CC>2 emissions due to operation of the A/C system. The incremental pull-down and steady-
state test results will be reported separately, as well as an average of these two results. EPA
believes that this new test cycle will be able to capture improvements in all areas related to
efficient operation of a vehicle's A/C system: solar control; efficiency improving components;
and efficient control strategies. Below is a depiction of the new test cycle. To assure
consistent results for the fuel consumption effect of operating the A/C system, the test is
always run in a warm condition, where an EPA Urban Dynamometer Driving Schedule
(UDDS) cycle is run at the start of the test sequence, with the A/C off and the solar lamps on.
Immediately following this precondition phase, the engine is turned off and the vehicle soaks
for 30 minutes with the solar lamps on. At the conclusion of the solar soak, the "pull-down"
(rapid cool-down of cabin temperature) phase begins. This phase utilizes the existing SC03
0 The pull-down period, is the time during which the cabin goes from an elevated temperature state (after having
soaked in the heat and sun) to the steady state interior temperature conditions requested by the vehicle occupants.
5-41
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
drive cycle to simulate dynamic, urban driving conditions. Finally, the highway fuel
economy test cycle, or HFET, is used to simulate a "steady-state" driving condition, while the
A/C system is maintaining the cabin temperature. Each element of this cycle exercises modes
of operation seen in everyday use where cabin cooling is needed. By running the vehicle
through each of these conditions with and without the A/C system operating, we seek to
understand the effect that soak, pull-down, and steady-state conditions have on the fuel
consumption for a particular A/C system design or technology. As shown in Figure 5-7, the
total time required to run this test on a single vehicle is approximately 4 hours (including A/C
on and A/C off phases).
80
Q.
I 60 1
o
Q.
CO
40 -
20
AC17 - MAC Efficiency Phase Timing
EPA 74 Preconditioning and MAC Cycle - Solar ON and AC ON
Solar and AC OFF
Vehicle and Site
Preparations EPA74
30 minutes f A -^
Solar Load = 850 W'm1 for 53 minutes
Hot Solar Soak srn, HWFpT
OFF
Analyze Dal
Preconditioning
30 minutes
Wi'windows
closed
Phase 1 Pnase2
Ban1 Bag 2
15 minutes
Site Conditions:
77 F, 50% RH
Road Speed Fan
505s
866s
600s
765s
Fan at Road Speed
Fan Speed at 4 mph
Vehicle Conditions:
Windows open or closed
Interior 77 2F
Cold or Hot Start
Fan at Road Speed
AC ON
1800 3600 5400
Time (seconds) - 2 hour block of site time
7200
Figure 5-7 AC17 Test
In the NRPM, EPA sought comment on the ambient conditions and system control
settings proposed for this test. The ambient temperature for the AC 17 test is 77+ 2 °F
average, and 77+ 5 °F instantaneous, with a humidity level in the test cell of 69 + 5 grains of
water per pound of dry air average, and 69 + 10 grains instantaneous. These ambient
temperature and humidity conditions for the AC 17 test were chosen because we believed that
they represented a common operating condition for A/C use, and that they would allow the
effect of many technologies to be demonstrated. The high temperature and humidity
conditions of the current SC03 test (95 °F and 100 grains of water per pound of dry air),,
while encountered in certain parts of the United States, does not adequately demonstrate the
5-42
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
impact of technologies such as variable displacement compressors, as their efficiency benefit
is more evident under lower cooling demand conditions. Several commenters noted that that
the temperature and humidity tolerances of the test cell conditions may result in voided tests,
due to the difficulty in maintaining these conditions throughout a 4-hour test interval. The
Alliance asked that humidity requirements be relaxed to avoid voided tests from the proposed
69 + 5 average and 69+10 instantaneous grains of water/pound of dry air. They also asked
that ambient temperatures be relaxed from the proposed 77 + 3 degrees on average and 77 + 5
degrees instantaneous for 95% of the time. They stated that SC03 test facilities were not
designed to operate at these temperature and humidity conditions at the required solar load of
850 W/m . Ford supported the Alliance comments on AC 17 temperature and humidity
requirements. However, we believe that widening these tolerances would negatively affect
the accuracy of the test - producing either too-high or too-low results for A/C-related CO2
emissions. As such, we are retaining the proposed tolerances for temperature and humidity as
proposed, but will stipulate that these tolerances apply only during the emissions
measurement portions of the test, and temperature and humidity deviations during the non-
emissions measurement portions of the test (e.g. preconditioning and soak) may exceed these
tolerances, as long at the duration of the deviation is no more than three minutes. In addition,
we will allow manufacturers to use a 30-second moving average on the temperature tolerance,
instead of an instantaneous temperature value. We will continue to investigate over time
whether temperature and humidity correction factors on the AC 17 test results are appropriate,
and if so, we will consider their inclusion in the future.
Commenters also noted that an allowable wind speed for the AC 17 test cell should be
specified for soak and idle conditions, as some air movement within the cell may be necessary
to assure proper control of the air temperature and humidity. The agencies agree with this
suggestion, and we will allow a nominal wind speed of 4 miles per hour or less during the
soak and idle portions of the test cycle. The agencies are also clarifying that the solar lamps
are off during the soak prior to the A/C-off portion of the AC 17 test (Phases 1 and 2).
Concerning the 30-minute soak time, this is a nominal time requirement, and not a precise
interval, as is the test involves coordination of the operator exiting and entering the vehicle at
the beginning and end of the soak period, as well as activation of the solar lamps. Also, if the
windows on the vehicle need to be partially open to allow for instrumentation and wiring to
be passed through to the interior, a temporary seal (e.g. foam or tape) can be used to close the
window gap.
The control settings for the "A/C ON" portions of the test (Phases 3 and 4 in Figure
5-7) are different for systems with automatic and manual climate controls. Automatic
systems will be set to a 72 °F target temperature, with blower (or fan) speed and vent location
controlled by the automatic mode. Manual systems will set the temperature selector to full
cold, blower speed at its highest setting, and the air supply set to "recirculated air" for the first
185 seconds of Phase 3. At the first idle of Phase 3 (186 to 204 seconds), the blower speed
will be set to achieve a nominal 6 volts at the motor, temperature selector will be set to
provide 55 °F at the center dash outlet, and the air supply set to "outside air". The
recommended temperature selector and blower control positions for manual systems will be
identified by the manufacturer.
5-43
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
5.1.3.7 Analysis of EPA's AC17 Testing
EPA has conducted 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 CC>2 emissions
which are expected from A/C technologies on the efficiency credit menu. EPA tested six
vehicles, including three pairs of carlines with some element of different air conditioner
systems. The vehicles and A/C technologies evaluated (and yet to be evaluated) in the EPA
test program are shown in Table 5-11.
Table 5-11 EPA AC17 Test Vehicles and Technologies
YEAR
2009/2010
2011
2011
2012
2009
2012
2009
2010
2011
2012
2011 or
2012
2011 or
2012
2011
2012
MAKE/MODEL
Dodge Journey 2.4L
Dodge Journey 2.4L
Chrysler 300C 5. 7L
Chrysler 300C 5. 7L
Dodge Caliber 2. 4L
Dodge Caliber 2.4L
GM Silverado
GM Sierra
Buick LaCrosse
Buick LaCrosse
Ford Fiesta Hatchback
Ford Fiesta Hatchback
Ford Edge or Ford Explorer
Ford Edge or Ford Explorer
SPECIAL HVAC DETAIL/FEATURES (if applicable)
New condenser design
Default to re-circulated air above
75 °F
Switch from orifice tube to TXV +
circulated air above 75 °F
default to re-
Reduced reheat
Reduced reheat
6 speed auto
6 speed auto with SFE (Super Fuel Economy)
3.5L
2.0L Eco-Boost
5.1.3.7.1 Overview:
In order to verify the appropriateness of the AC 17 test as a method to estimate the
relative efficiency of different A/C systems in a particular vehicle model, EPA is currently
carrying out vehicle testing at the Mercedes-Benz Tech Center located in Ann Arbor,
Michigan. The AC17 test is run in a SC03-capable test cell with a road-speed fan and solar
lamps, with ambient conditions of 77°F and 69 grains of water per pound of dry air (about
50% relative humidity).
As noted above, each segment ("A/C On" and "A/C Off) of the AC 17 test procedure
has four phases: a UDDS preconditioning drive cycle; a 30 minute soak period with a 4MPH
wind speed; a SC03 transient drive cycle; and a HWFET quasi-steady-state drive cycle. The
5-44
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
r\
test procedure is run twice: first with the A/C "On" and 850 W/m solar load during the soak
period and SC03 and HWFET drive cycles; and second with the A/C "Off and no solar load.
The incremental CO2 emissions related to fuel used to operate the A/C system is
obtained by subtracting the CO2 emissions when the test procedure is run with the A/C "on"
from the CO2 emissions when the test procedure is run with the A/C "off.
EPA ran AC 17 tests on a paired set of the same vehicle model of different model years
where an A/C system redesign has occurred between models. The purpose of the testing is to
verify the appropriateness of the AC 17 test as a method to estimate the relative efficiency of
different A/C systems. The AC 17 test was repeated between three to five times per vehicle to
validate the test-to-test accuracy and feasibility of the AC 17 test procedure.
Although the AC 17 testing program was ongoing at the time of this final rule, the test
results of the first three sets of vehicles appear to reinforce the value of the AC 17 test for the
purposes of this rule. EPA testing thus far shows low test-to-test variability, we were able to
quantify a CO2 increment between A/C "on" and A/C "off," and we were able to establish a
relative CO2 emissions difference between two A/C systems.
5.1.3.7.2 Summary of Testing To Date
Buick LaCrosse:
The 2011 Buick LaCrosse AC 17 test was repeated three times with A/C "off and
another three times with A/C "on." The average A/C "off CO2 emission result was 248 g/mi,
and the average A/C "on" CO2 emission result was 267 g/mi, thus resulting in a difference
("delta") of 19g/mi. These results show a very low test-to-test variability. Similarly, the
average A/C "off fuel economy was 33.5 mpg, and the average A/C "on" fuel economy was
31.1 mpg, or about a 7% reduction in average fuel economy).
EPA also tested a 2012 Buick LaCrosse that has a reduced reheat strategy and an A/C
economy mode that will turn "on" and "off the compressor. For the reduced reheat strategy
A/C system on this vehicle, the average A/C "off CO2 emission result was 221g/mi, and the
average A/C "on" CO2 emission result was 255.8 g/mi; thus resulting in a difference ("delta")
of 35g/mi. The reduced reheat strategy (which can generate a CO2 credit of up to 1.5 g/mi for
cars and 2.2 g/mi for trucks) should have resulted in a lower delta on the AC 17 test, but in this
case, the 2012 vehicle has an automatic start-stop feature, which turns off the engine when the
vehicle is stopped and the A/C is off. This engine-off feature resulted in the 2012 vehicle
having an A/C off CO2 emissions result that was 27 g/mi lower than the 2011 vehicle's, which
resulted in a larger A/C on-to-A/C off delta for the2012 vehicle. This does not necessarily
indicate that the 2012 vehicle's A/C system is less efficient than the 2011 system, but
illustrates that AC 17 is valid only when comparing vehicles with similar A/C and powertrain
systems. When the 2012 vehicle was run with the A/C on and the "ECO" mode enabled
(ECO mode is activated via a dash button, which when pressed, allows the engine start-stop
feature to function under certain conditions while the A/C is on), the AC17 result was 241.6
g/mi (14.2 g/mi lower than the non-ECO-mode result), resulting in a delta of 21 g/mi, which
is much closer to the 19 g/mi result observed on the 2011 vehicle. In the case of the 2011 to
5-45
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
2012 Buick LaCrosse comparison, an engineering analysis would be required to demonstrate
that additional technology (or technologies) present on the vehicle result in improved
efficiency of the A/C system.
GM Silverado:
The second pair of vehicles was a 2009 GM Silverado and a 2010 GM Sierra. Both
vehicles have the same platform. The 2009 GM Silverado had a manual A/C system and the
2010 GM Sierra had an A/C system with and automatic reheat reduction strategy. The
Silverado had average A/C "off CO2 emissions of 444 g/mi and fuel economy of 18.9 mpg,
and average A/C "on" CC>2 emissions of 481 g/mi and fuel economy of 17.4 mpg. This
corresponds to a CO2 emissions delta of 37g/mi. Again, the test-to-test variability was low.
The 2010 Sierra, had average A/C "off CC>2 emissions of 410 g/mi and fuel economy
of 20.8 mpg, and average A/C "on" CO2 emissions of 452.2 g/mi and fuel economy of 18.5
mpg, with low test-to-test variability. The CC>2 emissions delta is thus 41 g/mi.
Here, the AC 17 test between the redesigns was closer in value. However, the "more
efficient" 2010 system delta was still higher than the "less efficient" 2009 system, which on
the surface, seems counterintuitive. However, the A/C system settings on the AC 17 test for
'automatic' and 'manual' systems are not equivalent, and comparing the results between these
systems not valid for the purpose of demonstrating the effect of particular A/C technologies.
Where possible, we expect that manufacturers will use the AC 17 test for comparing the
performance of vehicles with identical or substantially similar A/C control systems (manual
vs automatic for example).
Further analysis of the AC 17 test results has shown that there are differences in the
A/C system operation on automatic- and manually-controlled systems which explain why
these differences can affect the A/C load applied to the engine, and the resulting CO2
emissions. EPA is examining data such as instrument panel temperature, compressor inlet
temperature, coolant temperature, engine control algorithms, recirculation control to
understand test result differences among similar vehicles, and trying to identify patterns
which may improve our understanding of how AC 17 test results and/or engineering analyses
will be used to demonstrate the effectiveness of advanced A/C technologies.
Chrysler 300
The third pair of vehicles was the 2011 and 2012 Chrysler 300C with a 5.7L engine.
The 2011 Chrysler 300C tested is a rear wheel drive vehicle, and the 2012 Chrysler 300C is
an all wheel drive vehicle. The 2012 Chrysler 300C A/C system has a default to
automatically recirculate air when the cabin temperature is 75°F.
The 2011 Chrysler 300C had average A/C "off CO2 emissions of 328.4 g/mi and fuel
economy of 25.4 mpg, and average A/C "on" CC>2 emissions of 358 g/mi and fuel economy of
23.4 mpg. This corresponds to a CC>2 emissions delta of 30 g/mi.
5-46
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
The 2012 Chrysler 300C had average A/C "off CO2 emissions of 348.7 g/mi and fuel
economy of 24.2 mpg, and average A/C "on" CC>2 emissions of 378.5 g/mi and fuel economy
of 22.1 mpg. The CC>2 emissions delta is thus 30 g/mi.
Again, for both vehicles, the test-to-test variability was low. Here, both systems have
the same CC>2 emissions delta. Possibly due the all wheel drive driveline, the A/C "off CC>2
emissions of the 2012 vehicle are higher than the 2011 vehicle, and therefore, the 2012
vehicle has a comparatively more efficient A/C system operation than the 2011 vehicle, as
both vehicles had the same delta of 30 g/mi CC>2.
In each of the three vehicle comparisons, there were confounding factors which
prevented a direct assessment of the A/C technology alone: in the case of the GM trucks, it
was automatic vs. manual control of the A/C system; in the case of the Buick LaCrosse, it was
ECO mode vs. non-ECO mode; in the case of the Chrysler 300, it was rear wheel drive vs. all
wheel drive. In follow-on testing, EPA will be testing vehicle pairs with A/C control
strategies, powertrains, and drivelines which are as close to identical as possible. The
preliminary EPA testing has shown 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. EPA also believes that 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 will analyze the data from the other vehicles as soon as they are
collected, and in the future, EPA plans to collect more data on this test procedure and monitor
the results of the reported results of AC 17 starting in MY 2014.
5.1.3.8 Options for Generating A/C Efficiency Credits
In MYs 2014-2016, to demonstrate that a vehicle's A/C system is delivering the
efficiency benefits of the new technologies on the credit menu, instead of running the Idle
Test, manufacturers will have the option to run the AC 17 test procedure on each vehicle
platform that incorporates the new, credit-generating technologies, and report the results from
all 4 phases of the test to EPA. In addition to reporting the test results, EPA is requiring that
manufacturers provide information about each test vehicle and its A/C system (e.g. vehicle
class, model type, curb weight, engine size, transmission type, interior volume, climate
control type, refrigerant type, compressor type, and evaporator/condenser characteristics).
For model years 2017 and beyond, EPA is eliminating the A/C Idle Test and threshold
requirement, and replacing it with the AC 17 test. Thus, for MYs 2017 and beyond,
manufacturers will run the AC 17 test to quantify the A/C-related CO2 emissions on a limited
number of vehicles. For model years 2017 through 2019, to access the A/C credit menu (i.e.
to be eligible to generate A/C efficiency credits), manufacturers will report the results of
AC 17 test results on the required number of vehicles to EPA as part of the certification
process. For model years 2020 and beyond, manufacturers will be required to demonstrate
that the results (delta between A/C on and A/C off) of the new model year vehicle are lower
than the results (delta) from a previously-tested 'baseline' vehicle. This comparison helps to
validate that the A/C technologies for which credit is generated are actually reducing A/C-
related CO2 emissions, To receive the full amount of A/C credit generated from the menu, the
5-47
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
difference between the new and baseline results should be equal to or greater than the sum of
the menu-based credits for technologies present on the new vehicle. The baseline vehicle
should be one with characteristics which are similar to the new vehicle, but not be equipped
with A/C efficiency-improving technologies, or be equipped with those technologies without
the technologies being implemented (e.g. forced cabin air recirculation). This baseline
vehicle would be one previously tested by the manufacturer, where AC 17 results were
reported to the EPA, and will be from the same platform, but from a prior (re)design.
We recognize that it may not be possible to find a baseline vehicle that is identical (in
terms of powertrain characteristics, as well as aerodynamic and parasitic losses) to the new
vehicle. The Alliance and others commented that any comparison to a prior redesign would
be an unfair comparisons because of other changes on the vehicle that may have occurred.
However, as we described in section 5.1.3 above and Chapter 5 of the EPA RIA, based on the
simulated behavior of A/C systems in a variety of vehicles, we believe that the amount of fuel
used to operate the A/C system is largely dependent on the compressor size and cooling
capacity of the system, and much less dependent on engine displacement or efficiency. As
such, we believe that it is technically appropriate for manufacturers to compare vehicles from
different generations of redesign cycles in order to demonstrate that their efficient A/C
systems can provide CC>2 and fuel consumption reductions commensurate to the amount of
credit that a particular vehicle can generate.
For cases where a comparison of a baseline vehicle to a new vehicle with additional
A/C technologies (which generate additional credits) is possible, the difference between the
baseline and new AC 17 test results must be equal to or greater than the amount of additional
credit generated for the new vehicle for the vehicle to receive the full credit (based on the
technology menu). In cases where the A/C technologies between the baseline and new
vehicle are identical, we expect that the difference between the baseline and new vehicle
AC 17 test results should be near zero. If the difference in AC 17 test results on this "same
technology" comparison was greater than zero (i.e. the "new" vehicle had greater AC17
emissions than the baseline vehicle), or in cases where no baseline comparison test result is
available (e.g. a brand-new platform has been created), we will require that manufacturers
submit an engineering analysis that justifies the generation of credits. The engineering
analysis would describe why the new vehicle had higher AC 17 results, or why a comparison
to a baseline vehicle AC 17 test result is neither available nor appropriate, and why the
generation of credits in either case is justified
However, starting in MY 2020, if the difference between the baseline and new AC 17
test results is less than the sum of the menu credit generated for the new technology (or
technologies), and an engineering analysis cannot justify a higher-than-expected AC 17 test
result, partial credit can still be generated. A manufacturer can use the credit scaling factor in
Equation 5-5, which is proportional to the ratio of the difference in the AC 17 test results
divided by the menu credit amount and can be applied to the new technology menu credits as
follows::
Equation 5-5 - AC17 Credit Scaling Factor
5-48
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Credit Scaling Factor= (Baseline AC17 Result-New AC17 Result) + Sum of New
Technology Menu Credits
For MY 2017 (and optionally for MYs 2014-2016), the AC 17 testing will first be
required on the highest-production-volume configuration from each platform for which
credits are generated. Because the new A/C test requires significant amount of time for each
test (nearly 4 hours) and must be run in SC03-capable facilities, EPA believes that it is
appropriate to limit the number of vehicles a manufacturer must test in any given model year
by limiting the testing to no more than one vehicle from each platform for each model year.
For the purpose of the AC 17 test and generating efficiency credits, a single platform will be
defined in a manner similar to that which EPA has used to define a "car line" (see 40 CFR
§600.002-08), and reads as follows:
"Platform " means a segment of an automobile manufacturer's vehicle fleet in which
the vehicles have a degree of commonality in construction (primarily in terms of body and,
chassis design). Platform does not consider the model name, brand, marketing division, or
level of decor or opulence, and is not generally distinguished by such characteristics as
powertrain, roof line, number of doors, seats, or windows. A platform may include vehicles
from various fuel economy classes, including both light-duty vehicles and light-duty
trucks/medium-duty passenger vehicles.
This definition was modified from the proposal based on comments received. In particular,
we agree with the Alliance that vehicle powertrain is not a key characteristic of a platform for
purposes of this rule, and the final definition of "platform" clarifies this.
EPA recognizes that a vehicle manufacturer may only utilize one or two A/C system
designs across many vehicle models within a platform, and it is not the intention of EPA that
a manufacturer measure the performance of each A/C system design on each model within the
platform, but simply that each A/C system design be tested on the highest expected sales
volume configuration within each platform, as defined above. For the first year in which a
manufacturer performs AC 17 testing - either as required in model year 2017, or as an option
to the Idle Test in model years 2014 through 2016 - what the manufacturer expects to be the
highest sales volume configuration from a given platform will need to be tested. In
subsequent model years, a unique A/C system design, where it exists and has not yet been
tested in this program, would be tested, continuing each model year until all A/C system
designs have been tested in their highest expected sales volume models.
For the purpose of this rule, a "unique A/C system design" will be defined as one in
which substantially-different A/C component designs or types and/or system control strategies
are used (e.g. fixed-displacement vs. variable-displacement compressor, orifice tube vs.
thermostatic expansion valve, manual vs. automatic climate control, single vs. dual
evaporator, etc.). A/C system designs which have similar cooling capacity, component types,
and control strategies, yet differ in terms of compressor pulley ratios or condenser or
evaporator surface area will not be considered to be unique designs. The test results from one
system design will apply to all design variants. EPA will require that manufacturers use good
5-49
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
engineering judgment to identify the unique A/C system designs which will require AC 17
testing in subsequent model years. Manufacturers would indicate the basis for their
engineering judgment at certification.
Starting in MY 2017, for each model year in which a manufacturer is using the AC 17
test (including optionally in MYs 2014-16), no more than one vehicle from each platform will
need to be tested on the AC 17 test. A manufacturer may choose to perform replicate tests (to
address concerns about test-to-test variability or to generate more robust data to support
credits for later use), but data from a single test is acceptable, As long as the necessary AC 17
tests are performed each model year, the credits generated for all model configurations within
a given platform can be carried over until there is a significant change in the platform design,
at which point a new AC 17 test on the highest sales volume configuration of the new platform
design will be required. The following flowcharts in Figure 5-8 illustrate the process for
determining the testing and reporting requirements for generating A/C efficiency credits.
Figure 5-8 Process for A/C Efficiency Credit Generation: Model Years 2014 through 2025
5-50
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Start
/Menu Credits for 2014X
20167, 2017-20197, or
2020+7
2014-2016-
AC17
Subsequent model year—
/Initial model year?, or a\
subsequent model yeartest
\on unique A/C system? ,--'
js there a unique /
system that differs
substantially from
prior AC17 testing?/
Credits for
platform carry
forward: no new
AC17 test is
required
Credits are
generated for
vehicle according
to menu value *
scaling factor
Credits are
generated for
platform according
to menu, and
AC17 results and
system description
are reported to
EPA
Credits are
generated for
vehicle according
to menu values,
and carry forward
until A/C system is
redesigned
End
5-51
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
/*\ /
^/ \ /
\, Credits for
^\ platform carry
?mq Initial mnriPlJ^, /Isthere a unique A/C\ forward to 2019:
-2019 Initial model yearV / system that differs \ ., no new AC17test
subsequent mode year \— Subsequent model year-**/ substantially from prior > No * is required unless
on unique A/C system?/ \ AC1 7 testing? / platform is
,/
I ^
Initial model year
/ redesigned
| Yes
Run AC17test (or
multiple tests if
COV is >= x%) on
highest sales
volume model for
platform
generating credits
i
,
1
r
Credits are
generated for
vehicle according
to menu values
-^_x
r
Run AC17test (or
multiple tests if
COV is >= x%) on
unique A/C system
on the highest
sales volume
model for platform
generating credits
i
^ ^J
r
Credits are
generated for
platform according
to menu, and
AC17 results and
system description
are reported to
EF
1
'A
r
End
5-52
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
//^\
-2025 initial model y
subsequent model }
on unique A/C syste
'•]•'
Initial model year
I
Run AC17test(or
multiple tests if
COV is >= x%) on
highest sales
volume model for
platform
generating credits
i
Credi
genera
platform E
to mer
AC17re
system d
are rep
EF
r
s are
ted for
according
u, and
suits and
Bscription
orted to
3A
/ \ Cred
/ \ platforr
\_ /there a unique A>C forward
eary. / system that differs \ .. no new f
l^ \ substantially from / > is require
-J^ \riorAC1 7 testing/ platfc
/ / redes
\ /
Yes
1
Run AC17test(or
multiple tests if
COV is >= x%) on
highest sales
volume model for
platform
generating credits
/ Compare /
/ AC1 7 result /
/ from new /
/ vehicle to /
/ baseline /
/ vehicle /
,/\
/ \
/ \
/ \
/ \
/ \
/ \
Aslhe difference betweeYk
/baseline and new AC17 test\
X^than the sum of applicable /
\ menu credits? /
No
f
7\
Prepare and submit and / \
engineering analysis /^an the higher-than\
which explains the new /expected AC17 result on \
AC17 result and ^_Yes — < new vehicle be justified >
justifies the full menu \ with an engineering /
credit amount \ analysis? /
^N /
\ /
I Proportional
No credits are
^ Generated for
/ Calculate credit / vehicle according
/ scaling factor / to menu value *
/ applicable menu / AC1 7 results and
/ credits / system description
EPA
i
tsfor
n carry
to 2025:
C1 7 test
d unless
rm is
gned
r
End
5-53
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
In addition, EPA is requiring that manufacturers provide detailed information about
the A/C systems in vehicles tested, both baseline and new, as well as a plot with the interior
temperature of both vehicles, to confirm that there is equivalent or better cooling system
performance in the new vehicle configuration. EPA requires that interior temperature be
measured at three locations: at the outlet of the center duct on the dash panel, and behind the
driver and passenger seat headrests. For the headrest locations, the temperature measuring
devices shall be nominally 30 millimeters behind the center of the headrest.
5.1.3.9 A/C Efficiency Credits and Quantification of Credits
EPA and NHTSA believe that it is possible to identify the A/C efficiency-improving
components and control strategies most likely to be utilized by manufacturers and will
continue to assign a CC>2 credit and fuel consumption improvement value to each. In addition,
the agencies recognize that to achieve the maximum efficiency benefit, some components can
be used in conjunction with other components or control strategies. Therefore, the system
efficiency synergies resulting from the grouping of three or more individual components are
additive, and will qualify for a credit commensurate with their overall effect on A/C
efficiency. A list of these technologies - and the amount of credit (and estimated fuel
consumption improvement value derived from the credit) associated with each technology - is
shown in Table 5-12. If more than one technology is utilized by a manufacturer for a given
vehicle model, the A/C credits or fuel consumption improvement values can be added, but the
maximum credit and fuel consumption improvement value possible is limited to 5.0 g/mi for
cars (equivalent to 0.000563 gal/mi) and 7.2 g/mi (equivalent to 0.000810 gal/mi) for trucks.
In the proposal, NHTSA sought comment on setting fuel specific conversion factors.
The agencies did not receive any comments on the use of fuel specific conversion factors.
The agencies believe that since both the CAFE target curves and the AC credits are derived
using the gasoline conversion factor, it is appropriate to use the gasoline conversion factor for
all fuels. If different conversion factors were used based on the type of fuel, there would be
misalignment between the A/C compliance credits, which would be based on the type of fuel
the vehicle uses, and the stringency of the target curves, which are based on gasoline.
Therefore, the agencies are finalizing the use of the gasoline conversion factor to determine
A/C improvement values for all vehicles.
5-54
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Table 5-12 Efficiency-Improving A/C Technologies and Credits
Technology Description
Reduced reheat, with externally-
controlled, variable-displacement
compressor
Reduced reheat, with externally-
controlled, fixed-displacement or
pneumatic variable displacement
compressor
Default to recirculated air with closed-
loop control of the air supply (sensor
feedback to control interior air quality)
whenever the outside ambient
temperature is 75 °F or higher (although
deviations from this temperature are
allowed if accompanied by an
engineering analysis)
Default to recirculated air with open-loop
control of the air supply (no sensor
feedback) whenever the outside ambient
temperature is 75 °F or higher (although
deviations from this temperature are
allowed if accompanied by an
engineering analysis)
Blower motor control which limit wasted
electrical energy (e.g. pulsewidth
modulated power controller)
Internal heat exchanger (or suction line
heat exchanger)
Improved evaporators and condensers
(with engineering analysis on each
component indicating a COP
improvement greater than 10%, when
compared to previous design)
Oil Separator (internal or external to
compressor)
A/CCO2
Emission
and Fuel
Consumption
Reduction
30%
20%
30%
20%
15%
20%
20%
10%
Car A/C
Credit and
Adjustment
(g/mi C02
and gal/mi)
1. 5 (30% of
5. 0 g/mi
impact) 1
0.000169
l.O/
0.000113
1.5/
0.000169
l.O/
0.000113
0.8 /
0.000090
l.O/
0.000113
l.O/
0.000113
0.5 /
0.000056
Truck A/C
Credit and
Improvement
(g/mi C02
and gal/mi)*
2.2 (30% of
7.2 g/mi
impact) 1
0.000248
1.4 /
0.000158
2.2 /
0.000248
1.4 /
0.000158
1.1 /
0.000124
1.4 /
0.000158
1.4 /
0.000158
0.7 /
0.000079
* This factor is a gasoline conversion from CO2 using 8887 g/CO2 per mpg, NHTSA
will set this constant independent of fuel.
5-55
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Even though EPA is finalizing a design-based A/C credit program that introduces
some minor revisions to the menu values from the MYs 2012-2016 rule, EPA continues to
believe that a full performance-based test procedure is the most appropriate way for
quantifying A/C credits. Performance-based procedures place no limits on the technological
choices made by a manufacturer to improve efficiency. Ideally, performance based standards
would be the most appropriate method of quantifying A/C credits, however there are many
challenges to accurately quantifying a small incremental decrease in emissions and fuel
consumption compared to a relatively large tailpipe emissions and fuel consumption rate. For
example, it would be nearly impossible to distinguish and measure the impact of a 0.5g/mi
improvement in tailpipe emissions due to an improved oil separator system incremental to a
tailpipe 250g/mi test procedure result. The 0.5 g/mi increment would be well within the noise
of a test measurement or test-to-test variability. Even if a number of the technologies were to
be packaged together to account for a 5.0g/mi improvement, this is still only 2% of the
tailpipe emissions value and still may be within test-to-test variability.
The other major challenge to quantifying credits is that it is not practical (from a
compliance standpoint) to measure the CO2 emissions from a vehicle with and without a
series of technologies that include hardware and software integrated in a complex fashion.
This could only be done with an "A" to "B" comparison where the "B" condition includes the
technologies and the "A" condition does not. Such A to B test comparisons require the
manufacture of a prototype vehicle that is in all respects identical to the certified vehicle with
the exception that the technologies being evaluated are removed. This would be impossible to
do for every vehicle certified for a fuel economy test. It would even be prohibitive for a
single vehicle demonstration for each manufacturer. This might only be practical on a single
vehicle research level program as was done in the EVIAC study. The comparison of the AC 17
test result to AC 17 result for the baseline vehicle with the older technology will likely show a
small change in emissions, based on the vehicle simulation results presented above. A more
direct comparison of individual technologies is likely to give even more accurate
quantification of credits such that the menu may no longer be required.
The EVIAC study successfully demonstrated that there are methods by which the
efficacy of technologies can be measured. In the EVIAC study, the efficiency of A/C
components were measured on a test bench where the conditions can be precisely controlled.
Test bench measurements are, by their nature, much more repeatable than chassis
dynamometer tests. They can also easily be used to do A to B comparisons of technology
effectiveness since components can be relatively easily swapped out. The limitations of test
bench measurements primarily lie in the fact that they cannot capture the impact of the
component integration into the vehicle. The test bench only measures the efficiency of the
A/C components, it cannot account for the controls strategy (for example), such as forced
recirculation, not defaulting to reheat, and smart cycling of fixed displacement compressors.
Another disadvantage of test benches are that there are few such facilities available in the
United States and typical OEMs do not possess such extensive test benches as they do not
manufacture A/C components.
One option to circumvent the limitations of both the test bench and the chassis tests
are to merge the two in a combined test procedure that will utilize the strengths of each to
supplement the weaknesses of the other. The test bench can generate the A to B comparison
5-56
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
portion of the credit on the hardware changes, while the chassis test generates the A to B
comparison of the (software) controls strategy changes.
An A/C test bench typically measures the efficiency of a system by measuring its
Coefficient of Performance (COP). The COP of a heat pump is the ratio of the change in heat
at the output to the supplied work (also equivalent to the SEER seasonal energy efficiency
ratio rating on a residential A/C unit).p The IMAC procedure employed the SAE procedure
J2765 in order to bench test systems in a fashion that reflects national average A/C usage.
This test procedure could be used to generate the efficiency of any production A/C system.
The challenge lies in the comparison to the baseline "A" system for the A to B comparison.
This could be done either with a defined hardware baseline system or a typical baseline COP
value agreed upon by the industry.
Combining the bench test together with a chassis test requires a model, simulation, or
some calculation procedure (algorithm) to convert the test bench results to fuel economy and
GHG emissions. There are a number of options for this model. The Lifecycle Climate
Performance or LCCP model (also known as SAE J2766), developed by General Motors in
partnership with SAE, NREL, EPA, is one such model, and was utilized for the IMAC
project. While the LCCP model took into account many factors concerning lifecycle
emissions and fuel use (including the energy needed to manufacture a particular refrigerant),
it may be possible to employ a portion that model, and only discern the effect of the A/C
system efficiency of annualized fuel use due to A/C operation. As updates to the LCCP
model occur, EPA will evaluate the appropriateness of using such tools to quantify the effect
of efficiency-improving A/C technologies. Another option is for the test bench to produce
charts like the one in Figure 5-4. This can then be used as an input into EPA's vehicle
simulation tool. Whatever the method, such a series of models can convert a system COP into
a change in fuel economy and CO2 emissions from the hardware changes in an A/C system.
The controls strategy changes in the menu will have to be measured with an A to B
comparison on the chassis dynamometer test procedure described above. To do this, the
manufacturer would test a vehicle with a baseline controls strategy compared with a modified
more efficient strategy. Though EPA has not yet conducted a test program to test the
feasibility of this concept, combining the results from the bench and dynamometer tests
should give a quantitative assessment of the credits from an improved A/C system compared
to a baseline system. Such an approach could be used in a manufacturer's engineering
analysis submission, to demonstrate effectiveness of a technology (or technologies) in the
absence of a of a baseline vehicle test.
The Alliance and others commented that the bench testing methodology is too
complex and costly (and thus impractical) to employ as a compliance mechanism. The
agencies tend to agree with this assessment. Due to the relative complexity (and expense) of a
p According to the second law of thermodynamics, the COP of a real heat pump system is limited to the Carnot
cycle efficiency, which is the ratio of the low Temperature to the difference between the high and low
temperatures (in Kelvin).
5-57
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
bench test and model demonstration, it would be practical for a manufacturer to do this testing
for only a small number of vehicle and A/C configurations in any given year. The EPA has
met with a few manufacturers and also received comments regarding test vehicle selection,
and they have informed EPA that on any single vehicle platform, the A/C systems usually
share similar configurations. Most full line manufacturers only have a handful of vehicle
platforms (in order to save on engineering and manufacturing costs). Therefore, this
compliance demonstration and engineering analysis should only have to be conducted
infrequently on a brand-new vehicle platform or A/C system design. AS described above,
based on the limited number of platforms and the relative infrequency of platform redesigns,
EPA expects that any manufacturer may ultimately only be required to do a compliance
demonstration and engineering analysis of A/C credits for a given platform perhaps one or
two times, depending on the number of unique A/C system designs used on the various
models within that platform in order to generate credits.
One clarification that is being added to this final rule is to note that air conditioner
efficiency is an "off-cycle" technology. It is thus appropriate for a manufacturer to employ
the standard off-cycle credit approval process described in Section II.F and III,C of the
preamble, as well as in the MYs 2012-2016 rule if the manufacturer believes it can
demonstrate that a greater amount of credit is justified. Utilization of bench tests in
combination with dynamometer tests and simulations (similar to the SAE EVIAC study) would
be an appropriate alternate method of demonstrating and quantifying technology credits (up to
the maximum level of credits allowed for A/C efficiency). A manufacturer can choose this
method even for technologies that are not currently included in the menu starting as early as
model year 2012 (2017 for CAFE).
5.1.4 Air Conditioner System Costs
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 CO2 emissions and fuel
economy as a result of A/C use. The GHG and fuel economy effectiveness is unchanged
from estimates used for 2016 model year vehicles in the 2012-2016 final rule.
In the MYs 2012-2016 rule, EPA estimated the direct manufacturing costs (DMC) of
direct/leakage reduction A/C controls at $17 (2007$) and of indirect/efficiency improvement
controls at $53 (2007$). These DMCs become $18 (2010$) and $55 (2010$), respectively, for
this analysis. EPA continues to consider those DMCs to be applicable in the 2012MY and
continues to consider the technologies to be on the flat portion of the learning curve. For this
rule, the 2012-2016 rule technologies represent the reference case in terms of controls and
costs. We have applied to those DMCs low complexity ICMs of 1.24 through 2018 then 1.19
thereafter to generate the indirect costs for this reference case. The resultant reference case
costs are shown in Table 5-15.
New for this rule, and consistent with the proposal, are additional costs for
indirect/efficiency improvement control as those 2012-2016 MY vintage systems penetrate to
the entire fleet. In addition, as new costs are assumed which are associated with the
alternative refrigerant—both the alternative refrigerant itself and the system changes to
5-58
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
accommodate that refrigerant. For the first of these—indirect controls—the agencies have
estimated the DMC at $15 (2010$) applicable in the 2017MY. The agencies consider this
technology to be on the flat portion of the learning curve and have used a low complexity
ICM of 1.24 through 2018 then 1.19 thereafter. For the alternative refrigerant, the agencies
have estimated a DMC of $67 (2010$) applicable in the 2016MY. The agencies consider this
technology to be on the steep portion of the learning curve because it is only now starting to
be used in a limited number of vehicles. For this technology, the agencies have used a low
complexity ICM of 1.24 through 2022 then 1.19 thereafter. For the alternative refrigerant
system costs (i.e., the hardware changes necessary to accommodate the alternative
refrigerant), the agencies have estimated a DMC of $15 applicable in the 2016MY The
agencies consider this technology to be on the flat portion of the learning curve and have used
a low complexity ICM of 1.24 through 2018 then 1.19 thereafter. The resultant control case
costs are shown in Table 5-16.
Note that these costs are expected to be incurred consistent with our estimated ramp
up of manufacturer use of A/C credits. For example, the direct credit for low GWP
refrigerant use is 13.8 g/mi in MYs 2017-2025, but we estimate that the average credit earned
by manufacturers would be 5.5 g/mi on cars in MY 2018 and 5.8 g/mi on trucks in that MY.
Table 5-13 shows the credits by MY as we estimate they will be used for both cars and truck.
Table 5-14 then shows how we have used these estimated credits to scale A/C-related costs by
MY for both cars and trucks. The percentages shown in Table 5-14 are included in the costs
shown in Table 5-15 and Table 5-16.
The total A/C related costs are shown in Table 5-17.
Table 5-13 Projected Average Estimated Use of A/C Credits in g/mi
Car (g/mi)
Truck (g/mi)
Direct
(Leakage)
Credit if
All R-134a
AC
Direct
Credit for
Low GWP
AC
Indirect
Credit
Total Credit
Direct
(Leakage)
Credit if
All R-134a
AC
Direct
Credit for
Low GWP
AC
Indirect
Credit
2016
5.4
0.0
4.8
10.2
6.6
0.0
4.8
2017
6.3
2.8
5.0
12.8
7.0
0.0
5.0
2018
6.3
5.5
5.0
14.3
7.8
5.8
6.5
2019
6.3
8.3
5.0
15.8
7.8
10.3
7.2
2020
6.3
11.0
5.0
17.3
7.8
13.8
7.2
2021
6.3
13.8
5.0
18.8
7.8
17.2
7.2
2022
6.3
13.8
5.0
18.8
7.8
17.2
7.2
2023
6.3
13.8
5.0
18.8
7.8
17.2
7.2
2024
6.3
13.8
5.0
18.8
7.8
17.2
7.2
2025
6.3
13.8
5.0
18.8
7.8
17.2
7.2
5-59
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Total Credit | 11.5 | 12.0 | 17.5 | 20.6 | 22.5 | 24.4 | 24.4 | 24.4 | 24.4 | 24.4
Table 5-14 Scaling of A/C Costs to Estimated Use of Credits
2012-2016
Rule (reference
case)
C
A
R
T
R
U
C
K
Leakage
Reducti
on
Low
GWP
Refriger
ant&
Hard war
e
Efficien
cy
Improve
ments
Leakage
Reducti
on
Low
GWP
Refriger
ant
Efficien
cy
Improve
ments
2017-2025
Rule(control
case)
C
A
R
T
R
U
C
K
Leakage
Reducti
on
Low
GWP
Refriger
ant&
Hard war
e
Efficien
cy
Improve
ments
Leakage
Reducti
on
Low
GWP
Refriger
ant
Efficien
cy
Improve
ments
2016
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
2017
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
1-85%
=15%
2.8/13.8
=20%
1-97%
=3%
1-85%
=15%
0.0/17.2
=0%
1-47%
=53%
2018
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
1-85%
=15%
5.5/13.8
=40%
1-97%
=3%
1-85%
=15%
5.8/17.2
=34%
1-47%
=53%
2019
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
1-85%
=15%
8.3/13.8
=60%
1-97%
=3%
1-85%
=15%
10.3/17.
2 =60%
1-47%
=53%
2020
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
1-85%
=15%
11.0/13.
8 =80%
1-97%
=3%
1-85%
=15%
13.8/17.
2 =80%
1-47%
=53%
2021
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
1-85%
=15%
13.8/13.
8 =100%
1-97%
=3%
1-85%
=15%
17.2/17.
2 =100%
1-47%
=53%
2022
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
1-85%
=15%
13.8/13.
8 =100%
1-97%
=3%
1-85%
=15%
17.2/17.
2 =100%
1-47%
=53%
2023
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
1-85%
=15%
13.8/13.
8 =100%
1-97%
=3%
1-85%
=15%
17.2/17.
2 =100%
1-47%
=53%
2024
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
1-85%
=15%
13.8/13.
8 =100%
1-97%
=3%
1-85%
=15%
17.2/17.
2 =100%
1-47%
=53%
2025
5.4/6.3
=85%
0 0/13 8
=0%
4.8/5.0
=97%
6.6/7.8
=85%
00/172
=0%
4.8/7.2
=47%
1-85%
=15%
13.8/13.
8 =100%
1-97%
=3%
1-85%
=15%
17.2/17.
2 =100%
1-47%
=53%
5-60
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Table 5-15 Costs of A/C Controls in the Reference Case (2012-2016 Final Rule) (2010$)
Car/
Truck
Car
Truck
Cost
type
DMC
DMC
1C
1C
TC
TC
DMC
DMC
1C
1C
TC
TC
A/C Technology
Leakage reduction
Efficiency
improvement
Leakage reduction
Efficiency
improvement
Leakage reduction
Efficiency
improvement
Leakage reduction
Efficiency
improvement
Leakage reduction
Efficiency
improvement
Leakage reduction
Efficiency
improvement
2017
$13
$46
$4
$13
$17
$59
$13
$32
$4
$9
$17
$41
2018
$13
$45
$4
$13
$17
$58
$13
$31
$4
$9
$17
$40
2019
$13
$44
$3
$10
$16
$54
$13
$31
$3
$7
$16
$38
2020
$13
$43
$3
$10
$16
$53
$13
$30
$3
$7
$16
$37
2021
$12
$42
$3
$10
$15
$52
$12
$29
$3
$7
$15
$36
2022
$12
$41
$3
$10
$15
$52
$12
$29
$3
$7
$15
$36
2023
$12
$41
$3
$10
$15
$51
$12
$28
$3
$7
$15
$35
2024
$12
$40
$3
$10
$15
$50
$12
$28
$3
$7
$15
$35
2025
$11
$39
$3
$10
$14
$49
$11
$27
$3
$7
$14
$34
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 5-16 Costs of A/C Controls in the Control Case (2017-2025 Rule) (2010$)
Car/
Truck
Car
Truck
Cost
type
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
DMC
DMC
A/C Technology
Leakage reduction
LowGWP
refrigerant
LowGWP
refrigerant
hardware
Efficiency
improvement
Leakage reduction
LowGWP
refrigerant
LowGWP
refrigerant
hardware
Efficiency
improvement
Leakage reduction
LowGWP
refrigerant
LowGWP
refrigerant
hardware
Efficiency
improvement
Leakage reduction
LowGWP
refrigerant
2017
$2
$13
$3
$1
$1
$3
$1
$0
$3
$17
$4
$2
$1
$0
2018
$2
$22
$6
$1
$1
$6
$1
$0
$3
$28
$7
$2
$2
$18
2019
$2
$32
$9
$1
$1
$10
$2
$0
$3
$42
$10
$2
$2
$32
2020
$2
$34
$11
$1
$1
$13
$2
$0
$3
$47
$14
$2
$2
$34
2021
$2
$42
$14
$1
$1
$16
$3
$0
$3
$58
$17
$2
$2
$42
2022
$2
$41
$13
$1
$1
$16
$3
$0
$3
$57
$16
$2
$2
$41
2023
$2
$39
$13
$1
$1
$13
$3
$0
$3
$52
$16
$2
$2
$39
2024
$2
$38
$13
$1
$1
$13
$3
$0
$3
$51
$16
$2
$2
$38
2025
$2
$37
$13
$1
$1
$13
$3
$0
$3
$50
$16
$2
$2
$37
5-61
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
LowGWP
refrigerant
hardware
Efficiency
improvement
Leakage reduction
LowGWP
refrigerant
LowGWP
refrigerant
hardware
Efficiency
improvement
Leakage reduction
LowGWP
refrigerant
LowGWP
refrigerant
hardware
Efficiency
improvement
$0
$1
$0
$0
$0
$0
$1
$0
$0
$1
$5
$11
$1
$5
$1
$3
$3
$24
$6
$14
$9
$15
$1
$10
$2
$3
$3
$42
$10
$18
$11
$15
$1
$13
$2
$3
$3
$47
$14
$18
$14
$14
$1
$16
$3
$3
$3
$58
$17
$18
$13
$14
$1
$16
$3
$3
$3
$57
$16
$17
$13
$14
$1
$13
$3
$3
$3
$52
$16
$17
$13
$13
$1
$13
$3
$3
$3
$51
$16
$17
$13
$13
$1
$13
$3
$3
$3
$50
$16
$17
DMC=Direct manufacturing cost; IC=Indirect cost; TC=Total cost
Table 5-17 Total Costs for A/C Control Used in This Final Rule (2010$)
Car/
Track
Car
Track
Fleet
Cost type
TC
TC
TC
TC
TC
TC
TC
Case
Reference
Control
Both
Reference
Control
Both
Both
2017
$76
$25
$101
$58
$2
$60
$86
2018
$75
$40
$115
$57
$46
$103
$111
2019
$70
$57
$127
$54
$73
$127
$127
2020
$69
$65
$134
$53
$82
$134
$134
2021
$68
$79
$147
$52
$95
$147
$147
2022
$67
$77
$144
$51
$93
$144
$144
2023
$66
$72
$138
$50
$88
$138
$138
2024
$65
$71
$135
$49
$86
$135
$135
2025
$64
$69
$133
$49
$84
$133
$133
TC=Total cost
The agencies received no public comments on A/C costs, though the EPA did have
confidential meetings with alternative refrigerant suppliers. Due to the confidential nature of
the information shared, the costs and supply discussions from these meetings are not relied
upon to determine the costs in the tables above.
5.2 Off-Cycle Technologies and Credits
EPA employs a five-cycle test methodology to evaluate fuel economy for fuel
economy labeling purposes. For GHG and CAFE compliance, EPA uses the established two-
cycle (city, highway or correspondingly FTP, HFET) test methodology. EPA recognizes that
there are technologies that provide real-world GHG benefits to consumers, but that the benefit
of some of these technologies is not represented on the two-cycle test. Therefore, EPA is
continuing the off-cycle credit program from the MYs 2012-2016 rule with some changes
such as providing manufacturers with a list of pre-approved technologies for which EPA can
quantify a default value that would apply unless the manufacturer demonstrates to EPA that a
different value for its technology is appropriate. This list is similar to the menu driven
5-62
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
approach described in the previous section on A/C efficiency credits. In meetings with
vehicle manufacturers prior to the proposal, the EPA received comments that the regulatory
process for generating off-cycle credits was too cumbersome to utilize frequently if at all, and
that the burden of proof to demonstrate a small incremental improvement on top of a large
tailpipe measurement was impractical. This is similar to the argument described above for
quantifying air conditioner improvements. These same manufacturers believed that such a
process could stifle innovation and fuel efficient technologies from penetrating into the
vehicle fleet. EPA generally agrees with these comments, and proposed and is finalizing a
menu with a number of technologies that the agency believes will show real-world CC>2 and
fuel economy benefits which can be reasonably quantified a priori. The estimates of these
credits were largely determined from research, analysis and simulations, rather from full
vehicle testing, which would have been cost and time prohibitive. However, actual vehicle
testing was used to either support or refine the credit estimates in cases where it was available.
In the proposal, the agencies requested comment on all aspects of the off-cycle credit
menu technologies and derivations. EPA and NHTSA received a number of comments and,
in addition, several stakeholders requested meetings and met with the agencies including
Denso, Enhanced Protective Glass Automotive Association (EPGAA), ICCT and Honda.
Overall, there was general support for the menu based approach and the technologies
included in the proposed list, but there were also suggestions to re-evaluate the definition of
some of the technologies included in the menu, the calculation and/or test methods for
determining the credits values, and recommendations to periodically re-evaluate the menu as
technologies emerge or become pervasive.
In the proposed off-cycle credit menu, credit values were fixed for most of the
technologies while others values were based on a step-function (e.g., x amount of credit for y
amount of reduction or savings) on the off-cycle credit list. In response to the proposal, the
agencies received comments requesting the use of a scalable credit value approach rather than
a fixed values or step-function derived values for high efficiency exterior lighting, waste heat
recovery (formerly termed "engine heat recovery"), solar panels (formerly termed "solar roof
panels"), and active aerodynamics. After much evaluation of and in response to these
comments, the agencies have revised the credit determination approaches for these
technologies by allowing scalable off-cycle credit values derived from a specific technologies
implementation affecting their relative reductions or savings. However, a by-product of
moving to this calculation strategy is the deviation, in some cases, from the proposed
methodology of subtracting a technology's 2-cycle test procedure benefits from the benefits
determined on the 5-cycle test procedure as the agencies, in their evaluation, determine this
approach was not easily or accurately to scalable. As a proxy, EPA employed a vehicle
simulation tool, and applied varying load reduction values to determine benefits shown during
5-cycle testing, where applicable, to develop tables and/or curves to provide sound data
properly scale credit values. This revised calculation approach is discussed in greater detail
for each applicable technology in the following sections.
Another complication that arises from scaling, is that extremely small credit values
can now be quantified. Although we are allowing scaling of the credits, we cannot accept a
request or grant credit for any level of credit less than 0.05 g/mi CO2. As proposed and
5-63
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
finalized, the agencies will be reporting CO2 values rounded to the nearest tenth of
gram/mile, as a result, any reported values below 0.05 g/mi of CO2 would be rounded down
to zero. Therefore, only credit values equal to 0.05 g/mi or greater will be accepted and
approved for any credit requested as part of the off-cycle credit program (e.g., scalable or
fixed; via the pre-defined technology list or alternate method approval process)..
Some commenters suggested that technologies should be added to the list such as high
efficiency alternators (Alliance, Denso, VW, Porsche, Ford), electric cooling fans (Bosch),
HVAC eco-modes, transmission cooler bypass valves (Ford), navigation systems (Garmin),
engine block heaters (Honda) and an "integral" approach utilizing a combination of
technologies (Global Automakers). Daimler commented that the agencies should provide
"congestion mitigation credits based on crash avoidance technologies."
Conversely, some commenters were opposed to adding any technologies to the menu
(CBD) and others suggested some of the proposed values should be re-evaluated (ICCT) or
that the values should be based on real test data, not simulation modeling (NRDC).
In most cases, there was either insufficient supporting data, dependence on unique,
manufacturer-specific designs or implementation, or dependence on driver interaction and
usage that led to our decision not to include these technologies within the menu of off-cycle
technologies. These comments are discussed in more detail in the Preamble Section II.F.
Finally, the agencies carefully assessed all of the comments and conducted additional
analysis in response to the comments, as well as to support the agencies' ongoing work. The
resulting adjustments to off-cycle credit menu values that are being finalized are detailed in
the following sections.
In addition to comments about the individual technologies, the agencies received a
number of comments on the proposed minimum penetration thresholds, the proposed cap on
the amount of menu based off-cycle credit that can be applied, and suggestions to allow the
proposed menu and credit values to be applied to MY2012-2016 vehicles. These comments,
and EPA's response to them, are discussed in preamble sections II.F. and III.C.5.
5.2.1 Reducing or Offsetting Electrical Loads
The EPA test cycles do not require that all electrical components be turned on during
testing. Headlights, for example, are always turned off during testing; including daytime
running lights (DRLs). Turning the headlights on during normal driving will add an
additional load on the vehicle's electrical system and will affect fuel economy. More efficient
lighting, electrical systems or technologies that offset electrical loads will have a real world
impact on fuel economy but are not captured in the EPA test cycles. Therefore, the EPA
believes that technologies that reduce or offset electrical loads related to the operation of the
vehicle or safety deserve consideration for off cycle credits.
To evaluate technologies that reduce or offset electrical loads, the EPA conducted an
analysis of the reduction in emissions corresponding to a general reduction of electrical
demand in a vehicle. Using EPA's vehicle simulation tool described in Chapter 2 of EPA's
5-64
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
RIA, the agency evaluated the change in fuel consumption for a 100W reduction in electrical
load for a typically configured vehicle. The impact of this load reduction was modeled on
both the combined FTP/Highway cycles (2-cycle), and over the 5-cycle drive tests. The
results of this analysis form the basis for a consistent methodology that the EPA applied to
several technologies to determine the appropriate off-cycle credits for those technologies. In
the NPRM, a single conversion factor was proposed to convert Wattage to the CO2 and fuel
consumption benefit. For the final rule, the agencies have determined that this conversion
should depend on the technology. Based on this determination, the solar energy capture and
high efficiency exterior lighting credit are now calculated differently from the waste heat
recovery credit. The method by which each technology converts and uses electrical energy is
discussed below.
5.2.2 Waste Heat Recovery
The combustion process that powers most of today's vehicles results in a significant
amount of exhaust heat. Most of this heat leaves the engine in the form of waste hot exhaust
gasses which are expelled from the vehicle through the exhaust system, or through hot coolant
which that transfers heat from the engine through the radiator to the atmosphere. Recapturing
some portion of this wasted heat energy and using it to offset the electrical requirements of
the vehicle will lead to improved fuel efficiency.
Regardless of the design of the heat recovery system, whether it is exhaust or coolant
based, the EPA assumes that any recovered energy will be in the form of electricity and will
be used to recharge the vehicle's battery (primarily for FtEVs or PFLEVs). This is consistent
with currently proposed waste heat recovery system designs. The GHG and fuel economy
benefit of generating a Watt of energy is estimated though a full vehicle simulation analysis.
For the vehicle simulation, EPA assumed that high-efficiency alternators will be
prevalent in most vehicles within the MY2017-2025 timeframe of this rule; thus, the
simulation includes a high-efficiency alternator. Figure 5-4 below shows a sample efficiency
map of a high-efficiency alternator. Based on this map (used in the proposal for waste heat
recovery), the global average alternator efficiency is 65%. For this final rule, in order to be
consistent with the analysis conducted by Ricardo to inform the efficiency of powertrain and
vehicle technologies,39 EPA used a global efficiency of 70% (from high efficiency alternators)
for use in its modeling calculations as presented below.
5-65
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Alternator Speed RPM
Figure 5-9: Alternator efficiency map (Delco-Remy, 200840)
Table 5-18 below shows the results of the revised simulation using 70% efficiency for
four vehicle classes. Reducing the electrical load on a vehicle by 100W will result in an
average of 2.5 g/mile reduction in CC>2 emissions over the course of a combined
FTP/Highway test cycle, or 3.2 g/mile over a 5-cycle test. A 100W reduction in electrical
load yields a reduction in required engine power of roughly 0.15 kW (=0.1 kW / 65%), or 1-
2% over the FTP/FIWFE test cycles.
Table 5-18: Simulated GHG reduction benefits of 100W reduction in electrical load over FTP/HW and 5-
cycle tests
Driving Cycle
FTP/Highway
5-Cycle
Electrical Load
100W Load Reduction
Base
2-Cycle Difference
100W Load Reduction
Base
5-Cycle Difference
5-Cycle/2-Cycle
Difference
Small Car
[g/mile]
156.8
154.2
2.5
217.8
214.6
3.2
0.7
Mid-
Size Car
[g/mile]
187.7
185.5
2.2
256.9
254.1
2.8
0.6
Large Car
[g/mile]
246.5
244.1
2.4
331
327.9
3.1
0.6
Pick-up
Truck
[g/mile]
416.6
413.9
2.7
544.5
541.1
3.4
0.7
Average*
[g/mile]
2.5
3.2
0.7
* based on a sales average
To determine the off-cycle benefit of certain 100W electrical load reduction
technologies, the benefit of the technology on the FTP/Highway cycles (2-cycle test) is
subtracted from the benefit of the technology on the 5-cycle test. This determines the actual
5-66
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
benefit of the technology not realized in the 2-cycle test methodology and in this case is 3.2
g/mi minus 2.5 g/mi, or 0.7 g/mi.q
We received two comments on this approach. The International Council on Clean
Transportation (ICCT) commented that they felt it is inappropriate to subtract 5-cycle benefits
from 2-cycle benefits for waste heat as well as other technologies (discussed below). Instead,
the ICCT suggested that the 2-cycle percentage benefits should be used for the load reduction
estimate. The Alliance commented in support of the approach of subtracting the 2-cycle
benefits from the 5-cycle benefits. However, they also wanted to have the technologies that
use the load reduction estimate to be scalable rather than as a single value (i.e., 0.7 g/mi CC>2
credit per 100 watts reduced). Other commenters shared the desire for more scalable credits
for this and other technologies. While these comments apply to all of the electrical load
technologies, it is fitting to discuss them here first.
Regarding subtraction of the 2-cycle from the 5-cycle benefits, the ICCT did not feel
the method of subtracting the 2-cycle benefits from the 5-cycle benefits was appropriate.
Instead, they recommended we use the 2-cycle percentage benefits to estimate any electrical
load reduction that does not occur on the test cycles. Supplemental comments from the
Alliance identified the inherent contradiction of ICCT's assertions to have "credits properly
reflect actual in-use reductions" while advocating for the use of 2-cycle testing, which
typically has lower electrical loads than the 5-cycle tests and the real-world. The agencies
agree with the supplemental comments from the Alliance and, therefore, we have decided not
to subtract the 2-cycle benefits from the 5-cycle benefits to develop a base load reduction
estimate and off-cycle credits.
We do agree with ICCT that utilizing the difference between 5-cycle and 2-cycle
benefits for all of the off cycle technologies that affected electrical loads may not be
appropriate. Based on this comment, we are only applying this methodology to waste heat
recovery as this technology will have 2-cycle benefits. Accordingly, the other load reduction
technologies, high efficiency exterior lighting and solar panels should have an alternate
method of calculation (since they will not have 2-cycle benefits) which is described in greater
detail in sections below.
We agree with the comments from the Alliance to allow scaling and are using the base
load reduction estimate directly for calculating the waste heat recovery credit. Accordingly,
we have developed the table and figure below that will allow for appropriate scaling of the
credit based on the load reduced. Therefore, we are finalizing a scalable approach using the
table and figure below based on these comments.
Table 5-19 Estimated electrical load reduction estimate and corresponding credit values.
qHowever, other technologies (for example, lighting and solar panels) providing benefits off-cycle that cannot be
measured on either the 5-cycle test nor the 2-cycle test be used as the credits without subtracting a 2-cycle value.
An example of this is provided later [WHERE?].
5-67
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Electrical
Load [W]
10
50
100
150
CO2 Credit
[g/rnile]
0.1
O.3
O.7
1.0
Off-Cycle Credit for Electrical Load
0.0
20
30
40
50
60 70 BO 90
Electrical Load [W]
100 110
120 130
140 150
Figure 5-10 Graph of estimated electrical load reduction and CO2 credit (g/mile)
We received comments from Honda requesting clarification on whether the waste heat
recovery value is the peak value or the average value over the test cycle. Honda
recommended that it should be based on the average value over a 5-cycle test. We agree that
this requires clarification and are clarifying the waste heat recovery credit is based on the
average value over 5-cycle testing.
Honda also requested the definition for waste heat recovery to be expanded for
conversion to mechanical and thermal energy in addition to "electrical energy." The
conversion of waste heat to thermal energy is already covered elsewhere under the active
engine and active transmission warm-up so we believe the additional inclusion of thermal
energy conversion as part of waste heat recovery is not necessary. The conversion of waste
heat to mechanical energy is more difficult to quantify since we do not have any data of these
5-68
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
systems on current vehicle applications. Therefore we have not included these mechanical
energy conversion systems on the table and we are finalizing the definition for waste heat
recovery specifying the conversion to electrical energy only.
Finally, comments from Borg-Warner and the Motor Equipment Manufacturers
Association (MEMA) mentioned that the term "engine heat recovery" was too narrowly
defined or ambiguous regarding the type of applicable technology. Therefore, they
recommended a more neutral approach and advocated for the term "waste heat recovery"
currently used by the industry, academia, and the Department of Energy. We agree with these
comments and have revised the terminology for this credit in both this section and in the
regulatory text.
The revisions to the terminology and definition, and other clarifications are reflected
in the definitions below under section 5.2.5.
5.2.3 High Efficiency Exterior Lights
The current EPA test procedures are performed with vehicle lights (notably, headlights
including daytime running lamps (DRLs)) turned off. Because of this, improvement to the
efficiency of a vehicle's headlights is not captured in the existing test procedures and is
appropriately addressed through the off-cycle crediting scheme. Further, since a typical level
of improvement can be quantified, it is appropriate to include this technology within the off
cycle credit menu.
Similar to the waste heat recovery, EPA conducted full vehicle simulation to
determine the impact of energy savings from high efficiency lights on fuel economy and CC>2
emissions. The methodology is identical to that described above with the exception that the
2-cycle results were not subtracted from 5-cycle test results (in response to ICCT's
comments). Rather, the energy levels with and without the technology were compared
directly on the 5-cycle simulation only. This results in a CC>2 reduction of 3.2g/mi per 100
Watt saved (or generated in the case of solar panels) as shown in Table 5-18 (in the NPRM,
this value was 3.7 g/mi per 100 Watts).
As with residential light bulbs, the technology available for vehicle lighting has
changed significantly in recent years. Vehicle manufacturers are commonly using advanced
technology LEDs in taillights and offering new light producing technologies for headlights. If
these technologies require less energy to operate, they will improve the overall fuel economy
of the vehicle and will be eligible for an off-cycle credit.
For the proposal, we referenced Schoettle, et al.41, which studied the effects of high-
efficiency LED lighting. In the draft TSD, Table 5-19 provided a summary excerpted from
that study of average lighting power requirements for both baseline and high efficiency lights
for late-model vehicles and Table 5-20 provided usage rates.
We used these two tables to develop a simple activity-weighted average of the
aforementioned categories which yielded an average nighttime power consumption (for the
categories in question) of roughly 180W for a baseline vehicle and 120W for a vehicle with
high efficiency lights as shown in the draft TSD that accompanied the proposal. This
5-69
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
difference of 60W (180W-120w) was discounted to SOW since 50% of all VMT occurs at
night based on MOVES activity data. Using this SOW and the base load reduction values of
3.7 g/mi CO2 benefit per 100 watts on the 5-Cycle test, we proposed a credit value of 1.1 g/mi
(e.g., (30 watts/100 watts) x 3.7 g/mi).
We received several comments suggesting that the value shown in the table for high-
efficiency low beams of 108.0 watts from the Schoettle et al. report (October 2008) was
overstated. Three commenters, the Alliance, Volkswagen and Honda, suggested separate
values for the low beam high efficiency lighting. The Alliance suggested a value of 52.4W
using the baseline wattage of 112.4W for a savings of 60W. Volkswagen suggested a value
of 54W and a baseline of 137W, based on an European application, for a savings of 83 W.
Lastly, Honda suggested a value of 66W also using the baseline wattage of 112.4W as well,
based also on a European application, for a savings of 46.4W. Since this report is slightly
older and these values are from actual vehicle applications, we agree with the commenters
that these suggested values may be more representative of today's vehicles. Therefore, we
used these numbers to revise our high-efficiency exterior light calculations. These high-
efficiency and baseline low beam wattages represent a percentage savings of 53%
(52.4/112.4), 61% (54/137) and 41% (66/112.4) respectively. Out of these, we used the most
conservative saving estimate of 41%, which is much larger than the 4% savings we originally
assumed. Therefore, we used 66W (41% of the 112.4W for baseline low beam lighting) for
the high-efficiency low beam menu value based on the comments and supporting data.
The Alliance also commented that the brake/stop lamps and center high mounted stop
lamp (CHMSL) lighting are enabled during the 2-cycle tests meaning that some of the real-
world benefit would also be seen on the 2-cycle tests. Since stop/brake and CHMSL already
have a very low usage rate, the benefit of high-efficiency lighting on these two lights would
be minimal (and as explained above, would be rounded to zero). Therefore, these two
lighting elements have been eliminated from the list.
Additional comments from the Alliance and Honda recommended that, rather than the
bulk approach that we proposed, we allow scaling of the credit according to the lighting
systems on the vehicles and that manufacturers be allowed to select individual lighting
components from the list to receive credit. To address these comments, a different approach
was required as the method above provides for a more absolute value based on a discrete
number of components and values. In addition, the approach above uses a 2-cycle/5-cycle
test comparison. Manufacturers currently do not operate lighting, with the exception of
stop/brake lights and CHMSL, on the test cycles and, as a result, this approach would have
required them to start enabling lighting during these test cycles.
For the lighting components used only at night, we used a night time VMT discount of
28.2%, to determine the credit for these components based on a more recent review of
MOVES VMT data as shown in Table 5-20 below. The values in the table were determined
by taking the total VMT distributed on a monthly basis and applying the sunrise and sunset
times on the 15* day of each month over the VMT distribution to develop a relationship
between VMT and the time of day on a monthly basis, and then taking the average of the
monthly night time fraction.
5-70
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Table 5-20 MOVES data showing fraction of VMT attributed solely to night time driving
month
1
2
3
4
5
6
7
8
9
10
11
12
All
Day VMT
1,647,881
1,776,185
1,931,025
2,083,232
2,129,737
2,195,311
2,171,185
2,109,682
1,959,926
1,781,577
1,641,451
1,582,657
23,009,849
Night VMT Night Fraction
1,022,867
894,563
739,723
587,516
541,011
475,438
499,563
561,067
710,823
889,172
1,029,297
1,088,092
9,039,131
0.383
0.335
0.277
0.220
0.203
0.178
0.187
0.210
0.266
0.333
0.385
0.407
0.282
For the components used during day and night, this simply becomes 1.0, which implies there
is no discounting based on night time only usage.
Therefore, we used the power demand estimates, with the revision to the low beam
lighting element, along with the VMT fractions and developed individual lighting credits for
each component on the high-efficiency exterior lighting list as shown in Table 5-21 below.
Table 5-21 Individual Credit Values for High Efficiency Exterior Lighting Components
Lighting Component
Low Beam*
High Beam
Parking/Position
Turn Signal, front
Side Marker, Front
Tail
Turn signal, rear
Side Marker, rear
License Plate
Base electrical load redux
Fuel savings per 100W
Total Available Credit
Baseline
112.4
127.8
14.8
53.6
9.6
14.4
53.6
9.6
9.6
100
3.2
1.0
High Eff
66
68.8
3.3
13.8
3.4
2.8
13.8
3.4
1
watts
g/mi
g/mi
Night Use Only %
91%
9%
100%
0%
100%
100%
0%
100%
100%
Day & Night Use %
0%
0%
0%
5%
0%
0%
5%
0%
0%
Nighttime VMT(MOVES Data):
g/mi CO2 Credit
0.38
0.05
0.10
0.06
0.06
0.10
0.06
0.06
0.08
28.2%
Savings %
52%
46%
78%
74%
65%
81%
74%
65%
90%
*Value for high efficiency wattage changed based on comments and supporting data
Using this table, manufacturers may use all of the lighting components on this list and receive
a maximum of 1.0 g/mi credit. Alternatively, as requested by comment, manufacturers may
select individual lighting components from this list to determine the credits for high efficiency
exterior lighting. To receive high efficiency exterior lighting credit using the pre-defined
technology list, manufacturers may only use the lighting elements and the values shown in
this table.
5-71
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
If a manufacturer has lighting elements that result in a higher benefit than shown on
this table, the manufacturer may use these values under the alternate case-by-case approval
process by performing according to the following formula below:
High Efficiency Exterior Lighting Credit =
(Baseline lighting wattage - high efficiency lighting wattage) x usage rate x VMT fraction x 3.2 g/mi CO2
100 watts;
where usage rate is the percentage shown below for the applicable lighting component, the
vehicle miles traveled (VMT) fraction is 28.2%, or 0.282, for lighting components used
during night time only and 100%, or 1.0, for lighting components used during the daytime and
night time. However, if a manufacturer has lighting elements that provide less benefit than
shown on this table, these lighting components are not eligible for the high-efficiency exterior
lighting credit. To implement this limitation, the agencies reserve the right to request a list of
lighting elements from the manufacturer to support the amount of credit requested, regardless
if it is via the pre-defined technology list or as part of the case-by-case approval process.
We received comments from Honda requesting a separate credit for replacing lighting
relays as well.42 Page 5 of that comment indicates that the implementation of all lighting
relays produces a small total savings of 1.9W. Therefore, we do not believe lighting relays
merit a separate credit due to the small amount of credit that would be generated. However,
they can be included in an assessment of high efficiency exterior lighting credit under the
alternate method approval process since this would create values other than that shown above
in Table 5-21.
We also received considerable comment regarding the inclusion of daytime running
lights (DRLs) on the list of high efficiency exterior lighting. The Agencies did not propose to
include DRLs on the high efficiency exterior lighting list, and are adhering to that decision in
the final rule. It is difficult to assign a power demand value to DRL since some
manufacturers may choose to use dedicated DRLs while some manufacturers may choose to
use the low beams as DRLs. In addition, some manufacturers may implement it on all their
vehicles while some will choose to implement DRLs on a portion of vehicles (or not at all).
The other primary reason for rejecting inclusion of DRL on the pre-approved menu
technology list is that DRLs are currently disabled during the 2-cycle testing and,
consequently, there is no basis for comparison to the 5-cycle test or real-world unless test
procedures are modified to require DRL enablement during standardized test procedure. As a
result, it is difficult to pinpoint a single strategy to assign power demand, assess fleet
implementation rates, and account for it on the standardized test cycles to develop a single
credit value. Therefore, we are not including DRLs in the list of lighting elements used to
grant a high-efficiency exterior lighting credit on the technology list menu. However, as
mentioned before, manufacturers may use the alternate case-by-case demonstration methods
finalized in today's action to request off-cycle credit for DRLs.
Finally, LEDs used for decorative or accent lighting are not eligible for off-cycle credits under
either the technology menu or through a case-by-case demonstration. This is because LEDs
are properly classified as optional accessories or "features".
5-72
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
5.2.4 Solar Panels
Manufacturers are beginning to offer the option to put solar cells on the roof of a
vehicle. The solar panel option on the new Toyota Prius is an example. The initial
implementation of this idea has been limited to cabin ambient temperature control (this
technology is covered under thermal/solar load control below), but manufacturers have raised
the possibility of using rooftop solar cells to charge PHEV and EV batteries and provide
energy to operate the vehicle, increasing the vehicle's all-electric range. This electrical
energy cannot be accounted for on the current EPA cycles - either the two cycle test or the
five-cycle test. Only HEVs, PHEVs and EVs are eligible for this credit.
Using engineering judgment, the EPA estimated in the NPRM that vehicles with a
solar roof would be parked in sunlight on average four hours a day, and that the solar panels
would have an output of SOW. The EPA also assumed that the solar cells will produce 50%
of their rated power of SOW (due to the solar angle, parking conditions, weather conditions,
etc.) with a battery efficiency of 80%. A vehicle with this configuration could save up to 80
Wh/day of electrical energy. The EPA sought comments on these assumptions and requested
more data to refine these numbers (See draft joint TSD section 5.2.1.3). EPA also noted the
possibility of scaling the credit for certain solar roof panels. Id.
The ICCT commented that this credit was not appropriate and was not supported by
actual data. However, in contrast, the Alliance comments stated that this level of credit was
appropriate based on theoretical calculations and experimental data.
In addition, the Alliance recommended that 1) the credit should be scalable (e.g.,
(solar roof panel output in watts divided by 50 watts) times 3.0 grams/mile), 2) the credit
should apply for solar panels in locations other than the roof, and 3) the credit should be
available for other vehicles, not just PHEVs and EVs.
The comments from Guardian stated that solar roof panel technology is "rapidly
evolving" to the point where the 50 watt threshold we proposed "will be quickly surpassed or
[is] being surpassed." To address this, they also proposed that we use a simple, formula-
based credit similar to the Alliance comments on scaling. Similar to the Alliance, Guardian
recommended that this credit should not be limited to just "roof panels if other locations can
provide appropriate output and, therefore, the term "roof should be removed to reflect this.
Finally, Guardian suggested that we use standard test conditions (STC) from the photovoltaic
industry to define how panel power is determined of 1000 watts per meter squared (W/m2)
direct solar irradiance and a panel temperature of 25 degrees Celsius.
First, we agree with the Alliance comments that we should scale solar panel credits.
As the agencies stated in the draft joint TSD, "EPA will also consider scaling this credit for
solar roof panels that provide more or less power than SOW." Second, the current definition
for "solar roof panels" does not specify that the panels must be on the roof, although the term
"roof implies this. Therefore, we understand the potential for confusion and will change the
term for this credit and the associated definition to "solar panels". However, since this term
also creates some ambiguity, we will clarify that the term "solar panels" is limited to
"horizontally-oriented, external solar panels with the potential for direct, uninhibited solar
5-73
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
exposure." This prevents someone from installing solar panels in less effective locations
(e.g., underneath the vehicle or in the vehicle passenger cabin) and claiming credit for solar
roof panels. Lastly, the reason we limited this credit to HEVs, PHEVs and EVs is this is
where we see the most benefit for this technology. It would not aid conventional vehicles
since this would amount to "trickle" charging of the battery and, since hybrid vehicles have
many methods and more substantial means of energy recovery, we did not see a need to
expand this credit beyond HEVs, PHEVs and EVs. We do see a benefit for this technology
on conventional vehicles when combined with active cabin ventilation, and these credits are
already included on the menu.
We also agree with the comments from Guardian regarding revising the terminology,
as discussed above, and defining the test conditions to determine the solar panel power output.
We performed a cursory literature search and verified the conditions that Guardian stated,
along with a specification for an air mass of 1.5 (AMI.5). As a result, we are including these
metrics as well as the revised credit terminology in the definitions section. Also, as Guardian
stated, the power output of 50 watts seems to have been surpassed and that solar panel outputs
of up to 150 watts are possible. Therefore, we are revising the solar panel credit formula to
allow for scaling to more efficient and larger panels.
Based on the comments from the Alliance and Guardian, and the agencies' own
suggestion in the draft Joint TSD, we revisited the credits for solar panels to provide for
proper scaling. Similar to high-efficiency exterior lighting discussed above, we needed to use
a different approach since the method used for the proposed solar roof panel credit assigned
was based on the credit scalar according to Table 5-22. This scalar represents an
improvement in CC>2 emissions for every 100 W of electrical load reduced in vehicles
equipped with conventional powertrains. Therefore it is inappropriate to use this scalar to
represent an efficiency improvement from solar panels. For this final rule, we use some of the
base assumptions in the proposal along with new information regarding methods for
quantifying the energy from solar panels to improve the calculation methodology.
To properly scale the credit, we estimated the energy generated by the solar panel and
stored in a P/H/EV battery. Then we used a number of vehicle simulation results from
Ricardo to determine a gram per mile displaced by running a vehicle off of electric power (as
in a battery). First, it is important to define the industry standard for rated solar panels, which
is:
P, = ri * (T)e * A
panel '|pv ^aysw ^
Where: Ppanei is the rated panel power output,
T|pv is the Photovoltaic cell efficiency
OsySW is the standard radiation flux (assumed to be lOOOW/m )r
A is the Solar Panel Cell Area in m2
5-74
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Next, we calculate the amount of energy that is captured by the solar panel on a yearly basis
and stored in the battery of the P/H/EV considering the following factors: the average solar
energy across the United States on a daily basis, the battery/motor efficiency, the amount of
time that the solar panel is exposed to the sun accounting for obstructed parking by structures,
clouds and inclement weather, and the size and efficiency of the solar panel. The formula for
this is as follows:
Epanei = EaVg * Days/Year * r|batt * Expsoiar * r|pv * A
Where:
Eavg is the national average solar energy flux per day (approximately 4.159
kWh/m /day)8, including the effects of weather, season, clouds etc
Days/Year is the number of days per year (365.25 days per year);
r|batt is the average battery/motor combined efficiency (assumed here to be 95%
for battery, 92% for motor, and 98% for power electronic for a total of
86%)*;
ExpSoiar is the amount of solar exposure for the solar panel or "derate" factor,"
assumed to be 79% and includes soiling and shading (e.g., trees,
parking, buildings).
To determine the average solar energy per day of 4.159 kWh/m2/day, we used the historical
data from the National Renewable Energy Laboratory (NREL). Specifically, we used the
State Average Insolation Values (2003-2005) Weighted by Region of Use Based on 2005
Electricity Use Patterns in kWh/m /day. We used all of the states except Alaska and Hawaii
since this would tend to skew the data in a certain direction since these states tend to be at the
extremes of solar exposure. The data is listed for several angles of incidence but we used the
values for a horizontal panel since the solar panels on a vehicle do not automatically move to
acquire an optimal angle for solar exposure. The values we used are shown below in Table
5-22.
s Estimated from http://rredc.nrel.gov/solar/old data/nsrdb/1961-1990/redbook/atlas/
4 Consistent with the assumption in the proposal, we are assuming that the power from the solar cell will be
stored in the battery for the most part. CITE TO PROPOSAL. A small portion can be used directly to power
accessories or even the traction motor during normal vehicle operation (in which case, this factor is not
required), but most vehicles spend most of their time parked. This analysis also assumes that the battery state of
charge is sufficiently low to be able to accept additional energy.
u Estimated from http://rredc.nrel.gov/solar/calculators/PVWATTS/versionl/derate.cgi. In order to determine
Expsolar (derate), we used the following factors: shading = 0.85, soiling = 0.95 (default), system availability =
0.98 (default), and all other factors 1.0, as most of these factors relate to stationary applications. We assumed
the suntracking factor of 1.0 since it would be already included in EavgFor the shading factor, we assume that any
vehicle purchaser who is willing to pay a significant premium for a solar roof will preferentially park it in an
area of high sunlight exposure.
5-75
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Table 5-22 State Average Insolation Values (2003-2005) Weighted by Region of Use Based on 2005
Electricity Use Patterns in kWh/m2/day for contiguous United States
State
Alabama
Arizona
Arka nsas
Ca 1 ifornia
Colorado
Connecticut
Delaware
District of
Co 1 u m b i a
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Ka nsas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Okla homa
Oregon
Pennsylvania
Rhode Island
South Ca rol ina
South Da kota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Insolation Va 1 ue
(2003-05 for a flat panel)
4.3
5.6
4.3
4.9
4.8
3.7
3.9
3.9
4.7
4.3
4.4
3.9
3.9
4
4.4
4
4.5
3.7
3.9
3.7
3.7
3.7
4.4
4.2
3.9
4.2
5.4
3.7
3.8
5.5
3.8
4.2
3.7
3.8
4.5
3.8
3.7
3.8
4.3
4.1
4.2
4.6
4.6
3.7
4
3.6
3.8
3.8
4.5
For our analysis, we used the average of the values across the contiguous states of 4.159
kWh/m2/day to represent the broad spectrum of solar conditions across the United States.
This is equivalent to a solar panel being exposed to the solar energy from the sun (also
known as solar radiation or flux) of 1000 watts per square meter for 4.159 hours on a daily
basis over the course of a year.
5-76
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Substituting in the values above, we get:
Epanei= 4.159 kWh/m2/day * 365.25 days/year * 0.86 * 0.79 * Tipv * A; or
Epanei = 1030 kWh/m2/Year * TV * A
Next, we convert the energy from the panel to a gram per mile of CC>2 equivalent to
determine the credit. In order to do this, we estimate the amount of tractive energy (energy to
drive the wheels in kWh) and the associated CO2 emissions for different subclasses of parallel
hybrid (P2) vehicles. Comparing the energy generated at the battery with the amount of
energy (and emissions) on the P2 hybrid vehicles, we estimate the amount of energy and
hence emissions that are displaced by the solar panels.
The following tables show the 2-cycle average emissions and tractive energy based on full
vehicle simulations and energy analysis of the hybrid vehicles as modeled by Ricardo (see
Chapter 3 of the joint TSD for a description of the Ricardo work):
Table 5-23 CO2 Emissions from Each Vehicle Type
Driving Cycle
FTP (CO2 g/mi)
Highway (CO2
g/mi)
Combined (CO2
g/mi)
Small-Size Car
125.44
150.75
135.69
Mid-Size Car
137.54
148.96
142.45
Large-Size Car
178.13
192.51
184.33
Pick-up Truck
267.95
306.30
283.95
Table 5-24 Vehicle Travel Distance per Energy applied at Wheel
Traveled Mile per
Energy applied at
Wheel [mi/kWh]
FTP
Highway
Combined
Small-Size Car
5.7461
5.5348
5.6510
Mid-Size Car
4.6861
5.1317
4.8867
Large-Size Car
3.8528
4.0623
3.9470
Pick-up Truck
2.6663
2.6553
2.6613
Table 5-25 CO2 Emissions per Energy applied at Wheel
CO2 per Energy
applied at Wheel
[g/kWh]
FTP
Highway
Combined
Small-Size Car
720.76
834.37
766.78
Mid-Size Car
644.52
764.42
696.12
Large-Size Car
686.29
782.03
727.54
Pick-up Truck
714.45
813.32
755.69
5-77
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
The sales weighted average CO2 emissions is 745.8 g/kWh. This is using the sales and VMT
schedules consistent with the rest of this rule. If we combine the emissions per unit energy
with the annual energy generated from a typical panel along with an assumption of 15,000
miles traveled per year for an average vehiclev, we get:
Amount of CO2 reduced by Panel = Epanei * 745.8 g/kWh * r|batt / 15000
mi/year; or
Amount of CC>2 reduced by Panel = 43.85 g/mi * r|pv * A
By rated panel power relationship (defined above), but rewritten in the form:
Ppanel/(1000W/m2)= TV * A
we get a scalable credit such that:
Solar Panel Credit, Csoiar = 0.04385 g/mi/W * Ppanei
This equation is a function of only the rated power of the panel. These are standard
specifications for solar panels and are provided by the panel manufacturers.
Therefore, for a 100 Watt rated panel, the credit would be 4.4 g/mi. This value is (per
the public comments) now scalable to the solar cell.
As an illustrative example, the 2012 Toyota Prius solar panel is currently used for
active ventilationw (equivalent to 2. Ig/mi credit). However, if the solar panel were to be used
to charge the battery it would get a higher credit: In the Prius, the approximate specifications
for the solar panel are an efficiency rating of 16.5% (0.165) and an area of 0.405 m2 thus
having a rating of 67 Watts.x This would qualify for a credit of 2.9 g/mi CC>2.
This value of solar panel credit is comparable to the 2.1 g/mi credit from the active
ventilation. However, the active ventilation is not required all year round (the remainder of
the power generated being wasted). In an effort to encourage more solar use on P/H/EV
vehicles, and to more accurately characterize the year-round potential benefits for use of the
solar technology on vehicles, the agencies are finalizing a credit scheme that will allow
benefits for active ventilation as well as for electrical generation. This was not included in the
proposal and is new to the final rule as a result of the Agencies' effort to make this credit
scalable.
v Consistent (though simplified) with assumptions made elsewhere in this rule
w http ://global.kyocera. com/reliability/file02. html
x http://techon.nikkeibp.co.jp/english/NEWS_EN/20090519/170318/: Area is 36*.15*.075 = 0.405m2, efficiency
is 16.5%.
5-78
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
There are three scenarios to consider regarding the interaction between solar panels
and active cabin ventilation: 1) using the solar panel solely for the purpose of charging the
battery, 2) using the solar panel to only power the active cabin ventilation system, and 3)
using the solar panel to charge the battery and power the active cabin ventilation. The first
two scenarios are simpler and more straightforward so we will discuss them first together.
The third scenario is more complicated due to power splitting.
If the solar panel is being used solely to charge the battery, the solar panel credit alone
will be granted and the equation above can be used to determine the amount of solar panel
credit. If the solar panel is being used solely to power the active cabin ventilation system,
only the active cabin ventilation credit (see section 5.2.13 below) will be available.
If the solar panel is being used to both power the active cabin ventilation system and
to charge the battery, the manufacturer may get credit for some combination of the solar panel
and active cabin ventilation system credit. However, the manufacturer would be required to
account for the amount of power required for the active cabin ventilation system, then
calculate the applicable solar roof panel power for battery charging and subtracting the power
for active cabin ventilation. Note that using the calculation below, a manufacturer could not
get more credit than accounting for the solar panel and active cabin ventilation credits
separately.
To account for the wattage that would be devoted for the active cabin ventilation, we
use the equation above for the panel power and the wattage needed for fans used for active
cabin ventilation. Based on information from Delphi, the power used to operate the fan motor
used in active cabin ventilation is typically 19 Watts. In order to calculate the amount of the
rated panel power that would generate the 19 W from the ideal solar flux of 1000 W/m , we
first estimated the fraction of the average sunlight in US, which would be used to generate the
required power.
Isun ~~ -tWg / Hhr,day / *PSySW
where:
fsun is a fraction of the average sunlight in US (dimensionless);
ay is an average daytime hours per day (assumed to be 12 hours/day).
After substituting appropriate values in this equation, we get 0.347 for fsun. Using this
value, we get the amount of the rated panel power that would generate the fan power from the
ideal solar flux of 1000 W/m2 is
" solar/vent invent / Isun
Where:
Psoiar/vent is the amount of equivalent rated panel power that would generate power for
the vent Pvent
5-79
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Pvent is the amount of power required to run the low speed ventilation fan in the
dashboard (assuming there is a route for the heat to escape the car). For this credit
calculation, the value of 19 W must be used.
With the value of fsun, Psoiar/vent turns out to be 54.8 W (using the fixed value of 19W
for the fan). Then, the remaining solar panel credit for battery charging is calculated as
follows.
Csolar/vent = 0.04385 g/mi/W * (Ppanel - Psolar/vent / 3)
Where:
Csolar/vent is the solar panel credit available for battery charging after the ventilation fan
power has been dissipated.
Note that Pvent is divided by 3 to account for the assumption that the active ventilation
is used only 4 months a year on average. This amount would be subtracted from the solar
panel credit menu value for full battery charging. Substituting, the partial credit equation
becomes:
Csolar/vent = 0.04385 g/mi/W * (Ppanel - Pvent / (3*0.347))
Due to the inherent uncertainties in some of the assumptions, and for the sake of simplicity,
we are setting the constant in the parentheses (3*0.347) to unity. Therefore the equation
becomes simply:
Csolar/vent '
= 0.04385 g/mi/W * (Ppanei - Pvent)
Using the Prius example to illustrate this, if the Prius used the solar panel to operate
the active cabin ventilation system fan motor and to charge the battery, the Prius could receive
2.1 g/mi for active cabin ventilation and additional 2.1 g/mi solar panel credit attributable to
battery charging. This 4.2 g/mi credit is higher than if the solar panels were used for
electricity generation alone (2.9 g/mi). This credit recognizes that the agencies are now
providing an incentive to use that additional power that might have been wasted in a
beneficial manner, reducing GHG emissions and using less fuel.
In summary, we are finalizing the credits for solar panels using the revised values,
allowing them to be scaled according to solar panel output, and allowing for combining the
credit with the active ventilation credits where the solar panels are used for both purposes.
The agencies are revising the terminology and definition for solar panel credits as discussed
above. In addition, as proposed, the solar panel credit is only available for HEV, PHEV, EVs,
and FCEVs (fuel cell), and is not eligible for incentive multipliers.
5-80
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
5.2.5 Definitions for Electrical Load Offsetting and Reduction Technologies
Waste heat recovery is a system that captures heat that would otherwise be lost
through the engine, exhaust system, radiator or other sources and converting that heat to
electrical energy that is used to meet the electrical requirements of the vehicle or used to
augment other load reduction technologies (e.g., cabin warming, active engine/transmission
warm-up strategies). The amount of energy recovered is based on the average value over 5-
cycle testing.
High efficiency exterior lighting means a lighting technology that, when installed on
the vehicle, is expected to reduce the total electrical demand of the exterior lighting system
when compared to conventional lighting systems. To be eligible for this credit the high
efficiency lighting must be installed on one or more of the following lighting components:
low beam, high beam, parking/position, front and rear turn signals, front and rear side
markers, taillights, backup/reverse lights, and/or license plate lighting.
Solar roof panels means the installation of horizontally-oriented, external solar panels with
direct solar exposure, uninhibited by portions of the or the entire vehicle, on an electric, fuel
cell electric, hybrid electric or a plug-in hybrid electric vehicle such that the solar energy is
used to provide energy to the electric drive system of the vehicle by charging the battery or
directly powering essential vehicle systems (e.g., cabin heating or cooling/ventilation), or
providing power to the electric motor. The rated power of the solar roof panels used to
determine the credit value must be determined under the standard test conditions of 1000
watts per meter squared direct solar irradiance at a panel temperature of 25 degrees Celsius
+/- 2 degrees Celsius with an air mass of 1.5 spectrum (AMI.5).
5.2.6 Active Aerodynamic Improvements
The aerodynamics of a vehicle play an important role in determining fuel economy.
Improving the aerodynamics of a vehicle reduces drag forces that the engine must overcome
to propel the vehicle, resulting in lower fuel consumption. The aerodynamic efficiency of a
vehicle is usually captured in a coast down test that is used to determine the dynamometer
parameters used during both the two-cycle and five-cycle tests. This section discusses active
aerodynamic technologies that are activated only at certain speeds to improve aerodynamic
efficiency while preserving other vehicle attributes or functions. Active aerodynamic features
can change the aerodynamics of the vehicle according to how the vehicle is operating, and the
benefit of these vehicle attributes may not be fully captured during the EPA test cycles.
Two examples of active aerodynamic technologies are active grill shutters and active
ride height control. Active grill shutters close off the area behind the front grill so that air
does not pass into the engine compartment when additional cooling is not required by the
engine. Nearly all vehicles allow air to pass through the front grill of the vehicle to flow over
the radiator and into the engine compartment. This flow of air is important to prevent
overheating of the engine (and for proper functioning of the A/C system), but it creates a
significant drag on the vehicle and is not always necessary. Thus, active grill shutters reduce
the drag of the vehicle, reduce CO2 emissions, and improve fuel economy. When additional
cooling is needed by the engine, the shutters open until the engine is sufficiently cooled.
5-81
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Active ride height control uses the chassis and suspension components, such as hydraulic
shock absorbers, to lower the height of the vehicle, thus reducing ground clearance, typically
at higher vehicle speeds. This lowers the relative amount of air traveling under the vehicle
while maintaining the amount of air around and over the vehicle. This reduces drag on the
vehicle requiring less power to maintain the same speed, and consequently reducing fuel
consumption.
As proposed, EPA is limiting credits to active aerodynamic systems only (not
passive). The reason for this is that passive systems are too difficult to define and isolate as a
technology. For example, the aerodynamic drag on the vehicle is highly dependent on the
vehicle shape, and the vehicle shape is (in turn) highly dependent on the design characteristics
for that brand and model. EPA believes that it would be inappropriate to grant off-cycle
credits for vehicle aesthetic and design qualities that are passive and fundamentally inherent
to the vehicle. Thus, passive aerodynamic systems are not an off-cycle menu technology, and
also could not be a candidate for off-cycle credits under the case-by-case demonstration
procedures.
To evaluate active aerodynamic technologies that reduce aerodynamic drag, the EPA
conducted an analysis of the reduction in emissions corresponding to a general reduction of
aerodynamic drag on a vehicle. Using EPA's full vehicle simulation tool described in EPA's
RIA Chapter 2, the agency evaluated the change in fuel consumption for increasing reductions
in aerodynamic drag for a typically configured vehicle. The results of this analysis form the
basis for a consistent methodology that the EPA applied to technologies that provide active
aerodynamic improvements.
Vehicle aerodynamic properties impact both the combined FTP/Highway and 5-cycle
tests. However, these impacts are larger at higher speeds and have a larger impact on the 5-
cycle tests. By their nature of being "active" technologies, EPA understands that active
aerodynamic technologies will not be in use at all times. While deployment strategies for
different active aerodynamic technologies will undoubtedly vary by individual technology,
the impact of these technologies will mostly be realized at high speeds. Since aerodynamic
loading is highest at higher speeds, EPA expects that active aerodynamic technologies will
generally be in use at high speeds, and that the 5-cycle tests will capture the additional real
world benefits not quantifiable with the FTP/Highway test cycles due to the higher speed in
the US06 cycle. Active aero may also depend on weather conditions. For example, active
aerodynamics may operate less in hot weather when air cooling is required to exchange heat
at the condenser. Also, active grill shutters may need to stay open during snowy conditions in
order to prevent them from freezing shut (potentially causing component failure). In fact, the
MOVES data indicates that only 68% VMT occurs between 40 °F and 80 °F.
Using EPA's full vehicle simulation tools, the impact of reducing aerodynamic drag
was simulated on both the combined FTP/Highway cycles and the 5-cycle drive tests. To
determine the fuel savings per amount of aerodynamic drag reduction, the fuel savings on the
FTP/Highway test cycle was subtracted from the fuel savings on the 5-cycle tests. This is
consistent with the approach taken for other technologies. Then, using the MOVES data, the
vehicle simulation results were adjusted for the temperature effects on active grill shutter
operations. Table 5-26 shows the results of the vehicle simulation. Also, Figure 5-11
5-82
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
represents this GHG reduction metrics in a graphical form. These results assume that the
active aerodynamics affects the coefficient of drag only, which is currently assumed to be
constant over a wide range of vehicle operating speed. However, if the coefficient of
aerodynamic drag is assumed to be vehicle speed dependent, then a different relationship
could result.
Table 5-26 Simulated Maximum GHG Reduction Benefits of Active Aerodynamic Improvements
Reduction in
Aerodynamic Drag (Cd)
1%
2%
3%
4%
5%
10%
Car Reduction in
Emissions (g/mile)
0.2
0.4
0.6
0.8
0.9
1.9
Truck Reduction in
Emissions (g/mile)
0.3
0.6
1.0
1.3
1.6
3.2
I
=
.a
ts
Performance Metrics for Active Aero Technology
7.0
6.0 -
4.0 -
3.0
2.0
i.o H
0.0
O'o
4 Car
• Truck
5% 10% 15% 20%
Aero Dynamic Improvement
25%
Figure 5-11 Simulated GHG Reduction Benefits of Active Aerodynamic Improvements
We are scaling the credit for active aerodynamics using Table 5-26 and Figure
5-11 shown above. A manufacturer would simply determine the aerodynamic benefit of their
active technology on a percent basis and find the corresponding CC>2 value in grams per mile
off using the data points in the table.
5-83
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
5.2.7 Definition for Active Aerodynamic Improvements
Active aerodynamic improvements are technologies that are automatically activated
under certain conditions to improve aerodynamic efficiency (e.g., lowering of the coefficient
of drag or Cd using SAE J2881, while preserving other vehicle attributes or functions).
5.2.8 Advanced Load Reductions
The final category of off-cycle credits includes technologies that reduce engine loads
by using advanced vehicle controls. These technologies range from enabling the vehicle to
turn off the engine at idle, to reducing cabin temperature while a vehicle is parked and thus
reducing A/C loading when the vehicle is restarted. Because the benefit of these technologies
is not fully captured on the combined two cycle tests, and the real-world benefits can be
reliably but conservatively calculated EPA has evaluated each technology and developed
automatic off-cycle credits for each technology individually.
5.2.8.1 Engine Idle Start-Stop
Engine idle start-stop technologies enable a vehicle to turn off the engine when the
vehicle comes to a rest, and then quickly restart the engine when the driver applies pressure to
the accelerator pedal. The benefit of this system is that it largely eliminates fuel consumption
at idle. The EPA FTP (city) test does contain short periods of idle, but not as much idle as is
often encountered in real world driving. HEV and PHEVs can also idle-off and are thus
eligible for this credit. EVs and FCVs do not have engines and thus are not eligible for this
credit.
As stated in the proposed Joint TSD, based on a MOVES estimate that 13.5% of all
driving (in terms of vehicle hours operating) nationwide is at idle, and compared to a 9% idle
rate for the combined (two-cycle) test, idle-off could theoretically approach an extra 50% of
the existing benefit seen on the FTP/HWFE test. Vehicle simulation data was used to
quantify the amount of fuel consumed in idle conditions over the FTP and FIFET test across a
range of vehicle classes. For each vehicle class reviewed, a FTP-HFET combined fuel
consumption was calculated and compared to total fuel consumption during the combined
test. The ratio of idle fuel to total fuel represents a maximum theoretical fuel consumption,
and hence GHG emissions, that could be reduced by eliminating idlingy. Table 5-27 shows
this below:
y Note that aggressive fuel cutoff upon vehicle decelerations are technically possible and could increase the total
amount of avoided "idle" fuel consumption; at the same time, the idle-off enable conditions might reduce the
total idle avoidance. Given the accuracy level of this methodology, EPA assumes these caveats to cancel each
other out.
5-84
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Total FTP fuel consumption (g)
FTP fuel consumed at idle (g)
Total HWFEfuel consumption (g)
HWFE fuel consumed at idle (g)
FTP-HWFE combined fuel consumption (g)
FTP-HWFE combined fuel consumed at idle (g)
potential % GHG reduction benefit
% FTP idle time
% HWFE idle time
FTP-HWFE combined % idle time
Real-world % idle time (via MOVES)
Real-world % GHG reduction benefit
Off-cycle GHG benefit
Assumed GHG for advanced vehicle (g/mi)
Off-cycle GHG benefit
Standard
Car
1044
68
675
0.0
878
37
4.2%
16%
0%
9%
13.5%
6.3%
2.1%
165
3.4
Large
Car
1276
71
862
0.0
1090
39
3.6%
16%
0%
9%
13.5%
5.3%
1.7%
235
4.1
Large
MPV
1412
69
970
0.0
1213
38
3.1%
16%
0%
9%
13.5%
4.6%
1.5%
255
3.9
Full size
Truck
1868
97
1240
0.0
1585
53
3.4%
16%
0%
9%
13.5%
5.0%
1.6%
365
6.0
Table 5-27: Calculations Used for Off-Cycle Credit for Engine Idle Start-Stop Technologies
Based on the data in Table 5-27 above, EPA suggested that engine idle start-stop
technology is theoretically capable of providing 3.8 g/mi credit for passenger vehicles and up
to 6.0 g/mi for trucks. However, cold and hot ambient conditions will prevent idle-off in all
cases. Based on MOVES data of VMT as a function of temperature (see Table 5-28 below),
the percentage of nationwide VMT driven above a 45 °F ambient temperature is
approximately 75%. Therefore, EPA and NHTSA proposed 75% of the theoretical savings
above will be appropriate for an idle off credit; equating to 2.9 g/mi for passenger vehicles
and 4.5 g/mi for trucks.
The comments from ICCT were critical of the underlying assumptions used to
determine the amount of credit in two respects : 1) the idle rate assumed for the 2-cycle tests
and 2) application of the real-world idle percentage to only the off-cycle credit value, not the
underlying idle time used to determine the amount of the credit.
First, the commenter stated that the!6% idle rate for the 2-cycle tests solely
contributed by the FTP and listed in the TSD should actually be 19.5% in total, with 19% of
the idle contributed by the FTP test and 0.5% contributed by the HWY test. Thus, when
applying the FTP/HWY weighting of 55%/45%, this produces a weighted idle rate of 10.7%,
not 9% used in the TSD.
Second, the commenter stated that we only applied the real-world idle percentage to
scale the engine idle start-stop credit to the credit value, not to the underlying idle time used
to determine the credit. This comment has technical merit. The agencies therefore we used
5-85
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
the MOVES model to estimate that 13.5% of VMT was during idling, providing an
opportunity for engine idle start-stop. Further, we assumed the engine would be running 25%
of the time due to cold temperatures, leaving 75% of the real world VMT for stop start. Using
vehicle simulation with the 13.5% real-world, idle-off time assumption, we estimated a
potential benefit of 3.8 g/mi for cars and 6.0 g/mi for trucks and applied the real-world factor
of 75%, resulting in a proposed credit of 2.9 g/mi for cars and 4.5 g/mi for trucks. Thus, the
commenter's assertion was that the 75% real world factor should have been applied to the real
world idle off time of 13.5%, yielding a true real-world idle off time of 10.1% (e.g., 13.5%
times 75%). As a result, according to the commenter, there is no benefit to grant an
applicable credit for engine idle start-stop.
We reviewed these comments thoroughly and agree that some of ICCT's comments
have merit. In particular, the comments regarding the 10.7% idle rate for the 2-cycle test and
the application of real-world factoring were appropriate. However, we disagree with ICCT
regarding the real-world idle time and the lack of benefits for granting engine idle start-stop
credits. Thus, we have revised the approach for determining engine idle start-stop credit as
described below taking into account the issues highlighted by ICCT.
For the 10.7% 2-cycle idle rate, when we consider the amount of time to reach proper
operating engine temperature, a small portion of the FTP was eliminated. Our in-house test
data showed that the average time to reach 90% maximum engine coolant was on average 324
seconds, and due to this, eliminating the first two idle periods of the FTP. As a result, the idle
rate we used for the 2-cycle test was 10.0% instead of the 10.7% suggested by the commenter.
Next we reviewed the estimates of the amount of idle time in the real world. We
reviewed the analysis for the estimate of the real-world percent idle time in MOVES and,
since new information has been added to the MOVES model since the NPRM, this number
has increased slightly to 13.76% from the previous estimate of 13.5%. To validate this
number, we reviewed other data and studies to see how it compares. In the Supplemental FTP
(SFTP) studies conducted in the 1990's, there was a very large vehicle driving activity study
conducted with instrumented vehicles (EPA 420-R-93-007, "Federal Test Procedure Review
Project: Preliminary Technical Report, May 1993). The study revealed that the real-world
percent idle rate (by time) was 22% and it is not certain how much driving patterns have
changed since then. Therefore, we looked at more recent data, noting that, in 2003, EPA
conducted another instrumented vehicle study in Kansas City. This study was much more
limited in scope than the 3 cities study used in the SFTP as each of the vehicles was only
instrumented for one day in Kansas City and the data from Kansas City was only collected for
three seasons (fall/winter/spring). This study found that the percent idle time was 17.7%.
Together, these two studies give evidence that idle rates in the U.S. could be higher than the
13.76% estimated from MOVES, and is probably higher than the 10.7% on the city/highway
test procedure. For the final rule, we are applying the more recent MOVES estimate of
13.76%, which we believe is a conservative estimate. As new data are collected, EPA will
5-86
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
continue to review idle rates to assess whether future adjustments in the credit values are
warranted.
Since operating conditions, such as cabin heating and cooling, can greatly affect
engine idle start-stop operation, we have re-evaluated the assumptions that were used in the
proposal for the percentage of vehicle operation in various ambient temperature conditions
that were used to estimate the percentage of vehicle operation that engine idle start-stop is
enabled. Based on a review of MOVES data shown in Table 5-28 below, we found that VMT
as a function of temperature is as follows: 1) 68.75% of VMT occurs between 40 deg F and
80 deg F (mid-range), 21.95% of VMT occurs below 40 deg F (cold range), and 9.69% of
VMT occurs above 80 deg F (hot range).
Table 5-28 MOVES data of vehicle miles traveled (VMT) as a function of ambient temperature.
VMT
1181.656796
4400.79767
12905.217
40874.20742
174939.1854
762497.0884
1915732.576
4924729.91
12353230.63
23259876.93
31418211.75
41033016.47
49426375.28
55404781.78
60396251 .48
63018086.25
68380740.42
73176481.47
72473451.14
67073984.17
54637578.9
39382139.05
24182451.73
7635253.418
1203687.536
593360.565
18352.30991
752904571 .9
tempAvg
-25
-20
-15
-10
-5
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
105
TotalVMT
Fraction
0.00000157
0.00000585
0.00001714
0.00005429
0.00023235
0.00101274
0.00254446
0.00654097
0.01640743
0.03089353
0.04172934
0.05449962
0.06564760
0.07358805
0.08021767
0.08369996
0.09082259
0.09719224
0.09625848
0.08908697
0.07256906
0.05230695
0.03211888
0.01014106
0.00159873
0.00078810
0.00002438
1 .00000000
Temp Range
VMT Fraction
0.21958689
(< 40 deg F)
0.68343503
(> 40 deg F, < 80 deg F)
0.09697809
(>80 deg F)
We also reviewed data from the Kansas City Study and note that it had a nearly
identical VMT-temperature distribution. Therefore, using this temperature distribution we
assumed: 1) all the mid-temperature range is available for engine-off/stop-start operation
since inhibiting factors (heater and A/C usage) are typically low in this range; 2) all the hot
temperature range requires A/C operation and would prevent engine-off/stop-start and,
consequently, none of this range is eligible for stop-start (note: it is possible that this is a
5-87
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
conservative estimate as smart A/C controls, potential cooling storage, and electric
compressors can allow engine to idle off for some hot idles while A/C is demanded); and 3)
there are several factors that are important to consider in the cold temperature range such as
average starting temperature, number of cold engine starts, time to reach sufficient engine
coolant temperature, and the average trip length. Additional information was reviewed to
refine the assumptions in the cold temperature range.
EPA also reviewed data from the Supplemental Federal Test Procedure (SFTP) study,
the EPA's MOVES model, EPA testing, and other sources to attempt to refine the estimates of
idle time in the cold ambient temperature range by looking at two key aspects: 1) the
percentage of time when the engine is cold and trip time is less than 5 minutes; and 2) the
percentage of time in the field that extended idle would occur to support cabin heating
demands.
For the percentage of time when the engine is cold and trip time is less than 5 minutes,
the SFTP Study showed that 49% of the time, the vehicle was running with engine
temperatures less than 180 degrees F, which would potentially make this portion of operation
unavailable for engine-off/stop-start because of the need to support cabin heating demands.
However, EPA test data showed that the average time to reach 90% of maximum engine
coolant is 324 seconds, which would eliminate only the first two idles in the FTP and, in
addition, would mean only trip lengths shorter than 324 seconds are ineligible for engine-
off/start-stop operation. Based on an estimate from MOVES data, 25% of the trips had a trip
time less than 5 minutes which would not achieve full warm-up and are ineligible for engine-
off/start-stop operation.
For the percentage of time in the field that extended idle would occur due to cabin heating,
based on an estimate from MOVES data, a majority of the starts (95%) have idle times less
than 5 minutes meaning that only 5% of the starts experience extended idle and are not
eligible for engine-off/stop-start operation.
Based on this information, we revised the cold temperature range to reflect the portion
of VMT that is not eligible for engine-off/stop-start operation by adding up the portions
ineligible for engine-off/start-stop operation. The estimated amount of time in the real world
when vehicles are not warmed up but idle time is greater than 5 minutes is 49% based on the
SFTP study, multiplied by 25% based on MOVES, and equaling 12.25%, is the value used for
the a-term in the equation below. The amount of time in the real world that is extended idle
of greater than 5 minutes based on MOVES activity data is 5% of the 22% VMT in the cold
range, or 1.1%, which is not eligible for engine-off/start-stop operation, and is the value used
for the b-term in the equation below. We did consider that the upper limit of 40 deg. F for the
cold temperature range was too low. However, it is possible that the upper range of 80 deg. F
5-88
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
for mid-temperature range could be higher since they are highly dependent on the relationship
between ambient temperature and humidity, and perceived passenger comfort. Therefore, we
did not change the upper range of 40 deg. F of the cold regime since these factors balance
each other.
Therefore, we revised our estimate of real-world % idle time by scaling it using these
values and the following equation:
Adjusted real-world % idle time =
Real-world % idle time x (0.6875 [mid- temperature range] + 0.2195x [l-(a+b)]
[cold temperature range])
Real-world % idle time x (0.6875 + 0.2195 x [1 - (0.1225 + 0.011)]
Real-world % idle time x (0.6875 + 0.19) =
Real-world % idle time x (0.8775) =
13.76x0.8775 = 12.07%
The above calculation assumes that the engine is warm after approximately 5 minutes
in cold weather. It also assumes that engine-off/start-stop operation occurs during these
conditions, even if heat is demanded by the passengers. For HEVs and PHEVs, this is a
reasonable assumption, as today's HEVs usually have a mechanism for idling off the engine
during cold temperatures. ICCT commented that this should not be applicable to all stop-start
systems. The agencies agree that for the calculation to be applicable to 12 Volt stop-start
systems, the vehicle should have some technology to continue to deliver heat to the cabin
even if the engine is not running. This can be done in a number of ways including an electric
heater circulation pump, secondary loops with heat reservoirs, or some other method of
maintaining heat transfer from the coolant. For this reason, the agencies are assuming that
future stop-start systems will include such technologies. In this final rule, the agencies have
removed the electric heater circulation pump from the table as proposed (and discussed
further in the next section). The manufacturer wishing to receive the full stop-start credit
must attest that the vehicle includes some technology to allow the engine to idle off while the
cabin heat is demanded. For those systems who do not include this technology, the real world
idle time calculation is modified by subtracting the credit previously proposed for electric
heater circulation pump of 1.0 g/mi for cars and 1.5 g/mi for trucks, and is reflected in the
credit values discussed below.
Below are the details of our model simulations using these values and recalculating the
stop-start, off-cycle credit. The off-cycle credit was calculated using EPA's in-house vehicle
simulation tool, known as ALPHA (Advanced Light-Duty Powertrain and Hybrid Analysis
Tool). The credit was calculated for cars and trucks, separately. Vehicle simulations were
conducted using ALPHA for small, medium, large cars and pick-up trucks. The simulations
5-89
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
were run only for FTP and Highway cycles with a baseline engine. First, the start-stop was
deactivated for the vehicle simulations. Vehicle speed trace and fuel flow simulation results
during the FTPcycle are shown in Figure 5-12 and Figure 5-13, respectively, for a small-size
car. Also, the vehicle speed trace and fuel flow during Highway cycle are shown in Figure
5-14 and Figure 5-15, respectively.
FTP Cycle Simulation
0 200 400 600 800 1000 1200 1400
Time (sec)
Figure 5-12 Vehicle Speed Trace during FTP Cycle
5-90
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Fuel Flow
3.5
2.5
o
o:
i 1.5
"Hi
0.5
0 200 400 600 800 1000 1200 1400
Time (sec)
Figure 5-13 Fuel Flow Trace during FTP Cycle without Start-Stop
Highway Cycle Simulation
"0 100 200 300 400 500 600 700
Time (sec)
Figure 5-14 Vehicle Speed Trace during Highway Cycle
5-91
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Fuel Flow
100
200
300 400 500
Time (sec)
600
700
800
Figure 5-15 Fuel Flow Trace during Highway Cycle without Start-Stop
Next, the start-stop was activated during the vehicle simulations. It must be noted that
the start-stop control algorithm was written such that the engine was not turned off during idle
for the first 300 seconds of FTP cycle to allow the engine to warm-up. Figure 5-16 and
Figure 5-17 show fuel flow traces during FTP and Highway cycles, respectively, for a small-
size car with the start-stop activated.
Fuel Flow
400
600 800
Time (sec)
1000
1200
1400
5-92
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Figure 5-16 Fuel Flow Trace during FTP Cycle with Start-Stop
Fuel Flow
2.5
•- 2
1.5
0.5
0
0 100 200 300 400 500 600 700 BOO
Time (sec)
Figure 5-17 Fuel Flow Trace during Highway Cycle with Start-Stop
The simulation results are shown in Table 5-29 and Table 5-30 below.
Table 5-29 Vehicle Simulation Results for Start-Stop in [MPG]
Driving Cycle
FTP
Highway
Combined
Start-Stop
Off
On
Improve
Off
On
Improve
Off
On
Improve
Small-Size Car
36.91
39.52
2.61
52.83
52.93
0.10
44.08
45.56
1.48
Mid-Size Car
26.39
28.82
2.42
43.93
44.04
0.12
34.28
35.67
1.38
Large-Size Car
21.28
22.95
1.67
32.34
32.42
0.07
26.26
27.21
0.95
Pick-up Truck
14.38
15.33
0.95
20.44
20.48
0.04
17.11
17.65
0.54
Table 5-30 Vehicle Simulation Results for Start-Stop in [CO2 g/mile]
Driving Cycle
FTP
Start-Stop
Off
On
Small-Size Car
240.8
224.9
Mid-Size Car
336.7
308.4
Large-Size Car
417.7
387.3
Pick-up Truck
617.8
579.6
5-93
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Highway
Combined
Improve
Off
On
Improve
Off
On
Improve
15.9
168.2
167.9
0.3
208.1
199.2
8.9
28.3
202.3
201.8
0.5
276.2
260.4
15.8
30.4
274.8
274.2
0.6
353.4
336.4
17.0
38.3
434.7
434.0
0.8
535.4
514.0
21.4
It is evident from these results that stop start has an effectiveness of about 4-5%. The
percentage of time the engine is at idle for the FTP/Highway cycles and real world are 10%
and 13.76% (MOVES analysis), respectively; allowing for the previously mentioned cold-
start warmup. In these calculations, the agencies determined that all of the 10% idle time in
FTP/Highway cycles is eligible for engine stop. However, for real world conditions, the
agencies determined a reduced engine idle time is available for engine stop; specifically
87.75%z of the real world idle time of 13.76%. Using these values, the agencies concluded
eligible engine off fractional values to be 10% for the FTP/HWFET test cycles and 12.07%
for real-world conditions. Table 5-31 shows these calculations.
Table 5-31 Percentages of Idle eligible for Engine Off
Cycle
FTP/HWFET
Real-World
Idle Fraction
10.00%
13.76%
Percentage of Idle
eligible for Engine-Off
100.00%
87.75%
Idle Fraction eligible for
Engine-Off
10.00%
12.07%
Applying the idle fraction eligible for engine off values shown in Table 5-29 to the
FTP/Highway combined cycles simulation values shown in Table 5-32, the agencies
calculated the following start-stop credit values for each vehicle segment shown in Table 5-
30.
Table 5-32 Start-Stop Credit for Each Vehicle Segment
Start-Stop Credit
C02 [g/mile]
Small-Size Car
1.8
Mid-Size Car
3.3
Large-Size Car
3.5
Full Size Truck
4.4
The impact of Stop Start is generally dependent on engine displacement because larger
engines generally have higher friction and pumping losses than smaller displacement engines
at idle, and therefore have high CO2 emissions and higher idle fuel consumption. The
differences in the credits that are available to each segment are based on the different engine
displacements typically used in each vehicle segment.
: This is due to temperature effects. Separate analysis was given earlier.
5-94
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Credits for cars and trucks are obtained by sales-weighted averaging the car credits for
model years 2017 to 2025, in Table 5-31. Note that sales-weighted averaging was used
between large-size car, which typically has similar engine displacements as smaller trucks,
and full size truck to determine the truck credit.aa
Table 5-33 Start-Stop Credits for 2017 to 2025
Year
Credit
Car
Truck
2017
2.5
4.4
2018
2.5
4.4
2019
2.5
4.4
2020
2.5
4.4
2021
2.5
4.4
2022
2.5
4.4
2023
2.5
4.4
2024
2.5
4.4
2025
2.5
4.4
Based on Table 5-33, the start-stop credits are 2.5 g/mile for cars and 4.4 g/mile for
trucks. These are values for vehicles equipped with idle-off cabin heat technologies. For
vehicles unequipped with such technologies, the credits are reduced to 1.5 g/mi for cars and
2.9 g/mi for trucks. These are the credit values that EPA is finalizing for use in the final off-
cycle credit menu.
5.2.8.2 Electric Heater Circulation Pump
Conventional vehicles use engine coolant circulated by the engine's water pump to
provide heat to the cabin during operation in cold ambient conditions. Since the coolant is
only circulated when the engine is running, very little heat is available to the cabin occupants
if the engine is stopped during idle in vehicles equipped with stop-start. Stop-start equipped
vehicles generally disable the feature during cold ambient temperatures to ensure cabin heat is
always available. However, stop-start operation can be expanded to much colder ambient if a
means of continuing to circulate coolant during idle stop is employed. An electric heater
circulation pump takes the place of the engine's water pump to continue circulating hot
coolant through the heater core when the engine is stopped during a stop-start event. Most
HEVs, and PHEVs are currently equipped with this technology; however, the more simple 12
Volt Stop start systems may not be. Therefore, by definition of this technology's function,
only vehicles equipped with stop-start technology, HEVs, and PHEVs are eligible for this
credit.
Because the engine does not generate any more heat when it is shut off during idle, the
amount of heat available to be moved to the cabin is limited by the thermal mass of the
engine. The heater core acts like a radiator to remove heat from the engine and deliver it to
the cabin. After some period of time, depending on engine mass, ambient temperature, and
desired cabin temperature, the coolant temperature would drop to a level where comfort
would not be maintained and the engine could cool off to a point where cold start features
aa Many of the assumptions made for the analysis were "conservative", others were "central". In this example,
an average vehicle (or high sales class) was selected on which the analysis was conducted. In this case, a smaller
vehicle may presumably be deserving of fewer credits whereas a larger vehicle may be deserving of more.
Where the estimates are central, it would obviously be inappropriate for the agencies to grant greater credit for
the larger vehicles since this value is already balanced by the smaller vehicles in the fleet. The agency will take
these matters into consideration when case by case applications are submitted for technologies that are
modifications to the ones listed on the menu.
5-95
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
would be needed (which increase fuel consumption). The stop-start control system would
turn the engine back on before either of these conditions is reached. The coolant circulation
pump is electrically powered and therefore uses some energy when in use.
EPA evaluated the effectiveness of this system using the same approach that was used
for start stop technology. Based on MOVES data, we assumed the percentage of nationwide
VMT below 45 °F is 25% and that vehicles with start stop systems will have to keep the
engine running for cabin heat if the ambient temperature is less than 45 °F, unless the vehicle
also has an electric heater circulation system. Therefore, we assumed a vehicle with both
systems could utilize the start stop technology 25% more of the time. However, while
reviewing our calculations for engine idle start-stop, we determined that there may be
conditions where engine idle start-stop may be enabled without the use of an electric heater
circulation pump or designs that can enable engine idle start-stop without the use of an
electric heater circulation pump. For example, Honda commented that they were planning to
implement such a system that would "to maintain all heating functions for more than 1
minute" when ambient conditions are as low as 30 deg F without the use of an electric heater
recirculation pump.
Therefore, we are eliminating the separate credit for electric heater circulation pump
and including the benefits of an electric heater circulation pump, or similar systems as implied
by Honda, within the credit for engine idle start-stop. Given the interaction and synergies
between the electric heater circulation pump and engine idle start-stop, we believe this is an
appropriate decision. For more information on how electric heater circulation pump has been
incorporated, see the discussion on engine idle start-stop above.
Finally, the Alliance and Honda commented that we revise our definition and
terminology for this vehicle. The Alliance pointed out that the main purpose is to maintain
cabin heating and "occupant thermal comfort" without using the conventional heater.
Therefore, this credit should be renamed to reflect this goal and purpose. Similarly, Honda
pointed out, as mentioned above, that they are planning on implementing a system that can
accomplish this without the use of an electric heater circulation pump. As a result, the
definition should be expanded to include any system that can maintain cabin heating and
occupant thermal comfort, not just an electric heater recirculation pump.
5.2.8.3 Active Transmission Warm-Up
When a vehicle is started and operated at cold ambient temperatures, there is
additional drag on drivetrain components due to cold lubricants becoming more viscous
which increases fuel consumption and GHG emissions. This effect is more pronounced at
colder temperatures and diminishes as the vehicle warms up. Components affected by this
additional drag include the engine, torque converter, transmission, transfer case, differential,
bearings and seals. Some components, such as the transmission, can take a long time to warm
to operating temperature. Automakers sometimes delay the application of very effective fuel-
saving measures such as torque converter lockup in order to help the transmission reach
operating temperature more quickly.
5-96
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Active Transmission warm-up uses waste heat from a vehicle's exhaust system to
warm the transmission oil to operating temperature quickly using a heat exchanger in the
exhaust system. This heat exchanger loop must have a means of being selectable, so that the
transmission fluid is not overheated under hot operating conditions. In cold temperatures, the
exhaust heat warms the transmission fluid much more quickly than if the vehicle relies on
passive heating alone. Other methods of heating the fluid can be implemented using electric
heat for example, but these are not included in this analysis because of the additional energy
consumption that would likely eliminate most of the benefit. This technology could also be
used for other driveline fluids such as axle and differential lubricant on rear-wheel-drive
vehicles or even engine oil, but only transmission fluid warming is considered here.
There is a lot of variability in which components are affected by cold temperatures and
for how long due to the type of vehicle and how it is operated. Active transmission warm-up
applied to a conventional front-wheel-drive vehicle will warm the transmission, torque
converter, and differential lubricants because in most cases these components share the same
lubricant. On a rear-wheel-drive vehicle such as a truck, active transmission warm-up would
only affect the transmission and torque converter. The rear axle and differential lubricant, and
the transfer case and front axle and differential lubricants in a four-wheel-drive vehicle would
not be heated. Additionally, a vehicle operated under a heavy load will tend to warm these
lubricants more quickly with or without active heating.
Using Ricardo modeling data and environmental data from EPA's MOVES model,
EPA calculated the estimated benefit of active transmission warm-up. The Ricardo data
indicates that there is a potential to improve GHG emissions by 7% at 20 °F if the vehicle is
fully warm. EPA assumed that given that this technology only affects the transmission (and
differential on a FWD vehicle) and that the technology does take some time to warm the
transmission fluid, one third of this benefit would be available, or 2.3%. EPA then assumed
the benefit would decay in a linear fashion to 0% at 72 °F. This simple relationship is
provided in Figure 5-18 below.
%GHG reduction via fast
warmup (trans/engine)
2
1
1
0
0
D 70
UTO n
0 20 40 60 80
Ambient Temp degF
Figure 5-18 Relationship showing linear decay of GHG improvement as a function of ambient
temperature.
Using MOVES data, EPA then calculated a nationwide VMT -weighted average
ambient temperature for all light duty vehicles. Based on the distribution data shown in
5-97
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Figure 5-18, the weighted average temperature was calculated at 58 °F and was assumed
uniform for all vehicle classes. Combined with the relationship assumed in Figure 5-19, this
weighted average temperature corresponds to an average benefit due to active transmission
warm-up of 0.58% of baseline emissions.
% VMT (from MOVES)
25%
28
39 50 61 72 83
Temperature deg F
94
104
Figure 5-19 Distribution of national VMT by ambient temperature
Finally, when this 0.58% reduction is applied to baseline emissions for various vehicle
class (as per the Ricardo-simulated 2010 baseline vehicles) the available credits, by vehicle
class, are calculated and shown in Table 5-34.
Table 5-34 Available credits (g/mi) based on fuel economy and CO2 benefits by vehicle class
Vehicle Class
Small Car
Midsize Car
Large Car
Large Truck
FTP (City)
FE70F
39.8
30.0
23.8
16.2
FTP (City)
CO2 70F
223
296
373
549
Benefit
g/mi
1.3
1.7
2.2
3.2
Using EPA's sales schedules (see TSD Chapter 1, 1.3.3, Tables 1-13 and 1-14) and
VMT's (see 5.2.8.1, Table 5-28 above) for the small car, midsize car, and large car vehicle
classes, we get average sales-weighted credit values of 1.5 grams/mi for cars, and a non-sales
weighted 3.2 grams/mi for trucks as trucks were not disaggregated by class. No benefit is
assumed during the FTP, so nothing is subtracted from this result. EPA believes an off-cycle
benefit of 1.5 and 3.2 grams/mile are possible using active transmission warm-up for cars and
trucks, respectively.
In their comments to the NPRM, the Alliance supported the credit value of 1.8
grams/mile for active transmission warm-up but recommended that the definition be
broadened to account for other methods of warm-up besides exhaust heat such as a secondary
5-98
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
coolant loop. This sentiment for an expanded definition was also expressed by Volkswagen.
Although we feel that waste heat from the exhaust system is one method, we are not opposed
to other methods that provide similar performance such as coolant loops or direct heating
elements that, albeit more costly, may prove to be more effective. Therefore, we agree with
the commenters that the definition can be expanded and will reflect this change in the
definitions section below.
The comments from Chrysler advocated for specific car and truck credits for active
transmission warm-up similar to other advanced load reduction strategies, such as engine idle
start stop and electric heater circulation pump. Based on these comments, we expanded the
range of vehicle classes/categories above and agree with the commenter that there is a clear
performance difference between active transmission warm-up systems on cars versus trucks.
Therefore, as mentioned above, we are adopting the specific car-truck credit values of 1.5
grams/mi and 3.2 grams/mile, respectively, for active transmission warm-up.
Finally, Honda's comments requested clarification on systems that use a singular heat
exchanging loop, rather than separate loops as we proposed, for active transmission and, as
discussed below in the next section, engine warm-up. Honda indicated that all of their
systems use a single heat exchanging loop for the transmission and engine, and, thus, would
potentially be eligible for an additive credit of 3.6 g/mi CO2, based on our NPRM credit
values. It is uncertain if a single heat exchanging loop would be as effective as two separate
loops for the transmission and engine, and ultimately eligible for a combined credit (e.g.,
under our revised credit values, a total of 3.0 g/mi CO2 for cars and 6.4 grams/mi CO2 for
trucks). The agencies currently do not have foundational data to support the effectiveness of a
single heat exchanging loop to increase the cold start fluid warming rate for both an engine
and transmission under various ambient temperature conditions.
Therefore, the agencies will not grant additive, default credit values for active
transmission and engine warm-up to systems using a single heat exchanging loop. Rather, a
manufacturer employing such a design will be able to initiate a credit request for such a
system would be made via the demonstration methods for technologies not on the defined
technology list. At a minimum, the request would need to demonstrate the performance of the
active transmission/engine warm-up for a single heat exchanging loop versus dedicated loops
for the transmission and engine. For such a request, if the manufacturer can demonstrate
single heat exchanging loop performance equivalent to separate, dedicated loops when
combined for both technologies, the manufacturer may be granted the revised additive credit
values of 3.0 g/mi CO2 for a car or 6.4 grams/mi CO2 for a truck, depending on the
applicable vehicle category. Otherwise, if the level of performance for a single loop system
falls short of the performance for separate cooling loops, the additive, active transmission and
active engine warm-up credit values above will be decreased proportionately to reflect the
lower performance of the single loop system. If the level of performance for a single loop
system exceeds the performance for separate cooling loops, the manufacturer may receive the
maximum, additive credit of 3.0 g/mi CO2 for a car or 6.4 grams/mi CO2 for a truck or,
alternatively, can seek credits above these values using the demonstration methods for
technologies not on the defined technology list.
5-99
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
5.2.8.4 Active Engine Warm-Up
Similar to active transmission warm-up, active engine warm-up uses waste heat from a
vehicle's exhaust system to warm targeted parts of the engine, reducing friction and cold start
enrichment requirements, and thereby increasing fuel economy. EPA assumed that similar to
active transmission warm-up, a similar magnitude benefit would also be applicable for active
engine warm-up. As a result, credit values for active engine warm-up are identical to those
for active transmission warm-up, and are additive if a manufacturer can demonstrate the
presence of both technologies (independent to one another, i.e., separate heating pathways) on
a similar vehicle. Active engine warm-up test data provided by manufacturers resulted in the
calculation of a similar emission reduction. Accordingly, the credit values of 1.5 grams/mi
for cars and 3.2 grams/mi for trucks also apply for active engine warm-up.
As discussed above for Active Transmission Warm-Up, the Alliance and Volkswagen
supported the credit value of 1.8 grams/mile for active engine warm-up but recommended that
the definition be broadened to account for other methods of warm-up besides exhaust heat
such as a secondary coolant loop. We agree with the commenters that the definition can
appropriately be expanded and will discuss this further in the section below on technology
definitions.
Also, as discussed above for Active Transmission Warm-Up, Chrysler advocated for
separate car and truck credits and Honda for combined credit for single loop systems for
active engine warm-up. Accordingly, our response above for active transmission warm-up
applies to active engine warm-up as well. Therefore, we are finalizing separate car-truck
credit values of 1.5 g/mi and 3.2 g/mi CO2, respectively, for Active Engine Warm-Up and
single loop systems must use the demonstration methods for technologies not on the defined
list.
5.2.8.5 Definitions for Non-Thermal/Solar Advanced Load Reduction Technologies
Engine start-stop is a technology which enables a vehicle to automatically turn off the engine
when the vehicle comes to a rest and restart the engine when the driver applies pressure to the
accelerator or releases the brake. Off-cycle engine start-stop credits will only be allowed if the
Administrator has made a determination under the testing and calculation provisions in 40
CFR part 600 that engine start-stop is the predominant operating mode. This technology may
be coupled with an electric heater circulation system (or a technology that has a similar
function), as described below, to receive maximum credit or may be implemented without an
electric heater circulation systems for a lower amount of credit. For systems that accomplish
the same level of performance as but do not utilize an electric heater circulation system, the
maximum level of credit may be granted provided that equivalent level of performance is
demonstrated by the requestor.
Electric heater circulation system is a system installed in a stop-start equipped vehicle, hybrid
electric vehicle or plug-in hybrid electric vehicle that continues to circulate heated air to the
cabin when the engine is stopped during a stop-start event. This system must be calibrated to
keep the engine off for 1 minute or more when the external ambient temperature is 30 deg F
and when cabin heat is demanded.
5-100
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Active transmission warm-up means a system that uses waste heat from the vehicle to warm
the transmission fluid to an operating temperature range quickly using a heat exchanger. This
reduces the parasitic losses associated with the transmission fluid, such as losses related to
friction and fluid viscosity, thereby increasing the overall transmission efficiency.
Active engine warm-up means a system using waste heat from the vehicle, to warm up
targeted parts of the engine. This reduces engine friction losses and enables the closed-loop
fuel control more quickly allowing for a faster transition from cold operation to warm
operation, thereby decreasing CO2 emissions, and increasing fuel economy.
5.2.9 Thermal (and Solar) Control Technologies
In the NPRM, EPA proposed a credit for technologies that reduce the amount of solar
energy which enters a vehicle's cabin area, reduce the amount of heat energy build-up within
the cabin when the vehicle is parked, and/or reduce the amount of cooling/heating energy
required through measures which improve passenger comfort. The State of California Air
Resources Board (CARB) has studied the effectiveness of many of these technologies, and
had proposed including them in their Cool Cars and Environmental Performance Label
programs.43 The National Renewable Energy Laboratory (NREL) conducted an extensive
research project as part of the SAE's Improved Mobile Air Conditioning Cooperative
Research Program (I-MAC). The purpose of this program was to study the effectiveness of a
variety of technologies which can reduce the amount of fuel used for the purpose of climate
control in light-duty vehicles. In this study, known as the Vehicle Ancillary Loads Reduction
Project, NREL estimated the effectiveness of window glazing/shades, paint, insulation, and
seat and cabin ventilation technologies in reducing A/C-related fuel consumption and
emissions.44 EPA has evaluated these technologies and assigned a credit amount for each,
based on their ability to reduce cabin air temperatures during soak periods and thereby reduce
the amount of cooling/heating energy required to improve passenger comfort.
NREL's studies estimated that when these technologies are combined, a 12 °C
reduction in cabin air temperature during soak will result in a 26% reduction in A/C-related
fuel consumption, or a 2.2% reduction in fuel consumption (and by extension, CC>2 emissions)
for each 1 °C reduction in cabin air temperature.45 If the A/C-related CC>2 emissions impact is
13.8 g/mi for cars and 17.2 g/mi for trucks, this 2.2% reduction in CC>2 emissions results in a
credit of 0.3 g/mi for cars (13.8 g/mi x 0.022) and a credit of 0.4 g/mi for trucks (17.2 g/mi x
0.022) for each degree centigrade reduction in cabin air temperature. There were no
comments submitted on this overall approach for the thermal and solar control technologies.
5.2.10 Glazing
When a vehicle is parked in the sun, more than half of the thermal energy that enters
the passenger compartment is solar energy transmitted through, and absorbed by, the vehicle's
glazing (or glass). 4 The solar energy is both transmitted through the glazing and directly
absorbed by interior components, which are then heated, and absorbed by the glazing, which
then heats the air in the passenger compartment through convection and interior components
through re-radiation. By reducing the amount of solar energy that is transmitted through the
glazing, interior cabin temperatures can be reduced, which results in a reduction in the amount
5-101
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
of energy needed to cool the cabin and maintain passenger comfort. Glazing technologies
exist today which can reduce the amount of solar heat gain in cabin by reflecting or absorbing
some of the infrared solar energy. NREL's study determined that cabin air temperature could
be reduced by up to 9.7 °C with use of glazing technologies on all window locations.
Technologies such as window films and coatings and absorptive or solar-reflective material
within the glazing itself are currently used in automotive glazings, both for privacy (e.g.,
tinting) and improved passenger comfort. One measure of the solar load-reducing potential
for glazing is Total Solar Transmittance, or Tts, which expresses the percentage of solar
energy which passes through the glazing. Lower Tts values for glazing result in lower cabin
temperatures during solar soak periods. EPA considers the April 15, 2008 version of the
International Organization for Standardization's (ISO) 13837 standard to be the appropriate
method for measuring the solar transmittance of glazing used in automotive applications.
A method for estimating the effect of the solar performance of glazing technologies
was developed by EPA and CARB, with input from NREL and the Enhanced Performance
Glass Automotive Association (EPGAA). This method utilizes the measured Tts of the
glazing used in a vehicle to estimate its effect on cabin temperature during soak conditions.
The contribution that each glass/glazing location on the vehicle has on the overall interior
temperature reduction is determined by its Tts (relative to a baseline level) and its area. For
purposes of this proposal, EPA considers the baseline Tts to be 62% for all glazing locations,
except for rooflites and rear side glazings of CUVs, SUVs, and minivans, which have a
baseline Tts of 40%. The relationship between the Tts value for glass/glazing and a
corresponding reduction in interior temperature is has been established using the data from
NREL testing, as shown in Table 5-35.
Table 5-35 Effect of Tts on Interior Temperature Reduction
Glass/Glazing
Position
All
All
All
All
Baseline Tts
for Glazing
Type(%)
62 (solar
absorbing)
62 (solar
absorbing)
75 (light
green)
75 (light
green)
Solar Control
Tts
40
40
50
60
Measured Breath
Air Temperature
Reduction (°C)
9
10
8
6
Estimated
Temperature
Reduction from
23. 8 °C Baseline
(°Q
15
16
8
6
bb
Glazing materials that are not subject to the requirement of >70% luminous transmittance, per 49 CFR
571.205, are often darkened "privacy" glass, and for this credit are subject to the lower baseline Tts of 40%.
5-102
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Using the NREL data and estimated temperature reductions, the linear correlation
between Tts and breath air (interior) temperature reduction was developed, and is shown in
Figure 5-11.
30
20
Figure 5-20 Correlation between Tts and Estimated Interior Temperature Reduction
From the slope of this correlation between the Tts value and reduction in cabin air
(also referred to as "breath air") temperature, a method for estimating the amount of interior
temperature reduction (in degrees Celsius) for a specific glazing location and its Tts
specification was developed, and is shown in Equation 5-6 .
Equation 5-6 - Estimated Breath Air Temperature Reduction for Glazing with Improved Solar Control
where TtsbaseUne = 62 for windshield, side-front, side-rear, rear-quarter, and backlite locations, and 40
for the rooflite location and rear side windows for SUVs, CUVs andMinivans which are typically
darkened privacy glass.
To determine the total amount of glass/glazing credit generated for a given vehicle, the
contribution (in terms of estimated temperature reduction) for each glazing location is
calculated using the glass manufacturer's Tts specification. The contribution of each glazing
location is then normalized to determine the effect each glazing location on the overall vehicle
temperature reduction. The method for normalizing the contributions is to multiply the
estimated temperature reduction of Equation 5-6 by the ratio of the glazing area of each
location divided by the total glazing area of the vehicle. The total vehicle temperature
reduction is the sum of the normalized contributions for each location. To calculate the
glazing credit generated (in grams of CO2 per mile), the sum of the total vehicle temperature
reduction (in degrees Celsius) multiplied by 0.3 for cars, or 0.4 for trucks.
5-103
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
We received several comments on the glazing credit. The ICCT agreed with the
proposed credit and the basis for the credit described in the draft TSD. In contrast, there were
multiple comments that fell into three main categories: 1) accounting for the overall glazing
surface area in the calculations and a minimum level of solar transmittance, 2) concerns
regarding metallic glazing and incentivizing this technology, and 3) granting of credit for
polycarbonate (PC) glazing technology.
The Alliance, Enhanced Protective Glass Automotive Association (EPGAA),
Guardian and Pittsburgh Glass Works (PGW) commented on the calculation accounting for
the glazing surface area where solar control glazing is applied. Each commenter
recommended that a factor accounting for the total surface area of the glazing be included in
the calculation or to account for increased effectiveness of solar control glazing for vehicles
with larger total glazing area. In addition, the Alliance suggested that a limit of 62% Tts be
used for a technology to be eligible for the glass/glazing credit. In proposed 40 CFR
§86.1866-12(d)(l)(i)(C), we included an equation to calculate the glazing credit as follows:
Credit =
Zx
where the "G" term in this equation represents "the total glass area of the vehicle, in square
meters and rounded to the nearest tenth." Therefore, the current equation takes into account
the total glazing surface area. As far as applying a limit of 62% for the glazing credit, we
agree with the commenter and the "T" term in the equation above represents the "the
estimated temperature reduction for the glass area of each window i." The T-term is
determined using the following equation:
TI = 0.3987 x (Ttsbase-Ttsnew)
where the "TtSbase" term is defined as "62 for the windshield, side-front, side-rear, rear-
quarter, and backlite locations, and 40 for rooflite locations and rear side windows for SUVs,
CUVs and Minivans which are typically darkened privacy glass." Therefore, the equation for
solar transmittance currently takes this into account and uses the 62% level as the baseline.
As a result, any glazing with a solar transmittance level of more than 62% (i.e., a higher value
implies that more solar energy is transmitted to the passenger compartment/cabin) would
result in a negative credit value. Accordingly, we are finalizing the equations above as
proposed since they address the commenter's concerns.
There were multiple comments with concerns regarding the use of metallic glazing
from the Crime Victims Unit of California (CVUC), California Manufacturers & Technology
Association (CMTA), California State Sheriffs' Association (CSSA), California Police Chiefs
Association (CPCA), California Narcotic Officers' Association (CNOA), CTIA - The
Wireless Association, Garmin, Honda and TechAmerica. Many commenters stated that low
Tts glazing uses metallic films or small metallic particles and that the credit for glazing may
unintentionally incentivize the use of this type of glazing, metallic glazing, which can
potentially interfere with signals for global positioning systems (GPS), cell phones, cell-phone
based prisoner tracking systems, emergency and/or electronic 911 (E911) calls, and other
5-104
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
signals emanating from within or being transmitted to the vehicle's passenger
compartment/cabin. In addition, some commenters cited this concern as the reason that the
California Air Resources Board (CARB) removed their mandate for metallic glazing from the
"Cool Cars" Regulation in California.
To address these concerns, we met with the Enhanced Protective Glass Automotive
Association, which represents the automotive glass manufacturers and suppliers, and
representatives from the automotive glass industry including PGW, Guardian, and AGC to
discuss the concerns with metallic glazing and the potential for signal interference and/or
radio frequency (RF) attenuation (details of this meeting are available in EPA docket # EPA-
HQ-OAR-2010-0799 and NHTSA docket #NHTSA-2010-0131). At this meeting, evidence
was provided to the agencies showing that, in general, any glazing material can create signal
interference and RF attenuation, and depending on the situation, RF attenuation and signal
interference can occur without the presence of metallic glazing material. There was no
statistically-significant increase in signal interference and RF attenuation when metallic
glazing was used, and there are deletion areas or zones without metallic solar control around
the edges and specific cut-outs in the metallic solar control films near the center of the dash
area to minimize signal interference and RF attenuation. Following the meeting, a list of
vehicles that currently use metallic glazing was also provided to the agencies demonstrating
that this technology is currently in-use without significant signal interference/RF attenuation
issues being raised.
In addition, we received comments from the California Air Resources Board (CARB)
in response to the comments on the Cool Cars Regulation. The CARB stated that the reason
they did not finalize a mandate for metallic glazing in the Cool Cars Regulation was primarily
the timing for when the signal interference and RF attenuation concerns were raised. They
also clarified that they were not requiring a specific type of glazing and that the performance-
based approach ultimately adopted in the Advanced Clean Cars Regulation accomplished the
same objectives as proposed under the Cool Cars Regulation. Finally, CARB performed
testing of signal interference and RF attenuation by CARB (see test results in EPA docket #
EPA-HQ-OAR-2010-0799-41752) echoing the findings of the automotive glass industry that
there is "[n]o effect of reflective glazing observed on monitoring ankle bracelets or cell
phones" and that any "[e]ffects on GPS navigation devices [are] completely mitigated by use
of [the] deletion window" placing either the device or the external antennae in this area.
CARB urged EPA to finalize the proposed credit values for glass and glazing as proposed.
Based on this information, the agencies are finalizing the proposed credit values and
calculation procedures for glazing. First, we are not mandating a particular technology for
glazing. The final version of the off-cycle technology menu is technology neutral with
manufacturers able to select the glazing technology based on desired performance. There are
other technologies capable of rejecting solar load from the cabin and suppliers, in their
comments, were keen to point out these alternatives. Second, we did not see evidence
contrary to the information that the automotive glass industry and CARB presented showing
that there would be significant adverse effects on signal interference and RF attenuation.
However, to allay the commenters' concerns, we will emphasize that manufacturers strongly
consider and evaluate the potential for signal interference and RF attenuation in their vehicle
design and glazing technology when requesting the solar control glazing credit.
5-105
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Next, the American Chemistry Council (ACC), Bayer Material Science, California
Manufacturers and Technology Association, CTIA-The Wireless Association, Garmin,
SAB 1C Innovative Plastics, and the Society of Plastics Industry all commented that benefits
of polycarbonate (PC) glazing should also be reflected in the amount of the menu credit for
glazing, and therefore that the automatic credit amount not be restricted to metallic glazing.
These commenters pointed to PC glazing's reduced thermal conductivity compared to glass,
which can reduce the amount of heat transmitted into the vehicle's passenger
compartment/cabin, as well as to the reduced weight of PC glazing compared to other
materials, potentially having mass reduction benefits as well. Some commenters advocated
for a separate credit for PC glazing equivalent to the overall glazing credit we proposed.
Further, SABIC Innovative Plastics supplied an equation to calculate thermal conductivity
similar to the one we proposed for calculating Tts.
As stated above, we are not mandating a particular technology for glazing and,
therefore, do not need believe that it is necessary to offer a separate PC glazing credit since
this credit covers all types of glazing technologies. Also, one of the main issues with
allowing a separate credit for PC glazing is that it is more effective when the vehicle is in
motion since air flow cools the glazing surface and less thermal conduction occurs. Thus, this
application is limited to a narrower operating regime and would have less effectiveness in that
regime considering that the cabin will be cooler and require external air flow and internal air
conditioning.
Additionally, we do not have information, at this time, to support the equation that
SABIC supplied to account for thermal conductivity. In contrast, for solar transmittance,
there are established ISO procedures (ISO 13837) that can be used and referenced to ensure a
consistent basis for information supporting the credit request. We need to have a similar,
established set of procedures to validate the equations, and substantiate a credit. Therefore,
we are not including the specific equations used to calculate thermal conductivity for the
defined technology list at this time. If manufacturers still believe that there may be some
additional benefit, they may apply for additional glazing credit using the demonstration
methods for technologies not on the defined technology list.
Finally, there was a comment from Honda advocating for the use of direct solar
transmittance (Tds) as a measure of solar/thermal control benefits rather than Tts. While there
may be some benefit to the use of Tds, we are not aware of sufficient data to determine 1) the
total effectiveness of Tds and 2) if the amount of glazing credit is appropriate based on the
effectiveness of Tds. Therefore, we are not including Tds technology as part of the glazing
credit on the defined technology list. However, this technology may be eligible to generate
off cycle credits based on the case-by-case demonstration procedure in the rules.
In summary, the credits, definitions and terminology for glazing will be finalized in
today's action as proposed.
5.2.11 Active Seat Ventilation
The NREL study investigated the effect that ventilating the seating surface has on the
cooling demand for a vehicle. By utilizing a fan to actively remove heated, humid air that is
5-106
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
typically trapped between the passenger and the seating surface, passenger comfort can be
improved, and NREL's Thermal Comfort Model predicted that A/C system cooling load
could be reduced, and a 7.5% reduction in A/C-related emissions can be realized.45 While
seat ventilation technology does not lower the cabin air temperature, it indirectly affects the
load placed on the A/C system through the occupants selecting a reduced cooling demand due
to their perception of improved comfort. Using the EPA estimate for the A/C-related CO2
emissions impact of 13.8 g/mi for cars and 17.2 g/mi for trucks, a 7.5% reduction in CO2
emissions with active seat ventilation results in a credit of 1.0 g/mi for cars (13.8 g/mi x
0.075) and a credit of 1.3 g/mi for trucks (17.2 g/mi x 0.075).
We received four comments on Active Seat Ventilation. The Alliance supported our
proposed credit values of 1.0 g/mi for cars and 1.3 g/mi for trucks. In contrast, ICCT felt
these numbers were modest based on the limited NREL dataset we used and other studies
from the University of Denmark and a journal on "Ergonomics" that showed up to a 6.4 deg
Celsius change in cabin temperature could be tolerated with cooled seats, equating to a credit
of 1.9 g/mi for cars and 2.6 g/mi for trucks. However, as ICCT also points out, this is highly
dependent on driver comfort perception and response and the real-world impact may be lower
than anticipated in the studies the commenter cited. Therefore, we are finalizing the menu
default credit values of 1.0 g/mi for cars and 1.3 g/mi for trucks in today's action as proposed.
In addition, the Alliance and MEMA commented that the definition for Active Seat
Ventilation was too narrowly defined since it only contemplated a suction-type system to pull
heat and reduce moisture from the seating surface. Specifically, they stated that the use of a
forced-air system to push heat and reduce moisture from the seating surface is just as
effective. We agree with these comments and are finalizing a broadened definition that
allows for forced-air as well as suction-type systems. Finally, the Alliance also suggested that
active seat ventilation technology need only be applied to the front seats in order to qualify for
the credit. We agree with this comment since some vehicles do not have seat heaters on the
rear seats or in the case of vehicles that only have two seats. Therefore, we will specify that,
at a minimum, the front driver and passenger seat, or in the case of a two-seat vehicle, driver
and passenger seats, must have active seat ventilation for a vehicle to be eligible for credit. In
summary, we are finalizing the credit values of 1.0 g/mi for cars and 1.3 g/mi for trucks, as
proposed, with the modifications to the definition for active seat ventilation technology.
5.2.12 Solar Reflective Paint
As the vehicle's body surface is heated by solar energy when parked, heat is
transferred to the cabin through conduction and convection. Paint or coatings which increase
the amount of infrared solar energy that is reflected from the vehicle surface can reduce cabin
temperature during these solar soak periods. While the amount of heat entering the cabin
through the body surface is less than that which enters through the glazing, its effect on cabin
air heat gain is measureable. NREL testing estimated that solar-reflective paint and coatings
can reduce cabin air temperature by approximately 1°C, whereas glazing technologies can
reduce cabin air temperature by up 10°C. Using the EPA estimate for credits due to cabin air
temperature reductions of 0.3 g/mi for cars 0.4 g/mi for trucks for each degree centigrade of
temperature reduction, a 1.2°C reduction due to solar reflective paint results in a credit of 0.4
g/mi for cars and 0.5 g/mi for trucks.
5-107
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
The comments on solar reflective paint were primarily supportive of the credit amount
we proposed of 0.4 g/mi for cars and 0.5 g/mi for trucks for a 1.2°C of temperature reduction.
The Alliance supported this level of credit, although they thought it might be ambitious due to
the worst case test conditions in the National Renewable Energy Laboratory (NREL) we used
as the basis for calculating the credit. The ICCT supported the credit since it met their
"general principle of being verifiable and additive."
Honda commented that the solar reflective paint credit should only apply to certain
colors on a sales-weighted basis since some colors cannot reflect at least 65 percent of the
impinging infrared solar energy as required under ASTM standards E903, El918-06, or
C1549-09 for measuring solar reflectiveness. In addition, Honda also stated that the credit
should only be applicable to the horizontal surfaces of the vehicle since these areas will be
exposed to the most solar loading where solar reflective paint would be more effective. The
credit for solar reflective paint is performance based (i.e. it is a scalar) so a manufacturer
would have to demonstrate what level of temperature reduction they are achieving.
Therefore, a lower level of credit would be granted under Honda's approach if a certain color
of paint demonstrates a lower level of temperature reduction. Regarding the comment on
application of the credit for only horizontal surfaces, it may be possible that other surfaces
(e.g., side quarter/door panels) under direct solar loading may be beneficial but this benefit is
limited by the orientation of the vehicle to sunlight. As a result, we can't guarantee that other
areas of the vehicle will be exposed, if at all, to the extent of the horizontal surfaces.
Therefore, we agree that the horizontal surfaces are the prime areas of benefit and will revise
the definition for solar reflective paint to state that only the horizontal surfaces utilizing solar
reflective paint will receive any credit under the technology menu. For the same reason, we
do not believe that there ever could be a demonstration that other-than horizontal surfaces
would generate off-cycle credits and therefore other-than-horizontal surfaces would not be
eligible to generate off-cycle credits under a case-by-case demonstration. In summary, we
will finalize the credit of 0.4 g/mi for cars and 0.5 g/mi for trucks for a 1.2 degrees centigrade
of temperature reduction in today's action with the revised criteria that only the horizontal
surfaces utilizing solar reflective paint will receive any credit.
5.2.13 Passive and Active Cabin Ventilation
Given that today's vehicles are fairly well sealed (from an air leakage standpoint), the
solar energy that enters the cabin area through conductive and convective heat transfer is
effectively trapped within the cabin. During soak periods, this heat gain builds, increasing the
temperature of the cabin air as well as that of all components inside the cabin (i.e. the thermal
mass). By venting this heated cabin air to the outside of the vehicle and allowing fresh air to
enter, the heat gain inside the vehicle during soak periods can be reduced. The NREL study
demonstrated that active cabin ventilation technology, where electric fans are used to pull
heated air from the cabin, a temperature reduction of 6.9 °C can be realized. For passive
ventilation technologies, such as opening of windows and/or sunroofs are and use of floor
vents to supply fresh air to the cabin (which enhances convective airflow), a cabin air
temperature reduction of 5.7 °C can be realized.45 Using the EPA estimate for credits due to
cabin air temperature reductions of 0.3 g/mi for cars 0.4 g/mi for trucks for each degree
centigrade of temperature reduction, a 6.9 °C reduction due to active cabin ventilation results
5-108
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
in a credit of 2.1 g/mi for cars and 2.8 g/mi for trucks. For passive cabin ventilation, a 5.7 °C
temperature reduction results in a credit of 1.7 g/mi for cars and 2.3 g/mi for trucks.
There were two main commenters regarding passive and active cabin ventilation: the
Alliance and ICCT. The Alliance felt that a significant credit was necessary to enable
implementation and, thus, supported the proposed credit values. Their only suggestion was to
broaden the definition for passive cabin ventilation from ".. .ducts or devices which utilize
convective airflow to move heated air from the cabin interior to the exterior of the vehicle." to
include the word "methods" (e.g., ".. .ducts, devices or methods...") since there are other
ways to perform passive cabin ventilation without the use of ducts or devices. For example,
the Toyota Prius and other vehicles will lower the side windows 1/2 to 1 inch to allow for
convection. This works absent a duct or device and the broadened definition would provide
for such methods.
The ICCT commented that the NREL report used as the basis for this credit lacked
sufficient data and identified the possibility of intrusion in the case of floor-level ventilation.
Therefore, ICCT stated that the credit for active/passive cabin ventilation should be deferred
to establish a real-world benefit and a proper verification benchmark can be established. In
response, the Alliance submitted supplemental comments presenting data from vehicles in the
existing fleet and a study commissioned by General Motors and conducted by NREL to
examine various active and passive cabin ventilation technologies. As stated by the Alliance,
this study demonstrated the ability to achieve cabin temperature reductions on five of the
twelve vehicles of greater than the 6.9 degrees Celsius in the NREL report cited in the TSD,
with temperatures reductions as high as 11.4 and as low as 7.2 degrees Celsius. The agencies
have carefully evaluated these studies and believe that they address the ICCT's concerns
regarding supporting data to support the level of credit proposed.
Therefore, we are finalizing the default credit for active/passive cabin ventilation as
proposed. In addition, as discussed above, we agree with the comments regarding the
definition for active/passive cabin ventilation and will expand the definition to include
"methods" that may be employed to achieve the same objective as "ducts and devices."
5.2.14 Summary of Thermal (and Solar) Control Credits
The amount of credit that a manufacturer can generate for thermal and solar control
technologies is shown in Table 5-36.
Table 5-36 Off-Cycle Credits for Thermal Control Technologies
Thermal Control
Technology
Glass or glazing
Active Seat Ventilation
Solar reflective paint
Passive cabin ventilation
Active cabin ventilation**
Estimated Breath Air
Temp. Reduction
up to 9.7 °C
N/A*
1.2 °C
5.7°C
6.9 °C
Credit (g CO2/mi)
Car
up to 2.9
1.0
0.4
1.7
2.1
Truck
up to 3. 9
1.3
0.5
2.3
2.8
* Active seat ventilation is not a temperature reduction technology, but rather a comfort control
technology, capable of reducing A/C-related emissions by 7.5%
5-109
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
** Active cabin ventilation has potential synergies with solar panels as described in Chapter 5.2 of this
joint TSD.
To generate off-cycle thermal control credits - up to a maximum of 3.0 g/mi for cars,
and 4.3 g/mi for trucks - a vehicle must be equipped with the thermal control technology, in
accordance with the specifications and definitions in this proposed rulemaking. If a
technology meets the specifications, its use in a vehicle will generate credits, in accordance
with the value set forth in the thermal control technology list. The one exception to a single
credit value for a technology is glazing technologies, where the method for determining the
credit is described in section 5.2.10.
5.2.15 Definitions for Solar Control Credit Technologies
Credit for solar control technologies can be generated for MY 2017-2025 vehicles
which utilize them. In the absence of a performance test to measure the affect of these
technologies, For all solar control technologies except glazing, EPA will rely on
manufacturers complying with a specification for, or description of, each technology to assure
that the emissions reducing benefits are be realized in real-world applications. Below are the
descriptions and specifications that EPA is adopting for the solar control technologies listed in
Table 5-36. EPA will use these definitions and specifications to determine whether the credits
are applicable to a vehicle.
• Active Seat Ventilation - device which draws air, forces air or transfers heat
from the seating surface which is in contact with the occupant and exhausts it
to a location away from the seat. At a minimum, the front driver and passenger
seat must utilize this technology for a vehicle to be eligible for credit. If the
vehicle only has two seats, then these seats must have active seat ventilation
for a vehicle to be eligible for credit.
• Solar Reflective Paint - vehicle paint or surface coating applied to the
horizontal surfaces, including the rear decklid and cabin roof, which reflects at
least 65 percent of the impinging infrared solar energy, as determined using
ASTM standards E903, El918-06, or C1549-09
• Passive Cabin Ventilation - ducts, devices or methods which utilize convective
airflow to move heated air from the cabin interior to the exterior of the vehicle
• Active Cabin Ventilation - devices which mechanically move heated air from
the cabin interior to the exterior of the vehicle
5.2.16 Summary of Credits
Table 5-37 summarizes the preapproved technologies and off-cycle credits available to
manufacturers. If manufacturers wish to receive off-cycle credits for other technologies, they
must follow the procedures laid out in section III.C.5 of the Preamble and in the regulations at
40 CFR §86.1869-12 (b) and (c).
Table 5-37: Initial off-cycle credit estimates (Maximum Available Credits)
5-110
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Technology
+High Efficiency Exterior Lights*
(at 100 watt savings)
+Waste Heat Recovery (at 100W)
+Solar Panels (based
on a 75 watt solar
panel)**
Battery Charging
Only
Active Cabin
Ventilation and
Battery Charging
+Active Aerodynamic Improvements (for a
3% aerodynamic drag or Cd reduction)
Engine Idle Start-Stop
w/ heater
$
circulation system
w/o heater
circulation system
Active Transmission Warm-Up
Active Engine Warm-up
Solar/Thermal Control
Adjustments for Cars
g/mi
1.0
0.7
3.3
2.5
0.6
2.5
1.5
1.5
1.5
Up to 3.0
gallons/mi
0.000113
0.000079
0.000372
0.000282
0.000068
0.000282
0.000169
0.000169
0.000169
0.000338
Adjustments for Trucks
g/mi
1.0
0.7
3.3
2.5
1.0
4.4
2.9
3.2
3.2
Up to 4.3
gallons/mi
0.000113
0.000079
0.000372
0.000282
0.000113
0.000496
0.000327
0.000361
0.000361
0.000484
* High efficiency exterior lighting credit is scalable based on lighting components selected from high efficiency
exterior lighting list (see Joint TSD Section 5.2.3, Table 5-21).
** Solar Panel credit is scalable based on solar panel rated power, (see Joint TSD Section 5.2.4). This credit can
be combined with active cabin ventilation credits.
#
In order to receive the maximum engine idle start stop, the heater circulation system must be calibrated to keep
the engine off for 1 minute or more when the external ambient temperature is 30 deg F and when cabin heat is
demanded (see Joint TSD Section 5.2.8.1).
This credit is scalable; however, only a minimum credit of 0.05 g/mi CO2 can be granted.
5.3 Full-Size Pickup Truck Credits
The agencies recognize that the MY 2017-2025 standards will be challenging for large
trucks, including full size pickup trucks that are often used for commercial purposes, and so are
taking steps to incentivize the penetration into the marketplace of "game changing"
technologies for these pickups, including their hybridization. EPA proposed and is adopting
per-vehicle credits for manufacturers that sell substantial numbers of mild or strong hybrid
full size pickup trucks. The credit is 10 g/mi and 20 g/mi for mild and strong hybrids,
respectively. EPA also proposed and is adopting a performance-based incentive credit for full
size pickup trucks that achieve significant emissions reductions below the target level that
corresponds to their footprint. The credit is 10 g/mi for pickups achieving 15% better CC>2
than their target, and 20 g/mi for pickups achieving 20% better CC>2 than their target. Access
to all of these credits in any given model year is conditioned on achieving a minimum
penetration of the technology in a manufacturer's full size pickup truck sales fleet:
• For strong hybrid credits: 10% in each model year 2017 through 2025.
• For mild hybrid credits: 20-30-55-70-80% in model years 2017-2018-2019-
2020-2021, respectively.
5-111
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
• For "20 percent better" performance-based credits: 10% in each model year
2017 through 2025.
• For "15 percent better" performance-based credits: 15-20-28-35-40% in model
years 2017-2018-2019-2020-2021, respectively.
A number of comments were received on the proposed minimum penetration
thresholds. These comments, and EPA's response to them, are discussed in preamble section
III.C.3. Credits are not available after 2025 for strong hybrids and "20 percent better"
performance, or 2021 for mild hybrids and "15 percent better" performance. Unlike the
hybrid credits, the performance-based credits have no technology or design requirements.
Automakers can use any technology as long as the vehicle's CC>2 performance is at least 15%
or 20% below its footprint-based target. A vehicle cannot receive both the hybrid and
performance-based credit. EPA and NHTSA are coordinating to allow manufacturers to
include "fuel consumption improvement values", equivalent to these EPA CC>2 credits, in the
CAFE program.
5.3.1 Full-Size Pick-up Truck Definition
As proposed, EPA is defining a full size pickup truck based on minimum bed size and
hauling capability, as detailed in 86.1866-12(e) of the regulations being adopted. This
definition is meant to ensure that the larger pickup trucks which provide significant utility
with respect to payload and towing capacity, as well as open beds with large cargo capacity,
are captured by the definition, while smaller pickup trucks which have more limited hauling,
payload and/or towing are not covered. A full size pickup truck is defined as meeting
requirements (1) and (2) below, as well as either requirement (3) or (4) below:
1) Bed Width — The vehicle must have an open cargo box with a minimum width
between the wheelhouses of 48 inches, measured as the minimum lateral distance between the
limiting interferences (pass-through) of the wheelhouses, excluding any transitional arc, local
protrusions, and depressions or pockets (dimension W202 in SAE Procedure Jl 100). An open
cargo box means a cargo bed without a permanent roof or cover. Vehicles sold with
detachable covers are considered "open" for the purposes of these criteria. And—
2) Bed Length — The length of the open cargo box must be at least 60 inches, as
measured at both the top of the body and at the bed floor (dimensions L506 and L505 in SAE
Procedure 11100). And-
3) Towing Capability - the gross combined weight rating (GCWR) minus the
gross vehicle weight rating (GVWR) must be at least 5,000 pounds. Or—
4) Payload Capability - the GVWR minus the curb weight (as defined in 40 CFR
86.1803) must be at least 1,700 pounds.
This definition is being finalized as proposed. The comments that were received on
the definition, and our responses, are discussed in preamble section II.F.3.
5-112
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
5.3.2 Hybrid Pickup Truck Technology
5.3.2.1 Mild Hybrid Technology
Often a mild hybrid is characterized by the addition of a belt-driven starter-alternator
of higher power capacity than a standard alternator. The drive belt system also typically has a
feature that enables the belt tension to be maintained at proper levels during generator
operation as well as when the starter-alternator is used to start the engine. Alternatively, an
axial motor can be mounted on the crankshaft, usually in the bell housing before the
transmission. This motor can be directly attached to the engine, or can be clutched to
decouple it from the engine. The vehicle uses a conventional transmission such as an
automatic, manual, CVT, or DCT with an appropriate conventional coupling such as a torque
converter or clutch.
The battery can be between 36V to over 150V nominal; generally the higher the
voltage, the higher the performance of the system. Most mild hybrid pickups are expected to
offer at least 100V of battery voltage due to the higher power requirement of these heavy
vehicles. Mild hybrids are capable of start-stop operation and regenerative braking, but unlike
strong hybrids they are not capable of any significant electric-only operation.
Mild hybrids are less capable than strong hybrids because of lower power capability,
but mild hybrids are lower cost and may be easier to adapt to some vehicles without making
major powertrain, chassis or body changes.
5.3.2.2 Strong Hybrid Technology
Strong hybrids can take several forms. One type has an integrated transmission-drive
motor system with a large, powerful electric drive motor-generator (often two motors). The
transmission usually is specifically designed to integrate the motor-generator(s) and often the
coupling between the engine and transmission such as a torque converter is removed with its
functions handled by the electric drive motor system. The transmission can also be replaced
by a power split device that uses a planetary gearset and two motor-generators, or a P2
arrangement can be used with a conventional transmission augmented by an electric motor.
Strong hybrids typically have high voltage battery packs over 300 V to provide the high
power necessary for their increased capability.
Strong hybrids are capable of start-stop operation, have significant braking
regeneration capability, and are often capable of driving exclusively on battery power up to
35-45 mph. They are also capable of launching the vehicle on electric drive alone, although
they typically cannot accelerate above 15-20 mph while operating on electric drive
exclusively.
5.3.3 Mild and Strong Hybrid Pickup Truck Definitions
In addition to meeting the definition for a full-size truck, a vehicle must meet
additional design and performance requirements to be eligible for the hybrid full-size truck
incentive. Mild and strong hybrids must have both stop-start capability and regenerative
braking. Additionally, the level of hybridization (mild or strong) is characterized by the
5-113
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
amount of energy recovered into the battery: at least 65% for strong hybrids, and at least 15%
but less than 65% for mild hybrids. These thresholds and the methodology for determining
the amount of recovered energy are discussed below.
5.3.3.1 Measurement of Recovered Energy
EPA is incorporating a metric - the total percentage of available vehicle energy
recovered over the test cycle - as a way to define levels of hybrid vehicles. For a given
vehicle and road load profile (characterized by ETW and A, B and C dyno test "coastdown"
coefficients), a theoretical amount of required braking energy can be calculated over the city
and highway test cycles. This maximum braking energy is the sum of the extra braking force
needed to slow the vehicle enough to follow the test cycle trace upon decelerations. Hybrids
recapture a portion of this energy by driving the electric motor (in reverse) as a generator,
which ultimately provides electrical power to the battery pack. Depending on the level of
hybridization, this amount of recaptured energy can range between a few percent of total
available braking energy, up to and potentially exceeding 100% of all braking energy (since
some manufacturers also charge the battery via excess engine load when it is beneficial to do
so).
This metric is a way to simplify the characterization of a hybrid as a "mild" or
"strong" hybrid. Batteries and motors must increase in scale to recover energy at a greater
rate. As the power rating of the motor and battery increases, a greater percentage of energy
can be recovered on rapid decelerations. So, all key facets of a hybrid system - the battery
pack size and power rating, the motor rating, etc. - are implicitly reflected in the percentage
of energy recovered.
The procedure involves calculating the available braking energy on the FTP city cycle
using the equation derived below. This value is compared to the actual energy recovered by
the vehicle during FTP city cycle testing. Since energy into and out of the hybrid drive
system battery is a standard part of emissions testing of hybrid vehicles, this procedure
introduces no additional test burden. However, energy flow into the battery must be separated
from the sum of energy into and out of the battery which is typically less than 1% of total fuel
energy used during the test.
The fact that some of the energy going into the battery may come directly from the
engine means that the measured energy flow over the FTP is not, strictly speaking, just the
recovered braking energy. Some manufacturers commenting on the proposal expressed
concern that this would make the categorization of mild and strong hybrids subject to gaming
by manufacturers seeking credits. They suggested that we replace this metric with one that
only integrates current flow during decelerations (with correspondingly revised thresholds for
percentage of energy recovered), and that we also add a second metric based on battery-
supplied tractive effort only.
We have evaluated these concerns and have concluded that the proposed metric
remains adequate for our purposes, and furthermore has the advantage of being simpler and
easier to measure than other metrics, such as measured current flow only during deceleration
periods with zero fuel flow, or only during applied braking. Even these metrics would not
5-114
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
completely remove the potential for inclusion of engine-to-battery energy flow as sought by
the commenters. The data that EPA collected on a 2-mode hybrid truck, discussed below,
indicates that there is a strong correlation between EPA's proposed metric and the energy
recovery metric suggested by commenters. Moreover, EPA believes that adding more
constraints in the characterization of hybrids, such as battery-only traction, would put too
much emphasis on the particular hybrid design strategy, and our preference is to remain
neutral on how these technologies are implemented. More fundamentally, we feel the total
cycle energy-to-battery metric, matched with the corresponding mild and strong HEV
thresholds we are setting, provides a fair indication of the degree of hybridization in the
design because, given the expense involved in using larger electrical components, we would
expect any energy flow directly from the engine to the battery to stem from real efforts to
optimize HEV design for performance and fuel economy rather than from gaming for credit
generation. This view is backed by the fact that the practice is common in today's hybrids
where there is no potential for credits. To keep from causing confusion, we are avoiding
calling the parameter that is derived from current measurement "recovered braking energy,"
instead simply calling it "recovered energy."
The measured energy into the battery is divided by the total calculated braking energy
to determine if the vehicle is a mild or strong hybrid. We proposed that the recovered energy
for a mild hybrid must be greater than or equal to 15% and less than 75% of the calculated
available braking energy, and that the recovered energy for a strong hybrid must be at or
above 75%. We based these proposed thresholds on available test data collected on hybrid
vehicles, none of which were large pickup trucks. Chrysler commented that the 75%
threshold for strong hybrids, though it may be appropriate for passenger cars, is too
demanding for large pickup trucks designed with powerful braking systems to safely handle
large towing loads. In response, EPA has conducted tests, using the methodology described
in this TSD section, on a Silverado 1500 2-mode HEV truck. This is the only large light-duty
truck currently on the market that is generally considered to be a strong hybrid.46 4? The
results over 6 repeat tests varied from 68% to 78%. Based on this testing, we believe 65% is a
more appropriate threshold than the proposed 75% for defining strong hybrids, and so are
adopting this threshold into the final regulations. We are retaining the proposed 15%
threshold for mild hybrids, consistent with comments received.
It should be kept in mind that these thresholds and the associated metric for evaluating
candidate hybrids are intended to provide a general, non-technology-specific parallel to the
hybrid technology overview in section 5.3.2 above. Their purpose is to clearly, fairly, and as
simply as possible define eligibility for credits. They are not meant to be used in any way as
an industry standard for these terms. We recognize too that technology evolution or new
information may make it helpful to reconsider these criteria, and believe that the mid-term
review may provide a suitable forum for doing so.
5.3.3.2 Spreadsheet documentation and calculation methodology details
Equation 5-7 defines the brake energy recovery efficiency (expressed as a percentage),
Or Tjrecovery-
Equation 5-7:
5-115
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
-_ recovered
'recovery T--
brake max
vered, the total energy recovered over the 4-bag FTP test (in kWh) is calculated in
Equation 5-8.
Equation 5-8:
/T
recovered
3600*1000
With i(t) defined as measured current into the battery (in amps) and V defined as the
nominal battery pack voltage. Current flowing out of the battery (discharge) is not included.
Battery current is measured via a current clamp probe, mounted directly upstream of the
battery pack. Both battery current and vehicle speed data should be collected at a sampling
rate of 10 Hz.
We received comments expressing concern that nominal voltage is a poorly-defined
term, determined in the industry in a variety of ways, and is therefore subject to gaming. In
response, and after discussion with industry representatives including developers of SAE
J271 1,48 we are defining nominal voltage in the following manner: Determine nominal
voltage of the battery by taking one battery voltage measurement immediately following
"key-on" for the FTP, taking a second battery voltage measurement immediately prior to
"key-off for the FTP, and then averaging the pre- and post-FTP voltages. The initial voltage
measurement may occur any time between "key-on" and up to 10 seconds following the "key-
on" event. The second voltage measurement may occur up to 10 seconds before the "key-off
event". Based on data we have reviewed from actual vehicle testing, we expect that this
straightforward methodology will be adequate for the purposes of this credit program, because
current flow at these times is typically very low.49 However, if a manufacturer' s test data
shows that the absolute value of the measured current to and from the battery during either of
the voltage measurements exceeds 3.0% of the maximum absolute value of the current
measured over the FTP, the manufacturer is expected to develop an alternative means of
determining nominal voltage, subject to EPA approval.
In order to allow verifiable measurements of nominal voltage, the manufacturer
wishing to make use of this optional credit provision will need to broadcast battery pack
voltage on an on-board diagnostics (OBD) parameter ID (PID) channel. Battery voltage is
already publically available on enhanced PIDs but the protocol is not consistent across
manufacturers. Making the data available on an OBD PID will make the procedure for
recording battery voltage consistent across vehicle manufacturers and will allow verification
of nominal voltage measurements during confirmatory and other testing by EPA.
5-116
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
max (kWh) is calculated by integrating required braking power (Pbrake) at each
point in the test cycle00 over the entire test, shown in Equation 5-9. For clarity, the prescribed
vehicle speed test schedule (not the recorded vehicle speed test data) is used in this
calculation.
Equation 5-9
brake_max o ^r\r\
joOO
Pbrake (kW) - the vehicle braking power required to follow the drive trace during
decelerations - represents the amount of braking force (expressed as power) in addition to the
existing road load forces which combine to slow the vehicle. It is expressed in Equation 5-10.
By convention, only negative values are calculated for braking.dd
Equation 5-10
P — P — P
brake accel_reqd roadload
Paccei reqd (kW), in represents the total applied deceleration power necessary to slow
the vehicle. It is calculated as the vehicle speed, v (in m/s) multiplied by the deceleration
force (vehicle mass * required deceleration rate), as shown in Equation 5-11.
Equation 5-11
^ ^dv
"accel_ reqd ~ V mETW ~T
Where:
(kg) is the mass of the vehicle based on equivalent test weight (ETW)
dv/dt (m/s2) is the required acceleration/deceleration for the vehicle to match the next
point on the vehicle test trace (as recorded at a 10 Hz rate)
Proadioad (kW) is the sum of the road load forces (N) as calculated from the
experimental vehicle coastdown coefficients, A, B and C. It is calculated in Equation 5-12.
00 These calculations assume a "4-bag" FTP schedule, or 2 consecutive UDDS cycles (cold-start UDDS with a 10
minute soak period and a second hot-start UDDS), as is common for testing HEVs for charge balancing
purposes.
dd All power terms are negative when power is applied to the vehicle (as in braking). Power provided by the
vehicle (such as tractive power - in the case of acceleration) would be positive.
5-117
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
Per convention, road load is negative as it always represents a deceleration (resistive) force
acting on the vehicle.
Equation 5-12
5.3.4 Pickup Truck Performance Thresholds for Advanced Technology Credits
This section describes how the agencies arrived at the threshold values of 15% and
20% better than the footprint target required to qualify a pickup truck for performance-based
advanced technology credits.
Based on the lumped parameter model (described in Chapter 3 of this joint TSD),
pickup truck hybrids are determined to be approximately 15% more efficient than non-
hybrids. However, this can vary over a range of efficiencies depending on the weight,
electrification level, HEV architecture, engine/transmission, utility ratings, control strategy,
etc. Rather than comparing directly to a given HEV technology, we have instead determined
the thresholds based on the year-on-year stringency of the standards (targets). Although we
discuss these thresholds in terms of GHG standards, a corresponding analysis could be made
based on fuel economy targets.
The targets (curve standards at a given footprint) become more stringent each year.
However, a typical vehicle model is redesigned every 5 years (6 years for some larger trucks).
When a vehicle model is redesigned, it is assumed that the emissions will not just meet the
footprint target, but rather exceed it so that in general the vehicle is generating credits for the
first two or three years of the product life, and using credits for the latter two or three years,
until the next redesign. Although no individual vehicle is required to meet its footprint target,
the manufacturer must meet its fleet obligation based on the footprint and sales volumes of all
the light-duty vehicles it produces. Therefore, under normal (business-as-usual)
circumstances, each manufacturer will be designing and redesigning some of their vehicle
models each year and some vehicles will exceed their targets (for about 2-3 years each) and
others will fall short (for about 2-3 years each), thus allowing the manufacturer to average its
fleet in order to comply each year. Recognition of this product development cycle is an
important element of the program structure.
In the following hypothetical example, illustrated in Figure 5-21, a recently redesigned
58 square foot pickup truck is certified in MY 2018. Its target is 308 g/mi. This truck will
not receive another redesign until 2022. Under normal circumstances, a typical vehicle would
likely be 10% better than the standard, which would make it a credit generator for three years
and a deficit generator for two years (consistent with the usual regulatory strategy outlined in
the previous paragraph). At 15% below target (262 g/mi) this truck will generate credits for
four years and deficits only in its last year. At 20% below target (246 g/mi), the truck will
generate credits for the full five year product development cycle.
Figure 5-21 2017-2025 Truck GHG Standard Curves, with Example Redesign of a 58 square foot truck
5-118
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
360
340
200
50
Vehicle Footprint (squarefeet)
•2016 2017 ---2018 2019 2020 2021 2022 ---2023 2024 2025
The analysis depends somewhat on the footprint selected. Table 5-38 shows the truck
footprint targets for each model year for three sample trucks with footprints: 58, 67 and 74 sq
ft, and three scenarios: 10%, 15% and 20% better than the target value. The table also shows
the number of years each of the sample trucks would take to start generating deficits. In the
10% scenario, the trucks create deficits in 3.4 years on average. In the 15% scenario, it takes
4.7 years and for the 20% scenario it takes 5.7 years on average. Based on this analysis, the
agencies have chosen the 15% and 20% thresholds, as these are significantly better than the
business-as-usual (-10%) scenario. The performance thresholds of 15% and 20% therefore
represent CO2 reductions greater than what EPA expects companies would typically plan for
during a redesign of these products, given the level of the standards and the CC>2 targets for
typical full-size pickup trucks. In addition, since the rate of improvement of using hybrid
technology on full size pickup trucks is approximately 15% (as noted at the start of this
section), this rate of improvement over target is comparable to what would be achieved by
applying hybrid technologies to the same vehicles. It is consequently reasonable to provide
an equivalent credit amount. These levels are also technically within reach of the companies
if they pull ahead technologies which they may not otherwise need until the later years of the
program, or in the case of the later years of the program, a pull-ahead of technologies beyond
what is needed for MY2025.
Table 5-38: Truck CO2 Footprint Targets for 10%, 15% and 20% Thresholds
5-119
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
2017
2018
2019
2020
2021
2022
2023
2024
2025
Footprint
58.0
315
308
299
290
268
255
243
231
220
67.0
347
342
339
331
307
292
278
265
252
74.0
347
342
339
337
335
321
306
291
277
2017
2018
2019
2020
2021
2022
2023
2024
2025
10% better than std
# of yrs before creating
deficits
58.0
283
111
269
261
241
230
219
208
198
67.0
312
308
305
298
276
263
250
238
111
74.0
312
308
305
303
301
289
275
262
249
58.0
4
3
2
2
3
3
67.0
4
3
3
2
3
3
74.0
6
5
5
4
3
3
Footprint
2017
2018
2019
2020
2021
2022
2023
2024
2025
Footprint
58.0
315
308
299
290
268
255
243
231
220
67.0
347
342
339
331
307
292
278
265
252
74.0
347
342
339
337
335
321
306
291
277
2017
2018
2019
2020
2021
2022
2023
2024
2025
avg
3.4
15% better than std
# of yrs before creating deficits
58.0
268
262
254
246
228
217
207
197
187
67.0
295
290
288
281
261
248
236
225
214
74.0
295
290
288
286
285
273
260
247
236
58.0
5
4
4
3
4
67.0
5
5
4
3
4
74.0
7
7
6
5
4
Footprint
2017
2018
2019
2020
2021
2022
2023
2024
2025
Footprint
58.0
315
308
299
290
268
255
243
231
220
67.0
347
342
339
331
307
292
278
265
252
74.0
347
342
339
337
335
321
306
291
277
2017
2018
2019
2020
2021
2022
2023
2024
2025
20%
better than
std
avg 4.7
# of yrs before creating deficits
58.0
252
246
239
232
215
204
194
185
176
67.0
278
273
271
265
245
234
223
212
202
74.0
278
273
271
269
268
257
244
233
222
58.0
6
5
5
4
67.0
7
6
5
5
74.0
8
avg
5.7
5-120
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
5-121
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
References:
1 Schwarz, W., Harnisch, J. 2003, "Establishing Leakage Rates of Mobile Air Conditioners,"
Prepared for the European Commission (DG Environment), Doc B4-
3040/2002/337136/MAR/C1. This document is available in Docket EPA-HQ-OAR-2009-
0472-0157.
2 Vincent, R., Cleary, K., Ayala, A., Corey, R., "Emissions of HFC-134a from Light-Duty
Vehicles in California," SAE 2004-01-2256, 2004. This document is available in Docket
EPA-HQ-OAR-2009-0472-0186.
3 EPA, 2009, "Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990-2007."
http://www.epa.gov/climatechange/emissions/usinventoryreport.html. This document is
available in Docket EPA-HQ-OAR-2009-0472.
4 General Motors Press Release, "GM First to Market Greenhouse Gas-Friendly Air
Conditioning Refrigerant in U.S.," July 23, 2010.
5 Malvicino, C., Seccardini, R,, et al, "B-COOL Project - Ford Ka adn Fiat Panda R-744
MAC Systems," SAE 2009-01-0967, 2009.
6 Ghodbane, M., Craig, T., Baker, J., "Demonstration of Energy-Efficient Secondary Loop
HFC-152a Mobile Air Conditioning System - Final Report," U.S. Environmental Protection
Agency Report EP07H001055,
http://www.epa.gOv/cpd/mac/l52a/EPASec%20Loop 07Jan08.pdf.
7 Ghodbane, M., Baker, J., Kadle, P., "Potential Applications of R-152a Refrigerant in
Vehicle Climate Control Part II," SAE 2004-01-0918, 2004.
8 Peral-Antunez, E., "Recent Experiences in MAC System Development - New Alternative
Refrigerant Assessment Technical Update," SAE Alternative Refrigerant System Efficiency
Symposium, Scottsdale, AZ, September 28, 2011.
9 Li, X., "Alternative Refrigerant AC6 Performance Evaluation and Optimization Potential,"
SAE Alternative Refrigerant System Efficiency Symposium, Scottsdale, AZ, September 28,
2011.
10 Society of Automotive Engineers Surface Vehicle Standard J2727, issued August, 2008,
http://www.sae.org. This document is available in Docket EPA-HQ-OAR-2009-0472-0160.
11 State of California, Manufacturers Advisory Correspondence MAC #2009-01,
"Implementation of the New Environmental Performance Label,"
http://www.arb.ca.gov/msprog/macs/mac0901/mac0901.pdf. This document is available in
Docket EPA-HQ-OAR-2009-0175.
5-122
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
12 State of Minnesota, "Reporting Leakage Rates of HFC-134a from Mobile Air
Conditioners," http://www.pca.state.mn.us/climatechange/mac-letter-082908.pdf. This
document is available in Docket EPA-HQ-OAR-2009-0472-0178.
13 Office of Energy Efficiency and Renewable Energy, U.S. Department of Energy,
"Transportation Energy Data Book: Edition 27," 2008. This document is available in Docket
EPA-HQ-OAR-2009-0472
14 EPA, "Fuel Economy Labeling of Motor Vehicle Revisions to Improve Calculation of Fuel
Economy Estimates - Final Technical Support Document," 179 pp, 2.1MB, EPA420-R-06-
017, December, 2006. This document is available in Docket EPA-HQ-OAR-2009-0472.
15 Ikegama, T., Kikuchi, K., "Field Test Results and Correlation with SAEJ2727,"
Proceedings of the SAE 7* Alternative Refrigerant Systems Symposium, 2006. This
document is available in Docket EPA-HQ-OAR-2009-0472.
16 Atkinson, W., Baker, J., Ikegami, T., Nickels, P., "Revised SAEJ2727: SAE Interior
Climate Control Standards Committee Presentation to the European Commission," 2006.
This document is available in Docket EPA-HQ-OAR-2009-0472.
17 Burnette, A., Baker, R., "A Study of R134a Leaks in Heavy Duty Equipment," CARB
Contract No. 06-342, http://www.arb.ca.gov/research/seminars/baker3/baker3.pdf.
1&
Society of Automotive Engineers, "EVIAC Team 1 - Refrigerant Leakage Reduction, Final
Report to Sponsors," 2006. This document is available in Docket EPA-HQ-OAR-2009-0472-
0188.
19 Minnesota Pollution Control Agency, "Model Year 2009 Leakage Rate List,"
http://www.pca.state.mn.us/climatechange/mobileair.html. This document is available in
Docket EPA-HQ-OAR-2009-0472-0161.
20 Society of Automotive Engineers, "EVIAC Team 4 - Reducing Refrigerant Emissions at
Service and Vehicle End of Life," June, 2007. This document is available in Docket EPA-
HQ-OAR-2009-0472.
91
Schwarz, W., "Emission of Refrigerant R-134a from Mobile Air-Conditioning Systems,"
Study conducted for the German Federal Environmental Office, September, 2001. This
document is available in Docket EPA-HQ-OAR-2009-0472.
99 _
"A/C Triage - Ensuring Replacement Compressor Survival, " AC Delco Tech Connect,
Volume 12, Number 5, January/February, 2005,
http://www.airsept.com/Articles/CompressorGuard/ACDelcoTechConnectJanFeb05.pdf. This
document is available in Docket EPA-HQ-OAR-2009-0472-0167. This document is available
in Docket EPA-HQ-OAR-2009-0472-0167.
5-123
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
23 Johnson, V., "Fuel Used for Vehicle Air Conditioning: A State-by-State Thermal Comfort-
Based Approach," SAE 2002-01-1957, 2002. This document is available in Docket EPA-HQ-
OAR-2009-0472-0179.
9zl
Rugh, J., Johnson, V., Andersen, S., "Significant Fuel Savings and Emission Reductions by
Improving Vehicle Air Conditioning," Mobile Air Conditioning Summit, Washington DC.,
April 14-15, 2004. This document is available in Docket EPA-HQ-OAR-2009-0472-0179.
9S
California Environmental Protection Agency Air Resources Board, "STAFF REPORT:
Initial statement of reasons for proposed rulemaking, public hearing to consider adoption of
regulations to control greenhouse gas emissions from motor vehicles," 2004. This document
is available in Docket EPA-HQ-OAR-2009-0472.
r\r
"Reducing Greenhouse Gas Emissions from Light-Duty Motor Vehicles," Northeast States
Center for a Clean Air Future, September, 2004. This document is available in Docket EPA-
HQ-OAR-2009-0472.
97
Lee, S., Lee, B., Zheng, H., Sze, C., Quinones, L., and Sanchez, J., "Development of
Greenhouse Gas Emissions Model for 2014-2017 Heavy- and Medium-Duty Vehicle
Compliance," SAE 2011 Commercial Vehicle Engineering Congress, Chicago, September
2011, SAE Paper 2011-01-2188.
9R
Forrest, W.O., "Air Conditioning and Gas Guzzler Tax Credits," Society of Automotive
Engineers, International Congress & Exposition, Detroit, Michigan, March 2002, SAE Paper
2002-01-1958.
29 Forrest, W.O. and Bhatti, M.S., "Energy Efficient Automotive Air Conditioning System,"
Society of Automotive Engineers, International Congress & Exposition, Detroit, Michigan,
March 2002, SAE Paper 2002-01-0229.
30 Johnson, V.H., "Fuel Used for Vehicle Air Condition: A State-by-State Thermal Comfort-
Based Approach," Society of Automotive Engineers, International Congress & Exposition,
Detroit, Michigan, March 2002, SAE Paper 2002-01-1957.
31 Rugh, J.P., Hovland, V., Andersen, S.O., "Significant Fuel Savings and Emissions
Reductions by Improving Vehicle Air Conditioning," Presentation at the 15th Annual Earth
Technologies Forum and Mobile Air Conditioning Summit, April 15, 2004.
^9
Northeast States Center for a Clean Air Future, "Reducing Greenhouse Gas Emissions from
Light-Duty Motor Vehicles," September, 2004.
33 EPA and DOT, "Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate
Average Fuel Economy Standards: Final Rule," May 7, 2010.
34 Society of Automotive Engineers, "EVIAC Team 2 - Improved Efficiency, Final Report,"
April 2006.
5-124
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
35 Memo to docket, "Meeting with Delphi and Presentation to EPA," March, 2009. This
document is available in Docket EPA-HQ-OAR-2009-0472-0196.
36 Society of Automotive Engineers Surface Vehicle Standard J2765, issued October, 2008,
http://www.sae.org. This document is available in Docket EPA-HQ-OAR-2009-0472-0171.
37 Barbat, T. et. Al., "CFD Study of Phase Separators in A/C Automotive Systems," SAE
2003-01-0736, 2003. This document is available in Docket EPA-HQ-OAR-2009-0472-0187.
38 Memo to docket, "Idle Test Data Submitted by USCAR to EPA," July, 2011.
39 U.S. EPA, "Project Report: Computer Simulation of Light-Duty Vehicle Technologies for
Greenhouse Gas Emission Reduction in the 2020-2025 Timeframe", Contract No. EP-C-11-
007, Work Assignment 0-12, Docket ID: EPA-HQ-OAR-2010-0799, Document ID: EPA-
HQ-O AR-2010-0799-1144
40 Bradfield, M., "Improving Alternator Efficiency Measurably Reduces Fuel Costs", Remy,
Inc., 2008.
41 Schoettle, B., et al., "LEDS and Power Consumption of Exterior Automotive Lighting:
Implications for Gasoline and Electric Vehicles," University of Michigan Transportation
Research Institute, October, 2008.
42 American Honda Motor Co. comments, NHTSA Docket ID# NHTSA-2010-0131-0239,
EPA Docket ID# EPA-HQ-OAR-2010-0799-9489.
43 State of California Air Resources Board, Draft Manufacturer's Advisory Correspondence -
Optional Test Procedure for Certifying Model-Year 2016 and Later Vehicles Under the Cool
Cars Alternative Performance Compliance Option, March 3, 2010.
44 Rugh, J., Farrington, R. "Vehicle Ancillary Load Reduction Project Close-Out Report,"
National Renewable Energy Laboratory Technical Report NREL/TP-540-42454, January,
2008.
45 Rugh, J., et al., "Reduction in Vehicle Temperatures and Fuel Use from Cabin Ventilation,
Solar-Reflective Paint, and New Solar-Reflective Glazing," SAE 2007-01-1194, 2007.
46 EPA Technical Memorandum from James Sanchez and Ben Ellies, May 21, 2012: "Energy
Recovery Testing for GM Silverado 2-Mode Hybrid and Other Hybrid Light-duty Vehicles".
47 EPA e-mail memorandum from Ben Ellies to Don Kopinski, May 23, 2012, "Data for GM
2-mode hybrid pickup truck".
AR _
E-mail correspondence from James Sanchez to Eric Rask, May 3, 2012.
5-125
-------
Air Conditioning, Off-Cycle Credits, and Other Flexibilities
49 EPA memorandum from James Sanchez to Docket EPA-HQ-OAR-2010-0799, July 23,
2012: "Analysis of ANL's JrtEV Data to Evaluate Methods for Determining Nominal
Voltage".
5-126
------- |